Well after 2 years into getting a UHU to run a cnc router I have learnt a lot of things.

Since I strongly believe that I can make it work against all odds,,,I decided to have my own thread continuing my path towards making the UHU a success on my machine!

here is the latest update!

After the power supply issues were sorted out I connected the 110V supply to the HP UHU, initially the motor locked and when I tried to run Mach and start the motor tuning the motor started vibrating - approximately the shaft was having forward and backward movements of around 40degree to 60 degree contineously, the motor also became very hot!

next the error issue continues as stated earlier - I set the error as per Henrik's suggestion to around 30 and it when I run the circle g code it came to error very soon after less than a turn. again it kept on oscillating.

after a few minutes of oscillating the oscillations stopped suddenly and the motor became free wheel.

Now i decided that I would not test this board and go ahead and build a new one as the older one had gone thru all the odds!

Now my question is

I had removed the 1uF 400v film capacitors across the larger capacitors in the power supply, are those causing the oscillations?

how to suppress the oscillations and the motor getting hot issue!

How exactly to determine if the motor is of a good quality? do you guys think that I should disassemble and check?

does any one of you think that moving to higher voltages might introduce interfering signals on the encoder line (my encoder cable is shielded and the shielding is grounded on the machine side to the mains ground.

one more thing is that many of my friends here state that the mosfets are very delicate, if so can any one guide me how to handle them exactly.

well thats it for now

Thanks for all the help folks!

Regards
IRfan

HI folks was finally happy today the machine moved very well!

the parameters set on the UHU were
P - 1000
I - 21
D - 255
T - 150
M - 0
H - 7000

C was default 10000 and I could not change it ?

problems

1. was able to set Error to 2000 and not below? this is very saddning, at this time the voltage at the drive was 105V, I did not measure the current, when I increased the current setting on the UHU the value improved to 1950, but still the UHU would fault when the g0 cycles were ending, I really dont know if this is the problem with the motor if it is really heavy.

2.G0 cycles in mach gave very jerky motions this is evident in the video below I initially thought that tthis would be mechanical problem but it persisted even after the belt was removed ie., the motor jerked in between!

YouTube - uhu cnc

I did not note down what IPM it was.

hope some one can help me here now.

RGDS
Irfan

http://www.youtube.com/watch?v=MlTBGdx7zHc

Hi Irfan,
First, the jerking. Looking at the video it seems like if the step-pulses from the computer is suddenly interrupted. This indicates that the computer running Mach3 has one or more processes running in the background that interferes with Mach3's timing. Energy-saving apps, multimedia-apps (Quicktime especially) and on board graphics drivers are all known to cause to this.

Try running msconfig.exe and turn off everything that is in the Startup-tab, then reboot and run drivertest.exe located in the mach3 folder, when it runs it should look something like the screenshot attatched to this message. The less "noice" there is on the "trace" the better - mine is quite good.

I'm pretty sure the jerking you're experiecing is due to the computer.

Now, the Error, you need to adjust the acceleration setting in Mach3 to match what the machine actually can acomplish. If you have the accel setting maxed out it may try to accelerate/deccelerate the machine faster than what it can keep up with. You really shouldn't need to set the Error-setting above 100.

I'm also a bit concerned about the C-setting, are you sure it says 10000? The manual says that default is 63 and allowd values is between 30 and 200.

Hang in there Irfan!

/Henrik.

Hi Henrik,

was at the inlaws today, will try the settings tomorrow or Tuesday, anywayz I am trying to change the computer ,

the one running Mach is a celeron 500MHz 256mb ram vintage and the new one is a celeron 1.7GHz 256mb ram.

hope it will do better on the new computer.

I will try setting a lower acceleration values in mach.

well its true I cannot set the C values? how they affect the performance I dont know!

well I am going tomorrow to get extra parts for 3 more HP UHU boards tomorrow.

cya later all!

RGDS
Irfan

Hi Irfan,
Although I have heard about people running Mach3 on a 500MHz it is way below the minimum requirements of 1Ghz - I'm actually surprised it runs at all....

The C-value is the recalculation interval expressed in tens of microseconds, that means how often the UHU-processor runs the PID-loop. A value of 10000means that the PID-loop is run at around 10Hz which really shouldn't work at all. The default value is 63 which equals to around 1.5kHz.

I have NO idea why your says 10000 or why you can't change it. Is there any chance you might have change the C-setting by accident? Can you try changing to UHU-processor to another one?

Hi Henrik,

had been to get the parts for the UHU, well it does work well in the 500MHzm i will convert that to just a UHU parameter setting PC from now on.

I will try changing the processor tomorrow and update you.

RGDS
IRfan

Hi Henrik,

had been to get the parts for the UHU, well it does work well in the 500MHzm i will convert that to just a UHU parameter setting PC from now on.

I will try changing the processor tomorrow and update you.

RGDS
IRfan

Thanks for the updates!,

Kreutz.

Hi Irfan,
Any progress (or setbacks) on the issue?

/Henrik.

HI Henrik,

I have been trying to get the E stop online and working toward that side. wanted to be safer than sorry! but its taking more time than what I estimated and I am also getting th e proximities set up for limits and home, meanwhile I had to build two more HP UHU's as currently I have only one working.

Last week was quite busy at office, with late nights and long hours. so work on the router got delayed further.

I fixed up the Y- axis motor yesterday and put the new timing belt in. Just completing what I can - when I can. The software tuning parts takes most of the time so kept it for this weekend.

will update you once I have more data to share with every one, and yes I made the heatsink clamps with 3mm thk hylam pasted using super glue with nuts in between.

one more thing .... henrik you were right on the computer end, the 500MHz celeron does give problems with the curve, the new computer dosent!

RGDS
Irfan

one more thing .... henrik you were right on the computer end, the 500MHz celeron does give problems with the curve, the new computer dosent!

That's what I wanted to hear... :) You are a hard working man, Irfan - keep it up but don't forget to sit back and enjoy once in a while too!

/Henrik.

Hi all,

progress on new UHU boards.

got 3 assembled ....

contactirfu,
I thought I'd post what we were talking about off line. If I'm right or not, we should get some additional help'
Irfanulla & I have been laboring what to do with the Jumper's "A" &"B". if one is using an outside power supply (5vdc) on the E-Stop Button circuit Here's the answer I wrote................. Be gentle

Irfanulla, If you look at the PCB, starting with the top side (component
side) with J8 & J9 closest to you. Looking closely where J8 & J9 are
indicated. Look real close just to the right of the right most pin of
J9 (pin#3), you will see the "A" designating Jumper A this is jumper A
and would be cut IF your using external 5vdc power on the E-Stop
Button.
Turn the PCB over & now looking at the solder side you can see J8 is
now on our right. the third pin to the left has the "B" Indicating
the trace that goes to U2 pin#3. this is the first jumper "B" that
needs to be cut if using external 5vdc on the E-Stop button. the
second "B" on the solder side, left to J9 pin#3, this trace goes to
C42, Cut this trace & you should be all set !
Now having said that, I'm thinking that we should be able to connect
our E-Stop Button, one side to the External GND (J9 pin#3) & the other
E-Stop contact to J9 pin#2, NOT CUT ANY TRACES & be good to go
Paul :stickpoke

Hi Folk -

as kreutz and paul were guiding towards making the E stop's work for me , I have a good breakthru and a setback

lets talk about the good stuff first, I was able to achieve 3750mm / min thats 150ipm on the x axis, then set up a newly assembled board

so here is the bad thing that happened to me!

fist tests were good on the new board, motor locked and everything alright

then to start up the computer I put off the powersupply and then restarted the PS . usually i turn on the logic side first and then the Power side, this time both got turned on at the same time!

I was happily chatting with my assistant that everything was goin so well, then it happened!

C23 abd C26 melted out, the c23 tantalum was charred beyond recognition and the MKT meted with a lot of smell!

the heat charred a part of the PCB as well, hae to yet see if I can recover the circuit.

hope I can. here are the pictures.

also one question - any one know how to set the max speed of the machine in Mach3....I was able to take it to 3750mm/min max, I know it can go more, so any ideas.

thats all for today.

more later this week

Hi Irfan,
Sorry to hear the drive let the smoke out...any ideas why yet? Any chance you had C23 mounted backwards?

Regarding Mach3, what is the resolution of your encoder? If you go to Config/MotorTuning and have the Velocity slider all the way up to max, the next thing you can do is raise Mach3's kernal frequency. As standard it is 25kHz but can be set all the way up to 100kHz (it is limited in the demo though). You set it in Config/Ports and Pins and you need to restart Mach3 and retune the motors after you change it.

You should never go to a higher kernal speed unless the velocity slider is maxed out and you need more because the higher you go the more stress is put on the computer and you may feel Mach3 gets slow to respond and such...

If you can't raise the kernal frequency you need to use the step-multiplier in the UHU. It will give you more speed but it will trade it for lower resolution.

So, what's resolution on the encoder?

/Henrik.

Irfan;

What about R28, did it get burnt? What voltage rate was your tantalum capacitor? What diodes are you using on D4 and D5?

The fact that you turn two power supplies on at the same time does not mean that the output voltages get there simultaneously, it all depends on the time constant of each power supply, so the better way to make sure you have logic voltage before high voltage, is a sequential turn on with a few seconds difference. The same is true for the turn off, High voltage should rise later and drop before the logic voltage does.

Practically your logic voltage should be part of the of the high voltage turn on contactor coil circuit, so when no valid logic voltage is present it inhibits the high voltage turn on.

You might recover the board. Clean it well, check/replace all related components including the Mosfet drivers. Then test it at low voltage (about 24V) before hooking it to the high voltage power supply.

Hi Irfan,
Sorry to hear the drive let the smoke out...any ideas why yet? Any chance you had C23 mounted backwards?

nope.



Regarding Mach3, what is the resolution of your encoder? If you go to Config/MotorTuning and have the Velocity slider all the way up to max, the next thing you can do is raise Mach3's kernal frequency. As standard it is 25kHz but can be set all the way up to 100kHz (it is limited in the demo though). You set it in Config/Ports and Pins and you need to restart Mach3 and retune the motors after you change it.

Resolution is 250 PPR, I will try upping the kernel frequency


You should never go to a higher kernal speed unless the velocity slider is maxed out and you need more because the higher you go the more stress is put on the computer and you may feel Mach3 gets slow to respond and such...



if thats the problem then slower would also be enough!



If you can't raise the kernal frequency you need to use the step-multiplier in the UHU. It will give you more speed but it will trade it for lower resolution.

