Interpolation problem on A2100 with BSD4

[cncwar]

Hello!
We bought Arrow VMC-1500 with A2100 and Kollmorgen BSD4 drivers. Machine works fine except one thing- it has a interpolation problem. Let say we are milling a square - the dimensions are as we want (within 0.015mm). But when it comes to do a circle there is disaster. In X and Y direction it keeps dimension but when i measure under 45 degree (with respect to X+ direction) it shows me -0,3mm of nominal diameter!! Similarly measuring under -45 degree it gives me +0,3 of nominal diameter. Coordinate-measuring machine confirmed this value. This is no mechanical problem as we check it. So i made few more test. The error depends on feed. With F=50 result are very good, with F=1500 result are as mentioned above. The error not depend on diameter - it is always the same.
I decide to play with two parameters: "Kv nominal" which can be find on SERVO -> AXIS DATA tab and "Kv adjust" on AXES -> AXIS RATES tab. The result is now i have got 0.07mm error (within big spread of Feed). Further increasing of Kv makes that axes are vibrating.
Before we bought this machine there was a hard disc exchange (someone put 250GB) because old was broken.
Any idea what else i can change to improve accuracy? Maybe BSD4 adjustment on pots?.

Thank You in advance.

cncwar

Similar Threads:

• A2100 startup problem
• A2100 I/0 Problem
• Problem shutting down A2100
• Acramatic A2100 file transfer problem
• Cincinnati a2100 Drip feeder problem HELP!

[Jim Dawson]

0.3mm is a lot of error, but by adjusting the Kv it is now 0.07mm. It sounds like you have adjusted the motors as much as you can.. You say there is no mechanical problem, but I would check for backlash again.

Before making any pot adjustments on the drives please read chapter 4 of the manual

http://www.kollmorgen.com/uploadedFi...ds4install.pdf

[cncwar]

thank You for fast reply. We did that three time. Using different tool methods and person. Besides adding a backlash we can see significant changes on surface ( in place where move direction are changing).

[Jim Dawson]

I have exactly the same problem on my machine, at -45°, 0.025mm error. I know that I have 0.1mm backlash in my Y axis, but the difference is that I have a very high speed motion controller and 1um linear scales on the table, not encoders on the lead screws or motors so it compensates for most of the backlash automatically, but is still limited by the motor response time.

If you have 0 backlash in X and Y then the only possibility is that the servo motors are not responding fast enough. Maybe adjusting the gain pot(s) on the board would help, or it could be that the servo update loop in the computer is just too slow, but nothing you can do about that. 0.07 mm error on circular interpolation is good on an older machine. If you want round holes, the best way is to bore them.

I'm sorry I can't be of more help.

[cncwar]

I made adjustment using pots on BDS4. First of all i had to compensate a zero speed drift. It wasn't big by the way (~1RPM). Then little adjustment of gain and range. Unfortunately my problem is still not solved. Regulation gave no result at all. For now- have no more ideas what to do.

[meckardt55]

You should have a nominal KV of 2.0 and you should adjust the gain for about 204.5 and then adjust the drive for a gain of 2.0 when moving the axis. The X, Y and Z axis MUST be matched at a gain of 2.0 when moving _+ 0.01.
The problem with scales is that the motor position is fighting the scale position because of lost motion in the thrust bearings. With a machine about 20 years old, it is time to replace the thrust bearings mostly in the X axis we have found.
The reason you you are having problems on a 45 degree angle is because your servo gain is not matched in the X and Y axis, one axis is moving faster than the other, classic gain mismatch.

MIKE CNC Machine Services INC | Cincinnati Milacron MachinesCNC Machine Services

[Jim Dawson]

You should have a nominal KV of 2.0 and you should adjust the gain for about 204.5 and then adjust the drive for a gain of 2.0 when moving the axis. The X, Y and Z axis MUST be matched at a gain of 2.0 when moving _+ 0.01.
The problem with scales is that the motor position is fighting the scale position because of lost motion in the thrust bearings. With a machine about 20 years old, it is time to replace the thrust bearings mostly in the X axis we have found.
The reason you you are having problems on a 45 degree angle is because your servo gain is not matched in the X and Y axis, one axis is moving faster than the other, classic gain mismatch.

MIKE CNC Machine Services INC | Cincinnati Milacron MachinesCNC Machine Services

Mike, thank you for that. It makes a lot of sense! I don't have a Cincinnati, but your advice still applies. Looks like I have a bit of work to do on my machine.

[mike_Kilroy]

What do you mean you do not have the Cincinnati?

Sent from my SM-G900V using Tapatalk

[mike_Kilroy]

Mike is the kv adjustment the same as our speed scale pot on the BDS 4?

