KMB1 Conversion

[Bob La Londe]

I am always amazed at how hard it is to find some things. On Friday I did a search for local stores that might have Heat Sink compound. Only Radio Shack showed any in -stock and it was the tiniest tube I think I have ever seen. (I've seen it before so I didn't go get it.) Instead I checked Amazon, and found 3oz tubes from Super Lube for about a buck more than the half eyedropper of compound from Radio Shack. Since my wife has an Amazon Prime account I paid for overnight and had it yesterday. Actually I bought two tubes. I should have enough heat sink compound for a lot of projects now instead of having to buy multiple tubes for a single project.

The DuGong drivers are here, but I am debating ordering the C32 breakout board that just connects up with RJ45 patch cables and using my C23 board on something else.

[Bob La Londe]

Wow! I have to say, "FREAKING WOW!"

Yes, you can quote me on that. I just got the first axis working on the new DuGong 16035 drivers. Once I figured out everything I was doing wrong and got it setup right I did testing upto 450 IPM where it faulted. I was testing on X which is not the heaviest axis, but its certainly not the lightest. At 300 IPM (a good speed for a mill) I had it running at acceleration rates upto 80iss. No, that's not a typo. I imagine I could accelerate it even faster. No faults at that acceleration, but all 2 tons of the machine was shaking with a torture testing code file running. I have a feeling my limitation here is that the servos only turn about 2000 RPM reliably under load at the voltage I am running. That would be about 400 IPM. I decided 300 was fast enough. Its fast enough to shake the machine with rapid reversals at even the 30iss I am leaving it at until I get the other two axis going. If I can get 300 and 30 with such a huge safety margin out of Y I'll be thru the roof.

HOLY FREAKING CRAP! Its like a whole different machine.

Of course I only have one axis going, and I have to wait for a new HEDS module to arrive because I broke one of the pins futzing around with the Z axis. I am pretty excited at the moment though.

[purple_jeep]

Damn that's great news! I am finally at the point where I am tuning my servos on the KMB-1 we have been retrofitting. atm we are using the original servo drivers with a Kflop and Kanalog to do the control and I was wondering what speeds I should be targeting.

Cheers

Chris

[skullworks]

I have been saying since day 1 that people who pull off these servos and replace them with steppers or try to run the OEM servos with Gecko's are severely downgrading the machine and not doing them justice.

If I remember correctly Hurco was using the same servos on the BMC20 & BMC30 machines with 400IPM rapids. (not sure what the ballscrew pitch on the BMC is.)

[Bob La Londe]

If I was running the original 88-89 VDC I could easily see it running 400 IPM reliably. As is it doesn't fault until 450.

Its the acceleration that blows me away though. I was testing at 300 IPM and 80iss with no faults. I ran just long enough rapids back and forth to reach full speed. I like the ServoConfig3 software too. The built in graph for tuning is really cool. I noticed for short moves its hard to get really perfect critical damping, but but for any move about 300 steps the graph looks like it came out of a text book. At 100steps it still looks ok, but not perfect. That's a mixed bag for me because I do a lot of 3D machining where short moves are the norm.

[Bob La Londe]

Well, I just hooked up the Y-Axis, and it went so smooth compared to X and Z. When you actually know what you are doing things sure go easier. Out of curiosity I decided to run a short reciprocating program on Y. 3 inches G00. Back and forth. At 400 IPM and 100iss it still didn't fault. I don't know what the acceleration limit is, but the feed limit was right there where the math said it should be. Now if only that darn replacement encoder head would show up so I can raise the quill back up, LOL. Thank goodness I can lower the table. I set it back to 300 and 30 for now. If the Z tests out similarly I'll probably leave the rapid feed at 300, but I'll certainly crank up the acceleration. Y is the heaviest axis having to carry the table and the saddle, but Z has to fight gravity. We shall see...

[skullworks]

On my KM3 - the Z is on the quill.

Now one issue with ballscrews is that when a Z servo is not powered you can have head/spindle drop when the weight is enough to back drive the ballscrew. My Universal Kwik-Switch 200 quill is effectively the same size as a Bridgeport Knee mill but a bit heavier due to the fact that it does not have the rack gear teeth cuts like a manual BP has and it has the U200 locking system vrs just an empty R8 taper plus the spindle is solid all the way up, not hollow for a draw bar. Even with a heavy tool and tool holder in the spindle and no servo brake the drag of the quill seal and motor are enough that I see effectively no power off spindle drop. Inertia wise - the Z axis is by far the lightest axis. I think you will be able to crank those numbers just fine - It really shows on a G83 cycle.

[Bob La Londe]

My Z is also the quill. I've actually had it ALL apart up on top. In fact I put in a new (used) quill and installed new bearings in it. I also had to hone the sleeve for the new quill to fit properly. That was scary. My quill does not fall with everything installed, but with the servo drive belt off it will move with just a little push. Its the reverse that always concerns me. Its so much different of a mechanism than any other axis with the sleeve as "way" and the roller guided elevator. By hand it definitely goes up a lot hard than it goes down. I do pump a lot of oil into that sleeve every 15 minutes though.