So, what's resolution on the encoder?


/Henrik.

250 PPR

Think the problem was the clamp I made, it might have shorted out the pins below, i had reduced the length of the screws bcos they were sticking down, i think the heat sink bcos suspended in my case - resulted in contact between the screw and some thing down there and hence the fire works!

I will mount the heatsinks the older way unless I have more problems!

RGDS
Irfan

Irfan;

What about R28, did it get burnt? What voltage rate was your tantalum capacitor? What diodes are you using on D4 and D5?


R28 and R22 are intact, infact the board was working well before it burnt.


The fact that you turn two power supplies on at the same time does not mean that the output voltages get there simultaneously, it all depends on the time constant of each power supply, so the better way to make sure you have logic voltage before high voltage, is a sequential turn on with a few seconds difference. The same is true for the turn off, High voltage should rise later and drop before the logic voltage does.

Practically your logic voltage should be part of the of the high voltage turn on contactor coil circuit, so when no valid logic voltage is present it inhibits the high voltage turn on.




ya it is switching the power with the contactor, meaning the DC is switched by the contactor after a delay with a resistor in between to avoid the inrush .

there are 2 resistors - one in between mains and the next between bridge and the caps. again i feel it was a short some where :(



You might recover the board. Clean it well, check/replace all related components including the Mosfet drivers. Then test it at low voltage (about 24V) before hooking it to the high voltage power supply.

I have assembled 2 more boards which I will try. then I will try recovering the other 2 boards.

moreover I love the board on the X axis performs like a charm!

RGDS
Irfan

D4 and D5 are UF4004

Hi Irfan,
With 250ppr (1000cpr in 4X mode) you'll get the following:

25kHz = 1500rpm
35kHz = 2100rpm
45kHz = 2700rpm
60kHz = 3600rpm
65kHz = 3900rpm
75kHz = 4500rpm
100kHz = 6000rpm

This is the speed of the motor, you need to factor in any reduction gear ratio and the pitch of the screw, rack and pinion, belt or whatever to get the actual speed.

Henrik,

max out RPM of my motor is 2000 @ 110V, so I will get it to 35kHz and relax, think I can live with that speed, if i am not then I will make a 1:1 belt drive from the present 1:2, but later!

Thanks for your help henrik!

RGDS
Irfan

HI All,

just had another board smoking!

the C19 (10uF tantalum 25V) cap burned away.

i am really wondering why it would happen. all the other parts seem to be ok, i have chkd the values of resistors and they seem ok too.

in the previous board too, it was the tantalum which smoked! the inscription on the tantalum sayz 106 25v, donno if that is enough or need to go to 35V?

donno where I am missing what!

any ideas what I should chk before I smoke the 3rd (final) board away!

RGDS
Irfan

Just an update to the earlier post, this time the 110v DC was not even on, the polarity was right as I double checked it myself before supplying power to the board.

Just an update to the earlier post, this time the 110v DC was not even on, the polarity was right as I double checked it myself before supplying power to the board.

Replace your 15v power supply, it is not regulating to 15 V and possibly letting higher voltage go through even for short times. Unless that batch of tantalums is bad, they should not explode at 15 volts. Adding a high capacitance to the output of the power supply, let's say 4700 to 10000 uf /40V in parallel with a couple of 1 uf metalized film capacitors, could help filtering out transients too.

I never had those problems with my prototype, probably because I use a regulated 15V lab. power supply.

my only bet on the 15v supply is that the x-axis board is working well, I will chk the 15v supply tomorrow and let ya know abt it.

lets solve it. may be a lot of learning is required but we can do it I suppose!

I will order a new 15v smps supply for 6 amps soon.

meanwhile the burning of the tantalum does not mean that I have to replace the IR21844?

RGDS
Irfan

my only bet on the 15v supply is that the x-axis board is working well, I will chk the 15v supply tomorrow and let ya know abt it.

lets solve it. may be a lot of learning is required but we can do it I suppose!

I will order a new 15v smps supply for 6 amps soon.

meanwhile the burning of the tantalum does not mean that I have to replace the IR21844?

RGDS
Irfan

Maybe adding a fixed resistive load to your existing power supply will solve the problem, SMPS often require a minimum load to regulate. 47 ohms 10 Watt resistor will load it enough (0.32 amps). The average load of the combined HP UHU boards is probably not enough. Please, add also extra capacitance and test the power supply voltage before attaching it to the HP UHU boards.

Regards,

Kreutz.

Replace your 15v power supply, it is not regulating to 15 V and possibly letting higher voltage go through even for short times. Unless that batch of tantalums is bad, they should not explode at 15 volts. Adding a high capacitance to the output of the power supply, let's say 4700 to 10000 uf /40V in parallel with a couple of 1 uf metalized film capacitors, could help filtering out transients too.

I never had those problems with my prototype, probably because I use a regulated 15V lab. power supply.
If his power supply does not have a softstart, it could kill a 25V tantalum on only 15V. Some tantalums ( solid electrolyte ) can be damaged by current surges. We had the same problem in our electronics. We used a voltage rating of 2 to 3 times the supply voltage and the problem vanished. Later on we used softstart power supplys. So if he would use 35V or 60V tantalums his problem could go away.

If his power supply does not have a softstart, it could kill a 25V tantalum on only 15V. Some tantalums ( solid electrolyte ) can be damaged by current surges. We had the same problem in our electronics. We used a voltage rating of 2 to 3 times the supply voltage and the problem vanished. Later on we used softstart power supplys. So if he would use 35V or 60V tantalums his problem could go away.

Good point.

If his power supply does not have a softstart, it could kill a 25V tantalum on only 15V. Some tantalums ( solid electrolyte ) can be damaged by current surges. We had the same problem in our electronics. We used a voltage rating of 2 to 3 times the supply voltage and the problem vanished. Later on we used softstart power supplys. So if he would use 35V or 60V tantalums his problem could go away.

HI LeissKG / Kreutz,

Thanks for the inputs! today I wanted to replace the tantalum with 35 volt ones and try again!, but today there was a power cut and I could not even start the changes, meanwhile while cleaning the first burnt PCB - I found out that the heat had burned the PCB away and melted the neighbouring mkt cap, so I had to take away the material by scraping it and it left a hole of around 8 mm.

well I have got more IRFP264's today for trying em out.

hope I have better success this time.

meanwhile any info / links on soft start Power supplies?

RGDS
Irfan

Hi ALl,

it was a success, the usage of 35v tantalums helped and the HP UHU is working just great!

I have added 4700uF caps in parallel to the 15 V supply, found out one more thing, most Toshiba laptops work on the 15 V 6 amps PS so thats what I have abit the lower wattage version giving only 5 amps.

will post more later

RGDS
Irfan

Hi Irfan,
I'm pleased to hear you was able to fix it, thanks for letting us know!

/Henrik.

Hi Henrik,

I have decided to activate the relay( used for bypassing the powersupply surge resistors) using a delay switch instead of the resistor, the reason was that today the voltage on the 3 phase PS went down to 195V AC and due to this the relay would not function and the power was coming out of the 25 w resistor and went down to 45 V

well hope this works well now! let me know if I need to take care of anything before connecting the delay switch, i am setting it for 5 seconds.

RGDS
Irfan

Hi Irfan,
That should work just fine, 5 seconds is way more than what is needed, it'll still works fine though. Now let's hope there won't be any more dissasters at your end!

well I would call em challenges though, its all a learning thing for me, and a great experience. After all these happenning I am no more worried for disasters, I am more worried about completing it in time!

RGDS
Irfan

Contactirfu,
Knowing you outside of the list I should have realized this sooner. You attitude toward a continued " learning curve " is in itself a bonus to the rest of us, You keep plugging, intent on the goal, and keep your head clear of crap and not unlike the "Ant" attain your goal.
"Congratulations My Little Brother"
Paul

Hi All,

I plugged in the OMRON 60 second delay switch and set it for around 6 seconds, now I see the wonder of it! once power supply snapped and came in again, it protected any case of burnouts with high power.

well so far soo good.

Today was the day when I had to chk for the working and the deadly "O" value that henrik keeps on discussing.

Well there was the devil hidden there!

I have got "O" values upto 8 sometimes. that makes it bad. i had to to set a conservative value of 3200mm/min for the jog speeds, I found it enough for my first build anyways :)

the X and Y work great together, now only to get the Z axis redesigned and working.
only thing is that it will probably cost me around 300USD more than originally thought!

going thru all odds now!

RGDS
Irfan

Hi All,

Today I put in my third board - and then powered the 15V supply - the LED's sort of just blinked and I think the short protection circuit of the Toshiba 15v power supply kicked in and the 15v supply went down. When I checked the third board I realized that there was a short in the 15V supply.

Now the help I need is that some one had a short on the PCB - an you guys tell me who and where they had the short and did it short the 15v supply?

I am planning to chk all the rest of the PCB's I have from paul for any shorts.

also if you guys know let me know what components would have caused the short - if it is not the PCB..

Thanks in Advance!

Regards

Just chkd all the PCB's none of them have the short on the 15V line!

defective tantalum! 10uF 35V

Hi Folks, I put up the third final board and then everything now is working - but just this one thing - the yellow colored light the one on "clr" is always on now.

I have a habit that I dont bother about things unless they show up! :D

so now the yellow light has showed up and I dont exactly know what it means .... I mean I have read it several times during the discussions . but may be i am just a lazy a**, :)

well can anyone explain to me if it is trouble to have the yellow light on?

RGDS
irfan

Hi Irfan,
The LED marked CL is the current limit LED. It will light up when the drive limits the current to the motor. Have you done the mods to the LED circuit as per Kreutz's instructions or is the board as it was originally?

The LED should normaly not be lit, especially not when the motor isn't accellerating hard or moving under very heavy (too heavy) load. What do you have your Vref set to?

/Henrik.

yup I did the Mods!

the V ref is set to .079 or .08 I think, .As I see there is no load on the motor, do ya think i should increase the v ref?

what should I moniter when I have set any abnormally high V ref - motor heating?

RGDS
Irfan

here is the router moving 3 axis with the UHU

<object width="425" height="344"><param name="movie" value="http://www.youtube.com/v/XXGDtxylumY&hl=en&fs=1"></param><param name="allowFullScreen" value="true"></param><embed src="http://www.youtube.com/v/XXGDtxylumY&hl=en&fs=1" type="application/x-shockwave-flash" allowfullscreen="true" width="425" height="344"></embed></object>

Hi Irfan,
Did you cut the PCB trace between R38 and U9? The reason I ask is because I think that might not be a good idea since, as far as I can see, also disconnects the anode of D3 from U9 pin#1. This then takes the slow-limit circuit out of operation. I have PM'd Kreutz about this a while back and he promised to look into it but as you know he's quite busy.

