New sinusoidal AC servo drive development

[Xerxes]

Going well so far. I have now collected parts and tools to get seriously started with new AC drive software. I already have an external MOSFET power stage board to start experimenting with software within days. Lately I have been designing some schematics and layout for prototype.

[ESjaavik]

ESjaavik, Very interesting soldering technique! But where do everyone find the right flux and the right solder to make it work? I believe all water like fluxes wouldn't be that good. Solders also have HUGE differences, I have noticed that my self. Cheapest just don't work.

Here's some help choosing the right one:
http://www.stannol.de/Englisch/Auswa...20englisch.pdf

I use Stannol 500-3429. It's a couple of years old, and I guess 500-6B is the equivalent current one. It has good cleaning properties and leave little residue. It's considered a no-clean flux. That means you clean it only to make it look better, not because there is any need to.

[JFettig]

Is it possible for these drives to have AC input rather than DC?

Is your drive ready and done? it kind of sounds like you have something running.

Jon

[andy55]

Is it possible for these drives to have AC input rather than DC?

the input to the powerstage of the servordrive is DC.
the surpluscenter servos are 100VAC = ca 140 V peak to peak.
To account for switching losses the DC bus level needs to be a bit higher than 140 V, maybe 180 or so ?

now this DC can and will ofcourse be generated from AC mains.
In countries with 110VAC in the wall you could even rectify the mains directly but that's considered bad practice I guess. (you want to have an isolating transformer in between)

AW

[GisMo]

Hey there,
what I'd like to see..If I am repeating what someone has previously said, I appologize, but what about drivers for retrofit to exising machines/small machining centers..I mean if you can do AC brushless with 1000IPM rapids, thats an Ideal retrofit. I am not too knowledable on the hardware side, nor do I know exactly how the AC motor are controlled? was is sine or square wave?(am I even close?) But why does an AC driver need DC powersupply? The driver generated the AC signal? Is that standard?
Anyways, I'm for the higher power, higher duty cycle drives, high resolution. I like the bigger machines.

Also, I would like to offer my help if needed with the control aspect of the drive, I can help with the modeling and PID control and testing..

GisMo

[andy55]

Hey there,
what I'd like to see..If I am repeating what someone has previously said, I appologize, but what about drivers for retrofit to exising machines/small machining centers..I mean if you can do AC brushless with 1000IPM rapids, thats an Ideal retrofit. I am not too knowledable on the hardware side, nor do I know exactly how the AC motor are controlled? was is sine or square wave?(am I even close?) But why does an AC driver need DC powersupply? The driver generated the AC signal? Is that standard?
Anyways, I'm for the higher power, higher duty cycle drives, high resolution. I like the bigger machines.
Also, I would like to offer my help if needed with the control aspect of the drive, I can help with the modeling and PID control and testing..
GisMo

Hi,

disclaimer: this info is from what I've learned by reading the app. notes on the microchip website. please correct me if I'm wrong...

There are three motor windings, the current in each one is a sinusoidal current, each one is offset in phase from the other by 120 degrees. To get the maxium amount of torque you want the stator windings to create a magnetic field that is oriented 90 degrees compared to the rotor magnetic field. The drive reads the orientation of the rotor from either the hall sensors or the encoders, let's say that the result is an angle alfa. Then it computes the three neccessary winding voltages:
so winding U= A*sin(alfa+90 + 0), winding V= A*sin(alfa+90 +120), winding W = A*sin(alfa+90 + 240).

These sinewaves are created by the H-bridges in the drives, usually driwen with pwm (class D-amplifier). The frequency of the sinewave will change with the rpm of the motor. if the motor is spinning at omega rad/s then alfa above will be alfa = omega * t.

There is usually then a current feedback that checks that the desired winding current was actually achieved. This is because neccessary voltage amplitude A needed to achieve a certain current amplitude will wary with the load on the motor.

For cnc you want your motors spinning at all kinds of different rpm's, so for rpm=0...4500 (max rating for sanyo denki P5), the drive will generate AC winding currents with a variable frequency between 0 Hz and 75 Hz (*)

(*) this may be incorrect. I've assumed that one mechanical revolution of the rotor corresponds to one electrical revolution of the angle alfa. In a motor with multiple poles, like the sanyos, I think one mechanical revolution corresponds to four revolutions of the electrical angle alfa. So we would have a max freqeuncy for the sinewave of 4*75Hz = 300Hz

So the reason why the drives need DC voltage is that the drive then modulates this DC voltage to create three phases of AC with 120 degree phase shifts between the motor windings and with a variable frequency depending on motor speed and a variable amplitude depending on motor load.

hope that made some sense, those that know better, please contribute !