FYI: When I hooked up Y I tested it at 100iss and 400ipm. It didn't fault. When I torture test an axis I run back to back reversing rapids just long enough to get upto full speed. In 3 inches it was hitting over 350 in the middle. Its hard to get exact numbers for sure because the readouts can't keep up. I may upgrade the control computer just to have crisper readouts and DROs. LOL. Its still running on a single processor XP box I bought used. My little Syil X4 Speedmaster is running on a W7 Pro64 box, and it definitely has better display management with Mach 3 running. I bought another Win7 Pro64 box for this machine, but if I can keep it running decently on the XP box the new computer is dogeared for a hybrid machine build I am working on.

Its almost anticlimactic to mention this, but its also huge. When the drive settles (which is really really quick after its tuned) its almost silent. No dithering whine. And yes I still have decent high frequency hearing.

[purple_jeep]

Wow those speeds are crazy! Do you know what your position error is over a move? You mentioned the software had the ability to graph for tuning so maybe it has position error?

This is where I am at with the stock servo drives and Kflop/Kanalog 250ipm. Still a fair bit of work to do on the tuning. Error is ~85 counts with 20,000 counts per inch.



Cheers

Chris

[Bob La Londe]

I still have my error count set fairly high. After I have all three axis working, I'l start cranking down the error limit for faulting and see what it will really do.

Your theoretical following error is a little over .004, but I bet in reality your parts come out better than that.

[Bob La Londe]

Now for a dose of reality. With reasonable following error count (100) set and actual error counts ( < +10/-10) much lower I can run at 200ipm and 30 iss. Currently the drivers are set at 2X and I get 10K pulse per inch. Both of those things threw me, because I thought it was a pulse multiplier setting. ( 2x or 4x) That of course didn't make sense because there was another pulse multiplier setting. Once I figured out it was about whether or not the drivers read just the rising edge or the rising edge and the following edge it explained a lot. Like why previously I was using 20K per inch and now I was having to use 10K per inch. LOL. I think I'll switch them to 4X and retune.

As I expected based on experience Z was problematic. It was much harder to tune and had tuning numbers dramatically different than X&Y. Before dropping the error count the fastest Z would travel was about 230-240 IPM, but oddley it would still accelerate at the same high rate as the other axis did while testing. I wonder if the motor is weaker than the other two or is somehow different and doesn't turn as fast at the same voltage level. It looks identical. I suppose at the very least I should check the brushes. Unfortunately that requires dismounting it. Fortunately its the easiest motor to pull of the three.

[bhurts]

Servos are rated at no load for speed. How much weight they carry has a big effect on how fast they can really spin. Assuming you're z is carrying the knee or just the head and gravity is most likely causing the slower speeds. Also does it have the same screw pitch and diameter as the other axis?

Ben

[Bob La Londe]

Servos are rated at no load for speed. How much weight they carry has a big effect on how fast they can really spin. Assuming you're z is carrying the knee or just the head and gravity is most likely causing the slower speeds. Also does it have the same screw pitch and diameter as the other axis?

Ben

Just a quill elevator on Z. No head and no knee. Yep. Same 5TPI ballscrew on all three. Looks like the same motor. Y carries the most weight.

[m_mcgranahan]

Hi,



I know the thread is pretty old, but it seems like I have some slop in my bed. I saw several mentions about adjusting things, but I can't seem to find pics or more info about how to adjust the ball screws, and all the other things that were suggested to cure slop in the machine.


Also, I was thinking about replacing the bearings in my KMB-1 spindle. I talked to Buddy at Accurite machine and he said it uses a fafnir 206PP with either medium of High preload on the bearings. (Makes them more precise he said)

Does anyone know where to get the high precision 206PP? Or is there an updated part number I could try instead?

Thanks

God bless...
Mark

[Bob La Londe]

Hi,



I know the thread is pretty old, but it seems like I have some slop in my bed. I saw several mentions about adjusting things, but I can't seem to find pics or more info about how to adjust the ball screws, and all the other things that were suggested to cure slop in the machine.


Also, I was thinking about replacing the bearings in my KMB-1 spindle. I talked to Buddy at Accurite machine and he said it uses a fafnir 206PP with either medium of High preload on the bearings. (Makes them more precise he said)

Does anyone know where to get the high precision 206PP? Or is there an updated part number I could try instead?

Thanks

God bless...
Mark

I replaced the bearings in a KM1 quill and put it in my KMB1. (I think it was a KM1) I was told they are both the same. My stock quill was badly scored. If I think of it when I am in the shop I'll look to see what I used. I bought two sets and put one inside the control cabinet for the future. I had to make a pin wrench spanner for the spindle nose bearing retainer. So I wouldn't lose it I made the other end into a spanner for the Kwik 200 spindle collar. I found "new old stock" New Departure bearings I think. Anyway, what I did was pull the spindle apart, see what was in the spindle, and replaced them with like bearings. I only run the machine at the factory 3600 RPM so I wasn't to worried about an upgrade. At the time I had an RF45 conversion running so I wasn't to worried about down time. Only took me a week or so to source bearings and get them in. Carefully lubed the new bearings with Kluber grease. Then I did a slow proper breakin over a half a day. Works decently.