Anyway, are you sure that you soldered the jumper wire to pin#2 on U6 and not pin#1?

Does the drive work otherwise, ie does the motor hold position and respond to step-pulses etc?

/Henrik.

Henrik, the drive does hold position and respond wonderfully!

but i have done none of the modifications you have spoken of, the only mod is the IN4148 diode shorting R54 and which then gets connected to ground - it is the same for all the 3 drives.

two work perfectly and this one has its cl led on!

the thing is that it is on even when the motor is stationary and not running the cl led is on, so I will try to connect it to some other motor and try.

hope everything turns out well.

RGDS
Irfan

Hmmm, ok...
Measure the voltage at PIN#1 of U9. If it's 0V or close to it check for solder bridges around the Q5 transistor and/or try replacing Q5. If it's ~5V check R26, R40 and C28 and/or try replacing U9.

/Henrik.

Thats a lot in a one liner Henrik, will try your modifications tomorrow.

Irfan,

Could you, please, fix the link to your latest video, a few posts back? I would like to see it....

Thanks,

Kreutz.

http://www.youtube.com/watch?v=XXGDtxylumY

click on the link above, alternatively

the video is live at this this thread

http://www.cnczone.com/forums/showthread.php?t=46998&page=11

Congratulations!!

Thanks,

Kreutz.

you are welcome Kreutz!

I only just picked up this thread, 10 points for sticking with it Irfan, good to see everything working !

I had trouble with tantalums myself, one just popped one day, even though the power supply was off, the main cap was still charged. I was talking to a friend after I gave him a demo of the machine, about 5 minutes after power down it just went bang and leapt off the board in little pieces...took me a little while to find what made the noise, only tiny legs left sticking out of the PCB :)

Do they have a shelf life ? mine were probably pretty old, never used though but maybe 5-10 years old.

Cheers.

Russell.

I dont know about shelf life - but have learnt that when ever using a tantalum - use the one rated for double the operating voltages,

Russell u seem to have a load of surplus electronics parts.

:D

RGDS
IRfan

Yeah, but I am slowly getting through them by blowing them up :D

Russell.

THe machine working alright now! have to yet chk the dimensional accuracy

chk this video out at thread

http://www.cnczone.com/forums/showthread.php?t=46998&page=12

yesterday I found some new things on the UHU - two of my boards error out in the following condition

1. No Motor Power supply applied.
2. The controller computer is not on.

The red lights come on after a period of 10 minutes.

however when there is power applied to the motors this does not happen.

well it hasn't caused any problems in machining - so I am not worried - just a concern to be shared of.

RGDS
Irfan

yesterday I found some new things on the UHU - two of my boards error out in the following condition

1. No Motor Power supply applied.
2. The controller computer is not on.

The red lights come on after a period of 10 minutes.

however when there is power applied to the motors this does not happen.

well it hasn't caused any problems in machining - so I am not worried - just a concern to be shared of.

RGDS
Irfan

Hello;

You are probably receiving noise on the step-direction interface which is being interpreted as step pulses. When the amount of steps adds to the programmed error limit the Red Led turns on. RF interference could be responsible for that behavior.

Best Regards,

Kreutz.

hmmm, I have used shielded cable and connected the shield to the BOB ground for the wires connecting the BOB to the drives, also this does not happen when motor PS is on, I am I missing some thing?

OK here is one new problem faced :

I have set the accln setting in mach to 200 on all drives and a velocity of 3000mm/min,

I am very satisfied on how the machine is responding. But there is a catch, the drives error out when I press the pause the program, this does not help. I am assuming this is due to the low error setting on the drives (30) at present. I am going to set it to 60 or 100.

I believe Gecko has it set to 128 or similar, so I should be good at 60 to 100 I think. I am extremely satisfied about the torque and how the machine performs, just WOW!,

but the only problem I am facing like Henrik is the error counts - At present I have decided to neglect those and continue working as I am just wood working now.

RGDS
IRfan

does this increase in error count change the Zero'ing of the axis?

Hi Irfan,
When you get O-counts it means that the UHU-chip has missed (or wrongfully interpreted due to noice) one ore more quadrature cycles. So it can not, for sure, know where the motor is. It may guess right but you can't really be sure. This means that it might very well loose it's home position.

This is why *I* will not accept anything other than 0 O-counts on my system, it HAS to work reliably or it's not working at all IMHO.

/Henrik.

Hi Irfan,
When you get O-counts it means that the UHU-chip has missed (or wrongfully interpreted due to noice) one ore more quadrature cycles. So it can not, for sure, know where the motor is. It may guess right but you can't really be sure. This means that it might very well loose it's home position.

This is why *I* will not accept anything other than 0 O-counts on my system, it HAS to work reliably or it's not working at all IMHO.

/Henrik.

I agree, there is no reason why we should settle for less. It is necessary to find the root cause of the O counts, and solve it at once.

Kreutz.

well well well,

I take bk my words and its with a sad face I am typing this. henrick you re right I faced the problem today, over-runs to a tune of 68 on X, 48 on Y and 0 on z (after one hr of cutting) bcos I was doing 2 d cutting. see pictures for your self - not good for even wood cutting. Very Sad.

I am moving to Plan B, I have asked my supplier to modify my torroid to give out 50V AC on the secondary and I will be moving to Gecko 320's which I already have. I know that I will be low on speed but then at least the business will start. and if some one later find a soln to this I will improvise.

since I have time till monday - this is what I am going to do.

1. Ground all the shields of the encoders to mains ground with shield near UHU floating
2. Enclose the US digital encoder and the 26LS31 driver in a metal enclosure - probably will be a pepsi can
3. have a 100uF 25 v capacitor in parallel with .1 ceramic on the differential driver board.
4. make a 23LS32 receiver on a general PCB and use it instead of the onboard 7*****1 decoder

hope things work out. But the pain thing that to everytime I chk it I have to run my machine for an hour to see the kind of over-runs I am getting.

thats it for now folks, pray for me that I will hve a solution soon.

RGDS
Irfan

HI Folks, I am trying to work on the Y axis motor to get the noise or what ever that is causing the over-runs.

1. hooked up the sheild wire at the encoder end to machine ground - no use instead it resulted in a very fast single over run count.
2. Lowered the acceleration value in Mach which led to reduced error count - 12 but it was still there.
3. Put in a 100UF cap on the differential driver board and no use.

Now I understand Henrik's view...

I have to just have some self pity.

To go over it.

wish me luck!

HI My power lines are not shielded and the power to the makita router is delivered by a wire which is also not shielded, might that be a problem, I am assuming not due to use of differential signals.

Hi Irfan,
It's with mixed emotions I type this... I'm sorry you are having problems, believe me, I know how frustrating it is, spending near three weeks trying to solve it. But I'm also kind of glad that you too have the problem because that means that it's not something I'm doing completely wrong.

The whole point with differential signals is to make them less sensitive to noice. Your router motor probably emits quite a bit of noice but really, with the differential wiring it 'should' not be a problem. Besides during my tests I didn't have any spindle motor, VFD or anything like that running.

Do you have the 2200pF caps installed? Have you tried both with and without them?

I'm having a really hard time figuring this one out. I had the damn thing working, I probably run it on the bench for 5-10 hours during several days without a single O-count, then put on new encoder and it stops working. Yes, it points towards the encoder but honestly I don't think there's anything wrong with it....

I'm still on the road so I can't make any more tests for a while yet.

/Henrik.

Hi Henrik, today i tried shielding the encoder and the differential driver using alu foil - each time I had it connected to ground once to the machine groound and next to the encoder gnd.

no use at all, one of th electronics engineer suggested to me that there might be noise on the PCB itself, but he has not gone thru the PCB schematics. Any way this is what I did in the images.

At this stage I am assuming that the problem might be the board layout as I have sorted all others.

henrik do you think making a 26LS32 circuit and send A B direct to HP UHU from the receiver circuit might help?

Kreutz would it be possible to analize the PCB layout for us?

RgDS
Irfan

yup still have the 2200pF cap on board.

Hi Irfan,
I'm thinking about noice on the PCB too. I wonder what would happen if you cut the short trace/jumper that connects motor GND and low voltage GND and have those GND's connected outside the board.... Kreutz, are you there?

If you have the time and motivation please try with an external diff receiver with a 26LS32 or other circuit, the more tests we can do to isolate the problem the better - I will do it too when I get back.

I also tried all kinds of shielding, had my diff driver rapped in foil, just like you but it made no difference. Then, when I removed the 2200pF it just worked.

Please try without the caps and see what happens!

/Henrik.

Henrik Tomorrow I am going to get my transformer hve a 54V secondary tap, so that I can hook up the Geckos, I am broke as hell until I get my next months Salary :)

I have to still buy a chinese spindle and hook it up too, because the router's noise has got neighbours complaining and thats gonna cost another 1000USD.

its all go go money no in in. I have to stop. Since I will still have my 78V Tap for the UHU's I will keep testing and keep all posted, hope I dont have to do all this. ITs still a lot of work.

ALso I have to build the 26LS32 circuit for the geckos, so will do it soon and test it on the UHU's too.

Keep u posted till then.

RGDS
Irfan

I'm thinking about noice on the PCB too. I wonder what would happen if you cut the short trace/jumper that connects motor GND and low voltage GND and have those GND's connected outside the board.... Kreutz, are you there?

Separating motor GND and low voltage GND is a good idea. I have been bitten by this in the past. You should try that the two cables do not run parallel to minimize interference. Another possibility is to separate the two GNDs with a noise suppression choke.

Klaus Leiss

Leiss what is a noise suppression choke? some sort of ferrite core winding?

Have to study the PCB in detail to find the gnd inter connect.

Hi Klaus,
Thanks for stepping in. On the HP-UHU board the low voltage GND and motor GND are 'separated' and connected to each other at one single point by a fairly short thick PCB trace. Do you think that cutting this trace and connect the GND's together outside the board will help?

Hmmm..cutting the connection trace and replacing it with a choke....that may be an idea, any chance of something going wrong by doing so? Any suggested value for the choke?

Irfan,
I don't have the PCB o files here but the GND's interconnect trace is to the left of OP-amplifer IC when holding the board so that the motor connector is facing away from you. It's marked on the board by a square around the two solder pads and there's a fairly thick trace going between them.