If/when my motors arrive and I get a chance to solder som bits together I will report on my own progress. I'm likely to build a drive based on dsPIC and IRAMS powerstage. Drive would have dir+pwm inputs thus designed to be run with a motioncontroller like unipwm from pico-systems or 5i20 card from mesa electronics (_not_ step/dir signals, I've had enough of them thankyou

AW

[chrisw765]

ok hope your all just hard at work making the drivers cuz i havent seen a post for a bit??? im waiting for them to finish my convertion??? thx pls

[Xerxes]

Yes, I have been developing the drive quite actively few last days. Drives aren't just built in a day. I will keep you posted.

[chrisw765]

ow i know there not built in a day,, there just hadnt been any posts by anyone in a few days just trying to keep the thread and intrest going.... im just exited,, ill shut up know and go take my meds...... j/k

[Xerxes]

Chris, it is good to have at least some activity in discussion.

ESjaavik: I have a question about your smd soldering technique. How small parts you have soldered with this flux method? I guess SOIC packages are easy since they have relatively sparse pitch but how about TQFP or smaller?

And about the drive project: I now have access to a 500 MHz oscilloscope so I can get pretty sure about absence or existence of possible RF interference on the circuit. To design a good drive it is absolutely necessary to check the shape of every signal on the board. However it will take at least couple of weeks to get something to test. I'm rather busy with my studies now, lots of exams coming.

[ESjaavik]

I have a question about your smd soldering technique. How small parts you have soldered with this flux method? I guess SOIC packages are easy since they have relatively sparse pitch but how about TQFP or smaller?

If you look at this picture:
http://www.cnczone.com/gallery/showp...00/ppuser/3485
Those small specks you see in front of the soldering tip, in the upper right corner, are 2 resistor arrays each 4 resistors!
Yes it will do TQFP. But only on a professionally made PCB with tinned pads. If you plan on doing this on a regular basis you should invest in a stereo microscope. The problem is not soldering but inspecting it. I would not leave TQFP to beginners but you'll do it after a few tens of SOJ's as you get the hang of it.

And about the drive project: I now have access to a 500 MHz oscilloscope so I can get pretty sure about absence or existence of possible RF interference on the circuit. To design a good drive it is absolutely necessary to check the shape of every signal on the board. However it will take at least couple of weeks to get something to test. I'm rather busy with my studies now.

It's a lot more difficult to use a 500MHz scope than to solder some TQFP's. A scope can lie you right in the face unless you know how to catch it lying. At 500MHz about everything is a capacitor, inductance, delay line and whatnot. You better make sure you have someone experienced aiding you. If you can get access to a spectrum anlayzer it's a useful tool to snoop for RFI. If you can't find any RFI after careful use of it, you are ready for CE approval. I've saved a lot of expensive test chamber time using it.

[lolailando]

Xerxes,
I emailed you or sent you a private message sometime ago and never heard from you.
Let me know if you didn't get the message.

[chrisw765]

i thought i new alot,"well i do just not in the electronic field",im jealous of your ability to make your ow drives,so my question is what do you do 4 a living and whhat school did u go to and what field or couse was it,im going threw a mid life and am thinking off going back to school and this is field id love to learn about

thx chris

[Xerxes]

If you look at this picture:
http://www.cnczone.com/gallery/showp...00/ppuser/3485
Those small specks you see in front of the soldering tip, in the upper right corner, are 2 resistor arrays each 4 resistors!

Wow! That's really impressive! Do you know good suppliers for that miracle flux in Europe? What do you think about ones sold by Farnell and Elfa?:

http://uk.farnell.com/jsp/endecaSear...comSearch=true

http://www.elfa.se/elfa-bin/thumbs.p...7&7168&2015996

At least this one would be very affordable:
http://www.elfa.se/elfa-bin/dyndok.pl?dok=2008230.htm

It's a lot more difficult to use a 500MHz scope than to solder some TQFP's. A scope can lie you right in the face unless you know how to catch it lying. At 500MHz about everything is a capacitor, inductance, delay line and whatnot. You better make sure you have someone experienced aiding you. If you can get access to a spectrum anlayzer it's a useful tool to snoop for RFI. If you can't find any RFI after careful use of it, you are ready for CE approval. I've saved a lot of expensive test chamber time using it.