Why do you think you need to replace your bearings? Surface finish? Noise? Rattles around when you grab the collar?

[Bob La Londe]

Wow, Its been over a year and a half since I upgrade to the DuGong drivers. I lost one on the Y axis at a little over a year. I replaced it and everything is good again. Y is the heaviest axis so in case deceleration (back EMF) was killing it I dropped acceleration to 25iss. I do have the "damper" board installed, but if the drive lasts longer I'll assume that was it. I bet between the table, saddle, and vises its moving well over a thousand pounds. The machine gets used about 3 days a week for prepping (square to length) blanks to go on the little 24K mills. It also does a few production jobs of its own where its running continuously for several hours, and about every week to week and a half I'll spend half a day using it as a second operation machine drilling hinge pin holes in hinged lead casting molds. In reality it probably runs (actually doing jobs) on average 20-30 hrs a week. Its the machine I use almost every time when a buddy brings a project over and says, "Bob can you help me with ..."

[Bob La Londe]

Sadly the electro mechanical timer that fires the air solenoid valve that operates the pneumatically driven machine oiler died. Wow, that's quite a sentence. LOL.

I had an electric oiler in my machine building parts cabinet that didn't come with a manual. Its 110V, but just about everything else on the labels is in Chinese. I decided it was time to wire it up and see if I could figure out how to program it. Turned out to be dead simple. Press and hold the set button until it beeps to set the pump time. Press and hold until it beeps again to set the delay time. Press and hold until it beeps one more time to exit program mode. Now all I need to do is figure out where to mount it. I am looking for something a little more convenient than the bottom back of the machine so its a little easier to get to an fill.

I have it set to oil for 15 seconds every f15 minutes. It retains it programming when powered down, so I'll just connected it to the 110V bus in control cabinet. After I have run it for a day or two I'll be able to estimate its oil usage and adjust it so its similar to the original oiler.

P.S. The original oiler still works as does the solenoid valve. I can fire them manually by working the micro switch with my finger.

[purple_jeep]

Sounds like a good solution, I ended up programming the motion controller (Kflop) to fire the solenoid rather than use the mechanical timer.

Sadly the electro mechanical timer that fires the air solenoid valve that operates the pneumatically driven machine oiler died. Wow, that's quite a sentence. LOL.

I had an electric oiler in my machine building parts cabinet that didn't come with a manual. Its 110V, but just about everything else on the labels is in Chinese. I decided it was time to wire it up and see if I could figure out how to program it. Turned out to be dead simple. Press and hold the set button until it beeps to set the pump time. Press and hold until it beeps again to set the delay time. Press and hold until it beeps one more time to exit program mode. Now all I need to do is figure out where to mount it. I am looking for something a little more convenient than the bottom back of the machine so its a little easier to get to an fill.

I have it set to oil for 15 seconds every f15 minutes. It retains it programming when powered down, so I'll just connected it to the 110V bus in control cabinet. After I have run it for a day or two I'll be able to estimate its oil usage and adjust it so its similar to the original oiler.

P.S. The original oiler still works as does the solenoid valve. I can fire them manually by working the micro switch with my finger.

[Bob La Londe]

Well, I wound up mounting the new electric oiler on the back of the machine. I looked for other locations, but the places I found were either less accessible, or would have exposed it to coolant spray and chips. I drilled and tapped some holes into the back of the main column, but level to the ground, not at 90 degrees to the back of the machine. Then I installed two bolts with two nuts. One tightens against the machine, and the other tightens against the back of the oiler pinching it between the nut and the head. This forces the oiler to hang pretty level rather than at an angle. I did mount it much higher than the pneumatic oiler so I could see it more easily and I didn't have to get down on my knees to fill it. For now I am leaving the air actuated operated oiler in place so all I have to do if I find a good timer to use is swap the oil line back.

[Bob La Londe]

The KMB1 is starting to develop some chatter when profile cutting in the Y direction. It doesn't matter which direction, and it doesn't matter if its conventional or climb milling. My first thought is something is bad in the X axis lead screw assembly. Either the ball nut or the bearings.

Then I wondered. Hoped more likely that maybe I just need to adjust the gibs. I have never adjusted them, and to be honest I never considered whether or not the previous owner ever did since I bought the machine from a third party on behalf of his widow.

Now how to adjust the gibs. Hmmm... The X axis seems pretty straight forward. There is a visible screw at each end of the gib. Not sure how to get the feel for it though. On my little machines I can disconnect the lead screw and push the table around by hand, but the table on this thing is significantly heavier than that. Then there is the Y axis. It looks like the gib and adjustment screws must be hidden under the wipers.