/Henrik.

Thanks Henrik, let me chk,, where is Kreutz, have to chk with him if what we are doin is right or wrong, more over there is no patience left with me. I am already gone by tuesday if we cannot resolve it then the HP UHU hit the cold storage.

Hi Irfan, how long are your encoder leads ?

Usually line drivers are used to eliminate noise for longer runs, I forget the distance but six metres (20ft) seems to ring a bell.

I just wonder if your differential circuitry is actually introducing noise, how hard is it to remove it for testing ?

I am just thinking out loud here, it seems like you and Henrik are having the same problem and have the same setup as far as diff. encoder signals.

My router did crazy things at one stage but this was due to no Brown out detector enabled on the AVR's, didn't know about that one, but the UHU is pre-programmed so I don't think that is the issue.

Do you have details on the line driver circuit you are using ?

Cheers.

Russell.

HI Folks,

Today took out the 220PF cap from near the diff receiver, now I have "0" error no matter how long I run the machine.

Now my straight theory would be that the cap connected to GND is introducing noise in the receiver and once that is taken care of we have zero error.

I am not sure why Henrik got his errors back, may be with out the cap its just too much for the receiver to handle ?

well Russell, the longest I have is 8 meters of encoder cable on the Z axis.

The line driver circuit is the same as the one US digital uses and as per below schematic
http://www.cnczone.com/forums/showpost.php?p=476579&postcount=18

RGDS
Irfan

Today took out the 220PF cap from near the diff receiver, now I have "0" error no matter how long I run the machine.


WOHOO!!!

It is alive! :cheers:

Good work and great to hear you have it sorted.

Russell.

HI Russell thank for the cheers, but I believe that its not the end, but the beginning.

we need to do a root cause analysis on this one.

RGDS
Irfan

Hi Klaus,
Thanks for stepping in. On the HP-UHU board the low voltage GND and motor GND are 'separated' and connected to each other at one single point by a fairly short thick PCB trace. Do you think that cutting this trace and connect the GND's together outside the board will help?

It seems the HP UHU Board has already a good GND separation. What one tries to achieve is that the Sensor GND and Power GND are separated, and that no sensor signal flows parallel to a power signal or above a power layer. The separate GND signals are connected at one point and this point ( star point ) is connected to the power supply. If this is not possible the power side is connected to the power supply so that no power current flows through your sensor side. So the question is, do you get the power for your encoder from the same part of the board that has the sensor input. If not, your motor current may shift your encoder GND from your sensor GND and you get wrong signals. Get your encoder power only directly from the power supply if this is your star point.

Cutting the GND connection on the board and moving the star point to the power supply can help if you get your cabling right ( either routed separately or the low power section is shielded ).

Hmmm..cutting the connection trace and replacing it with a choke....that may be an idea, any chance of something going wrong by doing so? Any suggested value for the choke?

We use usually a ferrite bead over a wire, in some cases a slightly larger inductor.

Klaus Leiss

For sure I am giving up on the HP UHU, its not at all stable, where as I chkd for overruns on a 1 hr, today when I had to run it for 3 hrs it had 6 counts of overruns,

there might be several reasons, but that enough for me - I cannot take in more of this.

I am overloaded - and will not continue

more over for 2 d cuts I have to set up different accln levels in mach at which any 3d cuts will create the drivers to error out.

..............

its very sad for me.

RGDS
Irfan

Irfan,

Oouch, you can't believe how much it hurts to hear you say that!

Although it was/is a bit strange and unconventional I was pretty sure we had fixed the problem by removing those 220pF caps and that my USDigital encoders simply aren't up to the task. Now I'm not THAT convinced anymore, if you don't have USDigital encoders that is.... ?

This is really strange....I've probably put 5-10 hours runtime in it with my AMT encoder without ever getting a single O-count (yet, as you know, by just changing encoder it's a dissaster) so I suspect something's up with your wiring or the encoder itself.

What's the modelnumber on your encoders? Is there ANY chance that you might have gone above the frequency that the UHU can handle? I found ~130kHz to be absoulte max with Multiplier set to 0.

I feel your pain Irfan, I really do.

/Henrik.

Henrik,

my encoders are US digital EP4's 250ppr.

I donno if I maxed out the frequency - but it was set to M=0, I am also assuming it may be my motors too!

they are PMDG 625watts with four brushes.

I really dont know if it can be the encoder wiring, I have got em enclosed fully in Aluminum foil, with the foil and the cable shield grounded to the 15V gnd.

I am sure henrik that removing that cap is just an eyewash - the thorn is stuck deep in eye!

I hope I am wrong here - buy you will know once you start making the cuts.

RGDS
Irfan

Sorry to hear it still isn't working Irfan, I wish I could fix it from here :(

One thing you could try though, you mentioned you had three Gecko's, if you plan to install them try an keep your wiring the same as you have it now, rather than ripping the whole thing apart and starting over, this way if the problem comes back then you know it is an enviroment problem and not the UHU's.

Of course if you have to change something to suit the Gecko's then do that. I hope your luck improves !

Cheers.

Russell.

Russell, thanks for the kind words,

I am sure the whole setup was put up by me without having the full insight into what I needed.

It was my fault and a costly one.

I am building the bigger mechmate and cannot invest in this one anymore. If at all I earn more from the mechmate then I am going to change the whole setup on this machine once and for all.

there is gonna be no more investment on this one. its enough!

and yes when i upgrade its gonna have the whole lot shielded wiring for router power and motor powersupply new motors and gecko's for sure!

thanks for all help folks!

RGDS
Irfan

Irfan,
Yeah, I should know by now that you've got the E4P - you've told me more than a few times...sorry about that. I really can't believe there's a "generic" problem with the USDigital encoders as they are being used all over the place and I've personally never read about problems like this. (But others have said they have so...)

You're probably NOT on the frequency limit since 250ppr encoders would give you 7800rpm before reaching 130kHz.

I'm really sorry you're giving up on this. I'm getting quite concerned that there indeed IS a problem with the drive that we're missing. There's just not enough people using the HP-UHU to tell for sure what's going on and with you quitting there's even less....(I DO understand though, I've been VERY close myself).

I don't know how many kits you've sold but perhaps you or Paul could e-mail everyone you've sold a kit to and ask them if they experience the same problem and what their setup looks like.

/Henrik.

Henrik, maybe I will come back - but now I have to earn some money - I literally have nothing in the bank!

I am off to build the mechmate and if that makes me some money I will be back working on this one.

RGDS
IRfan

I forgot,

i still have a non-expensive MOD to try - the 200ohm resistor which Kreutz advised me to have in between the encoder lines - I never put em in.

will try that this week end.

RGDS
Irffan

Hello;

I can feel the frustration, it is reasonable, after so many tries. I am still not convinced the drives and/or the UHU chip are to blame for the O counts, can't prove otherwise either :( .

Replacing the HP-UHU by Geckos, will solve one problem, that is, the Geckos will not show any "O" counts, but you will never know if the problem is really not there or if it is hidden by the Geckos.

Have you tried with another manufacturer's encoders?

After having finished the shop, I've also reached the bottom of my savings account (actually in red numbers). So I will need to wait a little more in order to build a couple of workbenches, and to rent an "all terrain" forklift in order to move the "pick and place" machine in.

So, any work on the DSPic interface, as well as my bipolar stepper controllers is in standby until I get to save a little more money to cover the research/development expenses...

Best regards,

Kreutz.

Hi,

Irfan really got (and have) me worried so I thought I'd do some testing here....wish I hadn't ;-)

I put my AMT encoder back on the motor this afternoon and wired up roughly 10m of encoder cable. I put ~5m around a VFD that drive an 1.1kW three-phase motor and the other ~5m on top of the servo-motor and HP-UHU drive.

If you look at the first photo attached you'll see the VFD with the encoder cable around it on the right side and the HP-UHU and motor on the bench. With the current limit set to 20A and powersupply to 130VDC I started to run my test-program. I also started the VFD and while running, I periodically changed the output frequency of the VFD up and down. I let it run like that for ~20minutes and then left it running while I went inside to watch some TV.

After another 45 minutes I went out to check on it and found the motor on the floor, together with the capacitorbank/dump circuit. The HP-UHU was hanging from the bench suspended in the cables.... The encoder connector had been janked out and the drive had faulted out (as it should) but the drive was still powered up.

The shield on the motorcable had shorted to one of the motor outputs and the result of that can be seen in photo number two. Atleast two of the MOSFETs are blown and probably atleast one of the driverchips. Fuse still intact though....

BUT, anyway...although I didn't think of taking a photo of the mess I DID check the O-parameter and it was still 0. Now, I don't know for how long it had been running before falling on the floor and I can't be SURE that the drive didn't short the logic supply and reset in the fall but atleast the O-count was 0 when I went inside after ~20minutes.

It looks like the motor survived, atleast it didn't land on the shaft, it seems, it still rotates freely and the UHU can still read the encoder so that's good. The drive shouldn't be to hard to repair - I hope.


Anyway, the primary purpose of the test AND this post was to say that I think there's either a generic problem with the USDigital / UHU combination OR the UHU is extraordinary sensitive to variations in dutycycle, phase shift, pulsewidth, timing or whatever and I just happen to have two encoders that doesn't suffer from any of it.

I've been where you are Irfan, (actually I still am conciddering I bought brand new USDigital encoders that doesn't work) but I actually DO think it's a problen with the encoder in my case and the encoder and/or wiring in your case since EVERY SINGLE test I've done after removing the 2200pF caps has been very successfull. (Except this one but that's another story)

Irfan, you DID remove both 2200pF and not just one didn't you?

And, as Kreutz says - IF this is a noice or wiring problem at your end, changing to Geckos MIGHT not help. They DO have a working filter on the input so if it IS noice then it might filter it out. The only way you'll notice if there's a problem is if/when the motor loses position. There's no CPU etc in the Gecko so you can't check things like that.

Anyway, I'm not trying to convince you to throw more money at this if you can't OR don't want to. I've already ordered the Renco encoders and I'll make sure to post the results once I get them. I was and still am quite worried though.....

Sorry for the ramble....

/Henrik.

Henrik and Kreutz,

thanks for the words - I did give a thought about this and decided not to give up on this. I am only going to delay this for a while.

the cables on my motors - router - are not shielded and I am assuming I need to upgrade to proper automation stuff.

I need to concentrate more on money making now. I have the place leased out for 3 years and already one year is up and I have not got back any investment and thats adding to the pressure.

I am just on Pause and I will come back.