At university they have a EMC test chamber mostly idling. I don't know if they would let me use it. At least I did some measurements for course work in there last week.

What kind of "antenna" do you use to catch emissions with spectrum analyzer? Or do you mesure only conductive interference directly from traces?

BTW, do you believe motor drives are something that must follow CE directives? I think motor drives could be classified as 'components' so directives wouldn't apply. I have seen both cases. Some drives claim to be CE compatible but at least one Siemens VFD manual says that it cannot be CE approved because its EMC performance will depend on motors and cables that are connected to it. Most small stepper/servo drives don't mention anything about CE or EMC.

lolailando, I'm afraid I haven't recieved your message. Please post it again in PM.

Chris, I'm full time student at Tampere university of technology. My major subject is in electronics.

[ESjaavik]

@Xerxes: Of those, I would try the X32-10i. Since it is not the same brand, I don't know if it's exactly the same. But the blurb says it has a very good wetting ability. I believe that is what makes the solder wick in between the pins and the pads and shy the non-metallic spaces between them.

I made small antennas and different small air wound coils. Make several of different dimensions, as each will have it's own resonance band and thus enhance certain frequencies. Make them small, as then the frequency they are "tuned to" will be in the GHz bands and above your most interesting bands. Or better: your lab may have probes with known spectral properties. Measuring directly on the traces could easily destroy the instrument. Don't do it.

And I don't believe drives must be CE marked. The machine must (if sold commercially) be CE marked. And components that are CE marked will have an edge in sales as when it have passed once, the integrator knows it have been done before. Sometimes he can also get his machine approved using just the test protocols and showing that he have used the same or better precautions that were used in the drive manufacturers test setup (filtering, shielding etc.). We sell equipment with and without CE marking, and there is no doubt which is the ones chosen when everything else come out equal.

Now I don't expect you to put your P out for chopping by putting a CE mark on a hobby project. But since it seems you have a possibility to test it in a real test chamber, why not do just that. It will add tremendously to your experience. Also if it shows a good test result regarding EMC, especially by being a "silent design", it is a pretty good proof it does not stress its components much. And thus it will be less prone to failure in the field. If as an example your deadtime is too short, you'll come out of the test with your tail between your legs. It will be a quite good radio transmitter. It will also have a high failure rate. By nature a high power chopper will put out a lot, so of course this must be compared to what is unavoidable.

And I wish you well with your studies. I think you will be rewarded with interesting jobs.

[m_j01]

Buy the cheapest AM radio you can find and you will have a very good tool to search for EMI...

[chrisw765]

allright im kick starting to form again """"so whats the latest??

[andy55]

xerxes showed me a pcb design the other day with discrete FETs for the powerstage...

I've toyed with the idea of using the IRMCK201 instead:
http://www.cnczone.com/forums/showth...t=14673&page=3

[Xerxes]

I just placed an order of components that I need for the real prototype. I also ordeder a can of that X32-10i flux. :)

Some progress has been made in programming. Already 1000 lines of code written which includes some setup code and interrupts. This drive will be a total redesign of electronics and a complete rewrite of code compared to my first test circuit (the one in the video).

Still lots of work to do. I want to maximize drive realiability so most of the code will be there just to ensure various things. If drive detects a fault it always would try to fail in the safer direction, i.e initiate controlled shutdown.

I'm planning to make internal program memory checking so it would even detect a faulty flash memory. I also want it to be upgradeable thru serial interface. These might be the most difficult parts since there is no good tutorials for dsPIC on flash programming and bootloaders. Help/links would be appreciated. :-)

Some preliminary specs (subject to change):

-Discrete thru-hole MOSFET design
-Optoisolated step/dir
-Differential encoder input
-Optoisolated RS232 communication for setup & upgrade
-Setup via serial port using general terminal emulation programs
-Possibly analog +/-10V torque command (non isolated).
-Possibly short circuit protection

[lolailando]

this page has a good review of free bootloaders and explanations:
http://www.etc.ugal.ro/cchiculita/so...bootloader.htm

Also it seems like many IDE also provide bootloader code.
hope this helps