BTW henrik once you finalize on the encoders let me know - i will pick up the successful ones.

RGDS
Irfan

henrik I removed just one 2200pF cap - did I have to remove both?

Irfan,
Yes! Absolutely! There's one capacitor per encoder channel, C10 and C13 - If you can find the time please, please remove both and give it a try...crossing fingers here...

Replacing motor and encoder cables with properly shileded ones is definetly a good idea and highly recomnended too!

Let us know!
/Henrik.

yup I already have shielded ones on the encoder cables - but the motor power cables,

whats the third wire on the power cable of your motor is it for ground? and where is it connected on the controller?

for a 8 amp router is a 1.5mmsq wire enough for power?

RGDS
Irfan

Hi Irfan,
My motor cable has four conductors, one is not used at all, one (the green/yellow) is connected to the motorhousing and then to ground/earth and two are for the armature obviously.

Try running each and every ground connection to and from one single point as shown in the very crude sketch I've attached. In my case this single point is screw on the heatsink for the UHU but it's temporary after doing all the testing. The only grounds I don't have wired to this single point right now are the GND and shield for the encoder - they are connected to GND in the DB15 connector.

This is usually what is recomended to do and IS conciddered good practise, however I didn't have it like this myself from the start and it STILL worked as long as DIDN'T have the 2200caps mounted and DIDN'T use the USDigital encoder.

So, please try removing the second 2200pF cap too and see what happens my gut feeling says that it'll work then.

Local rules and regulations should be followed but here and AFAIK all over 1.5mm² is OK for up to 10A and sometimes more than that. You should be fine with 1.5mm².

/Henrik.

Thanks for the info Henrik, My earthing is all driven to the mains earthing - since I have the motor ground to teh machine I have earth wire (4mmsq) running from mains earth to the machine and all axis.

Think that will work well.

I will remove all caps and chk on thursday which is a local holiday here. also inline are the 200ohm resistors on the encoder lines.

probably the never give up attitude has set in :)

btw - some times I have a doubt if the problems rests in my ball screws which might have a backlash?

measured backlash is zero - but over the distance I am not sure at all.

i have to chk that too :)

RGDS
Irfan

Yup 1.5mmsq is fine for 8 amps, especially if it is flexible cable, it has higher current carrying capacity than normal low strand count stuff like the stuff you wire houses with.

Just out of curiosity, what rating are your servo's, voltage and amps ?

Cheers.

Russell.

motors rated at 110V 7 amps continuous, HP UHU rated to 185V 20+ amps .
Powersupply is 96V 20AMPS

The motors are PMDC's with US digital EP4 encoders.

probably the never give up attitude has set in
Looks like it ;-) Seriously though, I've been very close to give up too, more than once, and if the Renco encoders doesn't do the trick it'll be the nail in the coffin for my HP-UHU's. Then I'll have to resort to industrial quality or wait for the new high voltage version of the VSD-E from Grantie Devices.

However, these are tests you can do without throwing more cash at it - the fact that you only removed one of the two caps gave me my hope back for sure. Make ONE change at the time when you go about to test. First remove the second cap, if that doesn't work check your grounding and then try the 200 ohm resisitor across the encoder pairs.

Backlash can be the cause of a lot of things but I PRETTY sure it can't be blamed for bad transistions on the encoder lines.

Again, good luck and make sure to let us know how it goes.

/Henrik.

Thanks for the re-assurances henrik, If not for you - I would have already lost this battle!

Kreutz thank you too - without you all this learning would never ever have had happened.

I shall keep all updated and post again on Thursday evening India time.

Folks today I plucked out all the 2200pF caps and tried it and voila no over-runs at all.

but WHAT the position is lost again. the Z which I consider has the best THK ball screw without any backlash almost 2 to 2.5 mm difference in the zero position for z axis, the x and y have their own displacement differences.

now that we have zero over-runs what to look for I dont know.

the program was running for about 20 minutes.

Ya Ya today is wednesday - I could not stop myself.

i am now sure that the caps showed the overruns and with the caps out they are just hidden?

may be may not be

but if I split the program and do the routing then everything seems to be alright, but I set the zero everytime I start a sub program,

like the below the centre was v carved first and then the stars and circle around.

it looked good but if done completely at once then it really get screwed up.

RGDS
Irfan

Hi Irfan,
No O-counts !! That sounds good - let's hope it stays that way even on long runs.

Backlash shouldn't show up as an acumulating error like what you are seeing AS LONG AS YOU DON'T have backlash compensation turned ON and have entered the wrong distance to compensate for. Best to make sure backlash compensation is OFF for now.

Do you have a dial test indicator? If so you can set it up on the Z-axis and run, say 50 moves of the same distance up and down and watch the needle on the indicator as it runs. Do you loose a little bit of travel for each return to 0?

If you don't have a dial test indicator do the above test a copule of times and check how much travel you loose each time thru the test. Is it always the same amount and in the same direction or does it vary? (+0.5mm frist time, -2mm second time etc)

What you can try is to change the direction prechange value in Mach3, go to Config/MotorTuning and raise the Dir Pulse-value. Restart Mach3 and check that the change did "stick". (There's been some problem reports lately about these settings not saving properly so it's best to make sure.)

Also make sure that the pulley on the motor and/or screw isn't slipping and that the encoder disc is tightly fitted to the shaft.

Start with simple programs, begin with X and Y and run a 200X200 square many times and see what happens. Run it with the router turned OFF first, then run it again with the router ON and see if it makes any difference.

Good luck and keep us posted!

/Henrik.

Irfan;

No "O" Counts = GOOD!!!.
The UHU chip will not hide encoder problems, so I guess your problem now is a mechanical problem, please, follow Henrik's suggestions.

Best regards,

Kreutz.

HI folks, yes I will chk the mechanicals on this week-end.

RGDs
Irfan

Hi Irfan,
My drive is back up and running again - it was that single MOSFET that was blown, nothing else :-) Well the PCB too of course....

Anyway, I did the test I was going to do the other day. Same setup as before, roughly 10m encodercable, half of which was around the VFD and the other half on the bench, across the motor and drive. The AMT-encoder was set to 500 lines (2000 steps/rev), motortuning in Mach3 set to 1800rpm max (60kHz step frequency), PulseWidth set to 0 and DirectionPrechange set to 0.

The drive was connected directly to the computer (no breakoutboard etc) and the LPT-port outputs 3.3V only, fluorecent lights in the ceiling and VFD running together with my testprogram for 2h38min. There were zero O-counts and the motorshaft was, as far I could see, exactly where it was when I started. (I only had a mark from a pen to judge from but it looked to be spot on).

Why it doesn't work with MY USDigital encoders is still a mistery but with this test being successfull I really think it has to do with the encoders and not the drive.

/Henrik.

I had a quick look around and found this thread :

http://www.cnczone.com/forums/showthread.php?t=8694

Which although the links don't work may indicate you are onto something about the encoders.

I did some more searching and found this on the US Digital site :

http://www.usdigital.com/products/interfaces/encoder/cable-drivers/pc4/

Basically according to the little paragraph at the top the encoders can sink 4mA and source 200uA which isn't much, I don't know if your encoders are the same or similiar but this may be worth checking out, mainly how the encoder inputs are handled on the UHU as far as current draw/sink. I don't have time to look any further (going on holidays tommorrow morning ...gotta start packing:)) but this should give you guys domething to look for.

Good Luck !

Russell.

Hi Russel,
I believe Irfan is using a TTL -> Differential converter based aorund the 26C31 IC, this should work fine with the 75115 receiver on the HP-UHU. The fact that the CPU doesn't register any invalid transitions indicates that the encoder IS working properly and that, if it's noice, it's probably on the step & direction lines or a mechanical problem.

In my case I have the E7P's with differential linedriver built in but for me it doesn't work as the CPU complains about invalid transitions. Even when feeding the A and B channel thru my own linedriver circuit (which works with other encoders). So my conclusion is that noice must be getting inside the encoder circuitry itslef and not the cable.

As I said, yesterday I ran the system for almost three hours, with 10m of encoder cable - no problems. BUT, as soon as I put the USDigital encoder on the motor it starts to act up. It doesn't matter if I use the linedriver inside the encoder or use my own. I doesn't matter if I use termination or not. It doesn't matter if I run at 40VDC or 130VDC motor supply. It doesn't matter if I have 1m or 10m encoder cable. It doesn't matter if run fast or slow. It simply doesn't work.

Have a great holiday!
/Henrik.

Hi Irfan,
Any progress or setbacks?

/Henrik.

henrik, its my brothers wedding - this friday - did not even put on my machine.

on the other hand - one of my friends manjeet mentioned to me that he had problems due to PP wires being too thin - which is same in my case to. I have purchased a .25 mm sq cs cable 20 core for the PP cable.

RGDs
IRfan

Hi Irfan,
OK, just make sure to keep us in the look once you do get back to the machine - the wedding is much more important though! Have a wonderful weekend!

/Henrik.

H.O. & All,
I'm in the process of starting over on my mill project. It would help allot if the following questions were answered/reanswered. Unless I'm missing something each axis requires it's own regenerative power dump. If I've been told that more than one drive can be attached to one RPD (regenerative power dump) ? in thinking about this it would seem that if axis drive servo X were slowing down to zero rpm and requiring the RPD to divert EMF power to the dump resistor. If at the same time axis Y is moving under load; milling for example would not the Y axis servo be effected by the power from X be diverted to the dump resistor ? It would seem, not understanding, that each axis should have it's own RPD ? ? I think I'm asking if 1.) each axis requires it's own RPD card & resistor pack ? Would you try & explain this to me. Irfanulla tells me that the terminals X1-1 & X1-2 are for possible connections to the optocoupler to give an external signal to BOB/computer etc., at some future time ? can you give me an example ?
of it's use ?
Paul

Hi Paul, basically the idea of the Regen power dump is to dissipate energy created by the servo when braking/reversing direction. The short story is that the servo acts as a generator and pushes the motor supply voltage up, the regen dump senses the increase in voltage and sends it to ground, usually by switching a MOSFET to ground via a resistor. Once the voltage goes down to an acceptable level the MOSFET is turned off again.

You only need one of these on the supply as it is common to all drives, I guess you could use one per axis but it would not really be needed.

I have heard of these being used on stepper systems as well.

Hope this helps.

Russell.

epineh,
Thanks, I had posted this basic question on two lists & between H.O.s & your answer I believe I've got it. Sometimes I just get stuck and can't work a electronic thought out to the point of understanding although I have all the pieces.....................
Thanks Again
Paul

No prob's Paul, there are a few schematics kicking around the place, I am away from home for a while so I can't upload any that I have. They are easy enough to find though but yell out if you need more help.

Cheers.

Russell.

Hi Folks - finally the problem was Resolved and it was so simple to do that.


The PP cable - which I had bought ready made was the problem - when I chkd the resistance on the cable it was 3 ohm on each. wow!

The one which I made myself had just 0.3 ohm.

I did carvings which were always a problem now and this time it was perfect and no offsets what so ever.

finally it was a good ending to all -

next thing to achieve is to get more IPM on the machine and make it take up more accln.

RGDS
Irfan

Irfan,
You have no idea how glad I am to read this. Congratulations!

Thank you for sticking to it! Except for Jozsi's lathe spindle I believe your machine is the only one up and running with the HP-UHU - so far. I know it's been a long and sometimes painfull journey but perhaps we have a pretty decent drive after all :-)

Let's hope my problems dissapear too with the new encoders, when ever they'll get here....

Please keep an eye on the O-count from time to time, just to make sure!

/Henrik.

Thank you Henrik, I am now ready for more HP UHU - if anything comes along sooner on Servo.

I am actually on the hunt for good motors.

The ones I have I cannot have high accln bcos the drives error out. - i believe its due to heavy mass on the motor rotors.

Yes it might all be a problem of matching stuff and just be lucky to get all good at once.

My best to you henrik.

RGDS
Irfan

Congratulations & Nice carving Irfan;)...Your clock start ticking from now, so show us your great ideas and wood working projects...;)

Regards
Khalid

Folks I am off to add another router to my shop - It will be a 50x100 bed -

it will be a mechmate - since I already have most for the machine this will be a fast build.

http://www.mechmate.com/forums/showthread.php?t=573

and yes I will be doing much more work - just keep watching.

RGDS
Irfan

Folks I am off to add another router to my shop - It will be a 50x100 bed -

it will be a mechmate - since I already have most for the machine this will be a fast build.

http://www.mechmate.com/forums/showthread.php?t=573

and yes I will be doing much more work - just keep watching.

RGDS
Irfan

Congratulations!! We will be waiting for the photos of the new machine.

Best Regards,

Kreutz.

Congratulations Irfan ! you only have what you deserve.

Good luck for the second build !

cnc2.

Great news Irfan !

To think it was a pesky PP cable (chair)

Good luck on the next build :)

Russell.

OK folks

I need to daisy chain the UHU's for E stop so that when one drive faults then all stop - Kreutz did email me a schematic - but it never went in my half dead brains - :)

I donno if I need to combine the e stop and the drive stop, at present I only want to chain up the UHU's so that they all fault when one does. how to do that -

next Q is when I set a accln as 170+ in Mach. the drives fault very soon - I am not able to go more on accln. what should I do? all 2 d cutting I have curves instead of sharp corners - and when I set the accln more then the drive fault.

Is the problem with motors? as they are just PMDC's with encoders? or I need more voltage? at present its 96V for 110V motors.

That's pretty much it for now.

Thanks for all the help folks.

RGDS
Irfan

Hi Irfan,
You should be able to just tie all three error I/O's together. They are pulled up to 5V internally and when one drive errors it will pull its error I/O to GND which will disable the remaining two drives.

Remeber that the 5V for the fault interface is connected to the 5V from the USB connector so either give it 5V via the USB connector OR the screw-terminals NOT both. Or you can cut the PCB traces as outlined in the schematics if you want to separate them completely.

The drawback here is that there's no way to know which drive "issued" the shut-down. I have a CUBLOC "PLC" in my system so I built a little interface and a bit of ladder-logic in the PLC so that they all stop when one faults but I can see which one it was that cause it.

Another thing I've done is to wire up a simple PC817 opto-coupler to the reset-jumper on the drive, that way I can reset the drive with a 24V signal without having long cables acting as antennas directly connected to the low level circuits on the drive.


Regarding the acceleration you need to either raise the current limit or re-tune the PID parameters so the motor gets "stiffer". Raising the voltage will only allow you to go faster - not accelerating faster. Acceleration requires torque and to get more torque you need to push more current thru the motor. You can also raise the following error limit (the E-parameter) but that doesn't really give you more acceleration it just prevents it from faulting.

HTH
/Henrik.

Please, Look at the following post: http://www.cnczone.com/forums/showpost.php?p=518309&postcount=3

Henrik and Kreutz - thank you - it was good that Me and Paul came up with the same issue in the same time.

and now I understand it good - I wil give it a shot later today.

RGDS
Irfan.

Hi Kreutz,

In my case the the Pc Vcc will be the BoB Vcc and the PC GND will be the BoB Gnd right? in that case do I need to cut the traces. I am assuming yes bcos that will be ext 5 v I will be applying? or do I dont need it and never use USB even to tune it with the traces as it is?

Henrik,

Thanks for that write up - I have to chk my current settings and increase them.

will try that today too. also let me know which PID settings influence the accln settings the most?

its a good thing that you wired all the resets direct - using the opto - I will also do that I will tired reseting all 3 everytime they fault. :)

I have my E stting to 200 already. I dont want it to go more

and do log down your Cubloc interface - some day I will switch to it too,

This CNC stuff is so exciting.

RGDS
Irfan

Hi Irfan,
Generally raising P will make it a bit stiffer but when you do that you usually need to follow up with the D and perhaps H too - they all act 'together'. In your case though you may already be saturating the output and if that's the case the only thing left is to increase the current.

If you have a scope you can look at the PWM coming from the UHU-chip if it comes close to 100% when accelerating you're saturated and need more current, if not re-tuning the PID will help.

/Henrik.

Henrik, if I am using a external 5 v then can I use the same to use to power the chips on the UHU instead of the USB power, in this case the 5v will be coming from the BoB.

Hi Irfan,
If you don't cut the A and B traces then the 5V and GND connection of the Fault interface is connected to the 5V and GND of the USB connector. If you supply the HP-UHU 5V from your breakeout board (and this 5V isn't the PC +5V) then you should NOT connect 5V from the PC to the USB.

So, to answer you question, yes, you can supply it with 5V from the breakoutboard but remember to not connect the USB cable from the computer.

/Henrik.

here is what I think it should be.

Hi Irfan,
Correct, the MAX232 interface is now being powerd by the BOB so you should not connect the USB cable.

I'm not really sure what the PC817 is doing in you schematic. Which side of the opto-coupler is connedted to the drives, the LED or the transistor? If it's supposed to disable the drives when you push the E-Stop button then the transistor side of the PC817 should be connected between the Fault I/O and GND - not between Fault I/O and 5V.

Remember, the drive disables/faults when the Fault I/O is pulled low.

/Henrik.

Henrik, thanks for the help,

this clearly means I need to add a relay in between the opto and the UHU's. the relay should be configured NC, when estop is pressed the circuit opens and so does the transistor side of the 817, in my case this makes it a high on the estop for Mach. I need to convert this high into a low by adding a relay in between.

instead of adding a relay in between I will add a 24V relay which will pull low when estop activates.

this will eliminate the 817.

will try this - now it is getting more clear to me.

Thanks once again

RGDS
Irfan

Henrik, thanks for the help,

this clearly means I need to add a relay in between the opto and the UHU's. the relay should be configured NC, when estop is pressed the circuit opens and so does the transistor side of the 817, in my case this makes it a high on the estop for Mach. I need to convert this high into a low by adding a relay in between.

instead of adding a relay in between I will add a 24V relay which will pull low when estop activates.

this will eliminate the 817.

will try this - now it is getting more clear to me.

Thanks once again

RGDS
Irfan

Irfan;

Let me explain a little about the reason behind each component in my E-Stop circuit.

The Auxiliary E-Stop voltage could be either DC or AC (see comment in bold below), it has to match the K1 coil specs. The E-Stop switches are closed in normal operation, and wired in series with K1 coil, so, if the auxiliary voltage fails, or any of them are pressed or open, K1 will de-energize.

K1 is a SPDT relay with N.C. (normally closed contacts) in the simplest of the configurations. As soon as K1 is de-energized, the contacts close, signaling a E-Stop event to each drive, and the CNC software.

C1 is a 0.1 uF ceramic capacitor, it forms a low pass filter in combination with R1 (10K) (in parallel with the pull up resistors on all the drives attached), so noise in the power supply line gets filtered before entering the LPT port pin., but its main function is to charge itself to VCC and discharge through the contacts of K1 when they close, the brief electrical arc will keep the relay contacts clean.

The diode D1 will take care of the surge voltage when suddenly de-energizing the relay coil in a DC circuit. When using AC to drive the coil, it should not be populated.

Additionally, a second contact pair (N.O.) of the K1 relay could be used in series with the Motor voltage power supply transformer's primary to de-energize the motors in case of E-Stop event.


Kreutz.

HI All,

today I lost my Z drive - I was trying to get more accln on my drives and in the process increased the current to max. the motors started getting hot and the driver too got hot, i am assuming the temperature shot up due to the revrse voltage - ..though I have a RPD set up (on a general PCB) I havent set it up to the correct voltage, as I was running a 3d file which had some z extremes, the z gave up and when I touched the heat sink it was all too hot!

I will try to print the RPD circuit and get that thing going - now I got to know the relation between RPD and accln.

RGDS
IRfan

If the mosfets have blown is there a chance that the IR21844's has blown too?

Irfan;

Let me explain a little about the reason behind each component in my E-Stop circuit.

.........................................................

Additionally, a second contact pair (N.O.) of the K1 relay could be used in series with the Motor voltage power supply transformer's primary to de-energize the motors in case of E-Stop event.


Kreutz.

I am going to use your circuit Kreutz, only one thing - is the Vcc (5V) from BoB needs to be connected to the HP UHU?

RGDS
Irfan

If the mosfets have blown is there a chance that the IR21844's has blown too?


Yes. Have you checked the mosfets yet?

I am going to use your circuit Kreutz, only one thing - is the Vcc (5V) from BoB needs to be connected to the HP UHU?

RGDS
Irfan

Yes, it needs to be connected, and USB connector's power traces cut at A and B sites (to make sure they are both not connected at the same time).

Regards,

Kreutz.

I think the Mosfet look clean - the heat on the heat sink was so much that I felt my skin burn.

But I chkd the resistance on the motor power input side - on the damaged board its 0.3ohm where as on the working board it starts varying due to the capacitance.

I am assuming its a short - and I already desoldered the mosfet (all) now only have to confirm if there is anything wrong with the IR21844's

RGDS
Irfan

Yes, it needs to be connected, and USB connector's power traces cut at A and B sites (to make sure they are both not connected at the same time).

Regards,

Kreutz.

But I was planning to remove the USB connector from the board and use only the power from BoB - but now I see that if I take that board out then it will be a problem to tune it.

will cut out the traces Kreutz.

RGDS
IRfan

But I was planning to remove the USB connector from the board and use only the power from BoB - but now I see that if I take that board out then it will be a problem to tune it.

will cut out the traces Kreutz.

RGDS
IRfan

Cut and leave the traces about 1 mm apart, so if you need to power the board's E-Stop circuit from the USB connection, in the future, you can re-solder them.

Note that the Rs-232 interface will still be powered from the USB connector after cutting the A and B jumper traces.

yup done, also i am replacing only the IRFP264n's and not the IR21844's

RGDS
Irfan

OK folks - i faced a problem and I got a solution too

THe Probem, - at around 0.09V setting for the current and running a 3d file which puts huge strains with rapid changes in directions resulted in the temperature of the mosfets going very high even with the heatsink. after a while the drives start to error out and if we still continue then the mosfets fail (I had to replace my Z axis stuff :(

Solution : I pulled up a table fan and put it on near the drivers - problem solved -

Inference - We need huge cooling for the drives.

I did check the working voltage when the 3d files was being cut - all voltages were like 92V to 95V and never went above. The motors are rated at 110V so I am assuming its not the voltage but the current which is causing the problems (heating up of mosfets)

donno what to infer here though.

RGDS
Irfan

OK folks - i faced a problem and I got a solution too

THe Probem, - at around 0.09V setting for the current and running a 3d file which puts huge strains with rapid changes in directions resulted in the temperature of the mosfets going very high even with the heatsink. after a while the drives start to error out and if we still continue then the mosfets fail (I had to replace my Z axis stuff :(

Solution : I pulled up a table fan and put it on near the drivers - problem solved -

Inference - We need huge cooling for the drives.

I did check the working voltage when the 3d files was being cut - all voltages were like 92V to 95V and never went above. The motors are rated at 110V so I am assuming its not the voltage but the current which is causing the problems (heating up of mosfets)

donno what to infer here though.

RGDS
Irfan

Hello;

You could get higher speed if you raise the motor voltage to 125-130Volts. High currents for long time will heat up the Mosfets, that is normal, you could use a better and bigger heat-sink (with lower thermal resistance) for passive cooling, or use forced air cooling, as you did.

What kind of heat-pads are you using between Mosfets and the heat-sink?, are you using heat-sink paste?. What is the ambient temperature?

Best Regards,

Kreutz.

Hello;

You could get higher speed if you raise the motor voltage to 125-130Volts.

What kind of heat-pads are you using between Mosfets and the heat-sink?, are you using heat-sink paste?. What is the ambient temperature?

Best Regards,

Kreutz.

Kreutz the speeds are programmed by the software(CAM soft) and are set at 72IPM, people are getting 500mm/s/s acceleration which I was trying to achieve.

Paste - I am using Alu oxide based silicone paste (pate 340 (http://www.rhodia-silicones.com/silicones/product_main.jsp?PRODUCT%3C%3Eprd_id=845524441860979&FOLDER%3C%3Efolder_id=2534374302025621))

I am using Mica pads as insulators - which I am not happy with. the paste goes on both sides.

ambient temperature is here (http://www.google.com/search?hl=en&client=firefox-a&rls=org.mozilla%3Aen-US%3Aofficial&hs=9UB&q=temperature+bangalore&btnG=Search)

RGDS
Irfan

Kreutz the speeds are programmed by the software(CAM soft) and are set at 72IPM, people are getting 500mm/s/s acceleration which I was trying to achieve.

Paste - I am using Alu oxide based silicone paste (pate 340 (http://www.rhodia-silicones.com/silicones/product_main.jsp?PRODUCT%3C%3Eprd_id=845524441860979&FOLDER%3C%3Efolder_id=2534374302025621))

I am using Mica pads as insulators - which I am not happy with. the paste goes on both sides.

ambient temperature is here (http://www.google.com/search?hl=en&client=firefox-a&rls=org.mozilla%3Aen-US%3Aofficial&hs=9UB&q=temperature+bangalore&btnG=Search)

RGDS
Irfan

Alu oxide paste is good enough, silver based pastes are better. Mica pads are OK, what is the thickness?. Do you have a photo of your heat-sink? Ambient Temperature is about the same as here..

What is your motor's rated max. speed?. If you work below the maximum voltage for the rated speed, you will never get that speed. Remember that the output stage PWM maximum duty cycles effectively drops the maximum voltage about 15% from the applied motor power supply.

If you increase power supply voltage you also increase maximum acceleration, because motor current will build up faster. Motor speed depends on motor voltage, torque (acceleration also) depends on current. What are your motor's specs?

Thanks,

Kreutz.

the heat sinks are here (http://www.cnczone.com/forums/showpost.php?p=396396&postcount=1530)

The motors max rated RPM is 1000, yup I know about the speed and I am happy at the speed I am getting - I can cut at 80 IPM which is really good for me.

based on your words I am thinking I need more voltage - probably I will buy a servo stabilizer and configure it to give 250V contineous which will increase secondary voltage of my transformer (or probably get a 10 amps variac and set it to give 240V)- or just change the present transformer ( which is also expensive)

will get thru some how to get more voltage though :)

RGDS
IRfan

the heat sinks are here (http://www.cnczone.com/forums/showpost.php?p=396396&postcount=1530)

The motors max rated RPM is 1000, yup I know about the speed and I am happy at the speed I am getting - I can cut at 80 IPM which is really good for me.

based on your words I am thinking I need more voltage - probably I will buy a servo stabilizer and configure it to give 250V contineous which will increase secondary voltage of my transformer (or probably get a 10 amps variac and set it to give 240V)- or just change the present transformer ( which is also expensive)

will get thru some how to get more voltage though :)

RGDS
IRfan

Thanks,

The heat-sinks seem a little small for the board's power (without forced air cooling), you will probably need to permanently attach a fan to move the air over and around them...

Best regards,

Kreutz.

OK will take care of that.

RGDS
IRfan

Folks my next question - can I use a variac to increase my AC input voltage?

my input voltage varies from 195 to 210Volts I need at least 230V output.

RGDS
IRfan

Folks my next question - can I use a variac to increase my AC input voltage?

my input voltage varies from 195 to 210Volts I need at least 230V output.

RGDS
IRfan

You can use one variac, the problem is that if you adjust it to get 230VAC, and the line voltage is 195VAC, if afterward suddenly the line voltage increases to 210VAC you will get 247.6VAC at the output. Then you must be very careful with the Regenerative energy dump device trigger level if it does not automatically sense line voltage fluctuation.

Your variac should also be rated to at least 1.15 times the KVA rating of the transformer you are using as Motor power supply.

The perfect solution would be an old type motorized variac type voltage stabilizer or a high power on-line UPS, the first solution was used a long time ago as line stabilizers for X-ray equipment (integrated into the equipment's motorized auto-transformers).

Kreutz.

The HP UHU's work well with a lot of cooling

here it is the proof - the following took 40 min and no drives erred - the accln was much higher this time 300mm/s/s

http://www.cnczone.com/forums/showpost.php?p=520815&postcount=188

thanks to all once again

RGDS
Irfan

Congratulations Irfan!
All your hard work is finally starting to pay off! I'm still waiting on my Renco encoders but your success makes me feel quite calm about actually get this thing going. Again, well done!!

/Henrik.

With your's , Kreutz and tenmetalman it would be impossible for me to do it!

well done to all of us.

RGDS
Irfan

Hi Irfan,
How is everything going? Drives are working? No lost motion or O-counts?

My encoders are on their way from the US, hopfully I'll have them the coming week - it HAS to work.... :-/

Drives are working good henrik, just that they need more cooling - continuous - I am assuming from what heat I am getting that U need forced cooling on your drives too. The heat sinks are just not enough, its not the speed but the accln value that gets them hot.

apart from this - they are working like a charm and not one work piece I have lost since forced cooling was put in.

I havent chkd the overrun counts any time after it all set in.

RGDS
Irfan

Hi Irfan,
Thanks for the update, makes me feel quite comfortable :-) Yep, the more current that passes thru the MOSFETs the more heat is produced, voltage (ie speed) has little effect on drive heating I believe. I'm surprised though that you need THAT much cooling, what's your Vref set to and what's your ambient temp, roughly?

Thank you again for the update!

/Henrik.

Irfan,

Have you posted a picture of your heatsinks here? I looked through the whole thread but could not find one. I'm interested to see what design you are using and to compare it with mine here.

I am constructing my electronics enclosure and will put a fan on the side panel, still wondering how much of an airflow it would need. I think in my case a 12cm 12V computer fan will do the job, and in the same time be very quiet. But that's just an assumption for now. :)

Thanks,

Todor

V ref set to .89 or .089 either of one - have started to forget already :)

I have set a high V ref , for the motors which are rated 7 amps continuous.

The motors get real hot when doing 3 d cuts though,

if not then I have to reduce teh accln so that heat produced is low - but at 300mm/s/s , I believe its already less :)

ambient is avg 79 D F.

RGDS
IRfan

Hi Irfan,
Thanks for the update, makes me feel quite comfortable :-) Yep, the more current that passes thru the MOSFETs the more heat is produced, voltage (ie speed) has little effect on drive heating I believe. I'm surprised though that you need THAT much cooling, what's your Vref set to and what's your ambient temp, roughly?

Thank you again for the update!

/Henrik.

It is normal, the heat-sinks are too small for the power handled by the boards which is more than 1 KW and higher peaks during acceleration on each of them.

There is another detail, for the heat-sink to work in still air, as intended, the fins should be pointing up, their function is to convey the heat out by radiation and convection, if you put it upside down, in still air, the convection heat tends to accumulate under the heat-sink's aluminum base, increasing the Mosfets' temperature. When you use forced air cooling the heat-sink position is irrelevant.

Regards,

Kreutz.

Irfan,

Have you posted a picture of your heatsinks here? I looked through the whole thread but could not find one. I'm interested to see what design you are using and to compare it with mine here.

I am constructing my electronics enclosure and will put a fan on the side panel, still wondering how much of an airflow it would need. I think in my case a 12cm 12V computer fan will do the job, and in the same time be very quiet. But that's just an assumption for now. :)

Thanks,

Todor

this is the link

http://www.cnczone.com/forums/showpost.php?p=396396&postcount=1530

OK,

So I see your heatsinks are not so small but the ribs are somehow close and are designed in my opinion to be used with a fan. Natural convection should be used with relatively bigger ribs that have more distance inbetween.

Of course the ribs should be in vertical position, I know I don't reinvent the wheel, it's just how I see things.

Todor

:) yes Todor - you are true about all you wrote, I in a haste did not consider many things - just wanted the router moving and fast.

But - future is bright - its all a learing experience for me -

have to make money to start making the changes though :)

RGDS
IRfan

Ok Now - I had to work today on some Teak wood doors -

As I am reactive and not pro-active - I decided to daisy chain the estops of the drives.

now Pin2 on J9 on all the boards were interconnected using a wire., to test it I set up high speed and accln so that the drives fault -

but surprise -one drive faulting did not get the other drives to fault too.

so I decided not to carve the teak wood today and consult you folks to get it right - before I spoil some more of the stuff!

ok folks tell me how to get it right - I am assuming that I am missing something here.

RGDS
IRfan

Irfan... buck up... I hope u will get the problem corrected... U are doing good R & D.. Keep it up man;)

Ah Khalid - there is no problem - just that I am missing a thing or two here :)

the problem being me and me only - I am not a proactive person :)

RGDS
Irfan

Hi Irfan,
Because the Fault I/O is isolated you also need to connect +5V to J9 pin1 and GND to J9 pin3 of all three drives. Then chain J9 pin2 of all drives together. If you don't do this the GND potential of one drive is not the same as the GND potential of the other drives.

Remember: If you haven't cut the A and B jumper traces the +5V and GND must come from the computer or you'll risk blowing something up if you connect the USB cable. Alternatively you can USE the USB-cable as the source for your +5V and GND. Connect it to one drive and then chain the drives together using the J9 terminals.

If you HAVE cut the A and B jumer traces OR you are SURE you won't connect the USB cable then the +5V and GND may come from a separate powersupply.

/Henrik.

Thanks Henrik, I will just connect one USB and chain the rest!

See I always miss things and thats where a real jumper will help - rather than cutting the track. :)

RGDS
IRfan

H.O.,Irfan
I just reread the posts about chaining the dive's I/O's together for the ? time & got it. Thanks
Paul

http://www.cnczone.com/forums/showpost.php?p=532618&postcount=201

Congratulations Irfan !

It's pleasant to hear it & see it moving.

Good luck with marketing !

cnc2.

Irfan,
when we were talking earlier I meant to tell you just how outstanding all of the work you had finished is. It just looks great ! the customer must have been pleased beyond words
Paul

OK Folks, off late the longer axis drive had started to error frequently and required me to set really low speeds on that drive - even when current is set very high (9 Amps) the motor being rated for 7 amps continuous.

I had a doubt on this drive bcos last month I had tested with this drive for extreme accln and deceln for a 3d tool path. without the fan for cooling this drive. the result I think is showing up now

So today I have replaced the Mosfets on this drive - was about to chk the drive but my office transport came it and I had to leave - I will post the results later when I have things moving

RGDS
IRfan

Folks,

I have did some thing not expect of me even by myself! :D

I replaced the FETs ok. but the problem of the axis faulting continued - now i was really furstrated as to why I was facing this problem.

I chk the UHU soft tuning I had the T set for 200 - low torque - major mistake at the cost of 4 FETS - which already were biting the dust for no fault of their's

hmm - have to be more careful now on :D LOL

I can only laugh at myself!

now the machine moves effortlessly at 3500mm/min at 200mm/s/s Accln - and yes without faults.

RGDS
IRfan

OK folks - I finally decided to have a new controller enclosure made for the HP UHU's - the present on is just too shabby. and dust sits every where - even on the drives and I occasionally dust them.

Henrik, if you are closer to finish your - kindly post some updated pictures on your thread .

the enclosure will take more than a week to be made and delivered - so in the meantime need to plan on the layout - any hints would be welcome.

RGDS
IRfan

Hi Folks - i have started rebuilding the controller - this time I have decided to do a minor change on the power side - previously i had 2 - 50 - OHM 25W resistors one each on primary and secondary of the transformer - now I am going to remove the one on the secondary and see what happens.

I am adding a resistor which will short the Caps on Estop and Power off. (this is accomplished using a relay working at 24V

all the controls are now working common on 24V

planning to wire in resets of UHU's to a common switch using 817 optos as Henrik has done it.

and there is a new cabinet - which has place to mount a future VFD for spindle

thats it for now - since I have lost my mobile phone and could not post any pics - will take some on my digi cam and try to post asap.

RGDS
IRfan

folks - when building the new controls for my machine - i face a peculiar problem.

One of the 220v relays is getting heated up.
the relay is rated at 220V and that is supplied to it via a relay (used for estop) 24v - the 220v relay inturn switches on off the 24v timer which is used to bypass the 50-OHM resistor on the main supply to the supply transformer TR1

any ideas and help are welcome.

RGDS
Irfan

How much current are you pushing thru the relay contacts and how much is the relay rated for? The voltage rating alone doesn't say much.... Also make sure that the coil is for 24VDC, measure the coil resistance and compare with the one that's not getting hot.

Henrik, the coil is rated at 220v - and I an using 220 mains to power it, actually it seems the relays from that manufacturer in general are of bad quality - I got a replacement which heats up too - just when connected direct to the wall socket - I am going to change the manufacturer.

RGDS
Irfan

See your separate post on this.
Al.

Hi,
I read old thread about uhu ,seams here you are still writting about. I received this days controller from Uli , and pieces on reichelt .de . I have soldered two controller board{from 6}.,and just i expected after i readed many experience before me..... i neeed helpp.
1. i cant test uhu board .. dont know..maybe encoder ..i put on pc those 2 green led are shining, error led is ok vhen i put error on uhu soft control, is ok pc see my board but i cant move my motor ..sometimes he goes away alone in one direction.
motor for testing is shinano kenshi dx050..encoder 200lines 90degrees 2 chanel 4 wire { i have other bigger but encoder on them are not ok for uhu-3 wire max 50 lines.."} please tell me is encoder is ok for uhu is TTL???i put some encoer specs atached here.
2.i tested voltage everithing is ok but i dont know german and as in picture in place of 0 volt is 5 on my board{dont know what say there about measuring where to put -of multimetter?)is a problem somebody cheked that diode and iss not ok 5 volts?{ i measured all pind with - on the uhu board an + on each pin all but there i got 5}
3. can you give me some layot schems and docs for encoder receiver emitter for differential lines?if my encoder is not ok i can update him with differental lines to work in uhu?
4.some simple step dir program for testing servo???or tell me step by step how to run from uhu soft{E 2000 G 30000??ANS HE WILL RUN?}
5.btw i want to build an wood router for sculpture..hope i succed and dont regret a lot of money sped without result
6 thx . i know i will find here help ..i watched you for about 2 years nice guis..

best regards: ben

forgot another thing to asc what about with those colored pins on ic's?
and please help to do breakboard for 3 axys..image detail ..{for me a beginer in electronic and this domain an image make more thamn 100 words:D and her e s my controller

forgot picture ...my first post ever:)

Hi Ben - I started a new thread for you - start posting there

http://www.cnczone.com/forums/showthread.php?t=73297

post pics of your motors and power supply too.

this is the inside of the new enclosure : looks neat right :)

Very nice Irfan! Clean and tidy, makes servicing and troubleshooting a lot easier. Make sure you document the system as well otherwise you'll be in the dark if something goes bad a few months down the road.

The HP-UHU does steal quite a bit of real-estate when mounted like that though it makes it a lot easier to connect and and adjust the current limit than the way I have mine mounted.

Again, good job!

/Henrik.

Thanks for your comments Henrik - after the conversion the touch screen is putting on too much heat - its an ELO SAW screen and even the glass gets hot.

I am proposing to add a recirculating fan inside the cabinet and a fan near the monitor which will run from the computer power supply .

hope that works :)

RGDS
Irfan

one more thing what you don't see is the forced cooling fans behind the HP UHU's,

there are 3 for 3, the only thing is that I need to have a maintainance schedule to chk on them.

HI folks - today I sourced a Voltage stabilizer which was earlier used for a Air conditioner- its rated at 4KW

I am planning to use the same to power the motors so that I have a pretty nice voltage - today I chkd it - while the input was 180 the output was 200. So I a sure it would work for me - just needed your inputs if there are any concerns arising out of the use of this voltage stabiliser.

RGDS
Irfan

Hi,

today i got my 3 HP-UHU-Kits :-)

what is the maximum tested voltage ?
My Maho has DC-Servo with a nominal voltage of 170V DC

are there any complete documentation on board rev. 10b ?
the partlist i've found - but nothing more .-(

last but not least: is there anyone in german speaking area who has build HP-UHUs ?

Regards
Frank

Frank - there is more information on the wiki - which is enough to build and test them. there is enough information on mine and henrik's thread to get your boards built and completed, just don't go about have the 170V at one go first test it at low voltages and then try em for more. be sure that the 15V reaches the boards before the motor voltage,

almost all is covered in the threads here.

Enjoy!

Hi Irfan,

thanks for the reply - last night i soldered the 3 pcbs - nearly ready

only nearly because:
i got 3 kits with different parts

i.e. the resistors with a value of 470R are only in one kit - the 2 other have resistors rated with 300R (R2, R6 - R9 and R43, R44 - matching pieces)
was there anything changed i've not found ?

and another question:
i have no source for a TLP127 (SMT optocoupler)
what is a comparsion type ?

Frank

I think paul will be able to set right the differences in the resistors, do email Paul at uhu.highpower(at)gmail(dot)com

did you not order the TLP127's from Paul?? I personally dont know anything that is similar.

please double check the resistor values - most have got errors due to soldering wrong parts in wrong place.

RGDS
IRfan

Hi,
Paul will have to answer what's supposed to be in the kits and/or if something has changed or if there's been a mistake.
If you indeed have 300ohm resistors instead of 470ohm you can use them for R2, R6 and R7, R43 and R44 without any problems.

I'm 99% sure that you won't have any problems using them for R8 and R9 as well. It will put a bit more strain on whatever is driving the step and direction inputs but I'm pretty sure it'll work just fine. Atleast you shouldn't be able to damage anything by trying.

As for the TLP127 you can get them from DigiKey in single quantites, part number TLP127FCT-ND, the LTV-815S and LTV-355T should work as well.

/Henrik.