Sent from my SM-G900V using Tapatalk

[cncwar]

No. Kv nominal and Kv adjustment are both to change using operator screen (on Axes and Servo screen).
When i got this machine the Kv and Kv adjust was mismatched. I was like Kv =2.2, KV adjust=0.36 for X axis and Kv=2.2, Kv adjust = 0.5 for Y axis
Now i set Kv = 2.5 and Kv adjust = 0.5 on both axes and have got circularity error = -0.02. So i guess it is because of mechanical chain ( bearings ballscrew etc.).

[mike_Kilroy]

No. Kv nominal and Kv adjustment are both to change using operator screen (on Axes and Servo screen).
When i got this machine the Kv and Kv adjust was mismatched. I was like Kv =2.2, KV adjust=0.36 for X axis and Kv=2.2, Kv adjust = 0.5 for Y axis
Now i set Kv = 2.5 and Kv adjust = 0.5 on both axes and have got circularity error = -0.02. So i guess it is because of mechanical chain ( bearings ballscrew etc.).

So it sounds like this is a read only setting you put in to tell the a2100 what position loop gain to run. OK. So it may NOT SET your actual gain - it may be simply a setpoint telling a2100 WHAT GAIN to run. Some controls can ADJUST the gain based on setting this number (within a valid range) while others rely on YOU setting Kv to what you WANT, then YOU adjusting it properly to hit that value.

I suspect your interpolation issues are not mechanical but simply mis-adjustment of our 20 turn "command scale" pot on front of BDS4...

How do you KNOW what Kv (gain) you REALLY are running? There is a screen that shows that realtime while jogging or moving each axis on the a2100... If you cannot find that screen then you could at least watch following error value while moving: it needs to be the same on all axes at same IPM speed... so if you jog at 10 IPM on X, and have Kv set to 2.0, I would assume that means 2IPM/mil so you should have a following error of 0.05". If not same on all axes you want to coordinate, then circles will always be egg shaped as you describe. This actual gain is adjusted by turning the drives command scale pot. So you can see if someone turned the scale pot too far one way or the other, the automatic gain setting in the control, if it has it, will not work.

If you have already tested this and verified you have same actual gain, THEN you should begin looking elsewhere.

Reason I asked Mike to comment on Kv is I try to stay out of the a2100 control; we supplied the Kollmorgen (then allowed private label with Vickers and siemens of our drive) and still today repair them and retrofit them daily, So we know the BDS4 in depth, but let the wizards like Mike meckardt55 answer the control questions! Good luck!

[meckardt55]

If you press DISPLAY on the A2100 front screen then SERVO SET and go to servo 2 (X) or servo 3 (Y) you can see the values and adjust the gain and balance of each servo to match. You should set each servo to 205.4 rpm/volt then adjust the COMMAND SCALE on the BDS4 to get a gain of 2.0 when moving the axis at any speed. This takes 2 people to make this adjustment, one to shout out the gain number and one to adjust the pot on the drive.

MIKE CNC Machine Services INC | Cincinnati Milacron MachinesCNC Machine Services

[maver1ck]

If this is not a Cincinnati Milacron machine that RPM/volt setting probably does not apply as that was based on the Cincinnati's configuration.
We don't have enough info to know what yours should be so you should.contact the manufacturer to determinw what is correct for your machine. Even the gain suggested earlier doesn't necessarily apply to your machine.

Posting for help in the Cincinnati forum will by default get you suggestions that apply to Cincinnati's, which could make your machine worse if you blindly follow them. I still don't remember seeing what type of machine you have in this thread.

[mike_Kilroy]

post #1, first line:

Hello!
We bought Arrow VMC-1500 with A2100 and Kollmorgen BSD4 drivers. Machine works fine except one thing- it has a interpolation problem.

[maver1ck]

I noticed that after I posted, but was hung up on the later post saying he does not have a Cincinnati.

[Jim Dawson]

What do you mean you do not have the Cincinnati?

Sent from my SM-G900V using Tapatalk

Well, 45 years ago I had a V3 Cintimatic NC (paper tape), does that count?

[Jim Dawson]

What do you mean you do not have the Cincinnati?

Sent from my SM-G900V using Tapatalk

Well, 45 years ago I had a V3 Cintimatic NC (paper tape), does that count?

I noticed that after I posted, but was hung up on the later post saying he does not have a Cincinnati.

I believe the OP (cncwar) does have a Cincinnati but I don't. However getting the gain signals synced up makes perfect sense for any DC servo machine. The procedure may be a bit different on my machine with SD drives, but the concept is still the same. I can hang a 4 channel scope on it and watch both axis as I adjust. I should have thought of this before, I do motion control for a living.

[mike_Kilroy]

Well, 45 years ago I had a V3 Cintimatic NC (paper tape), does that count?

My vote is paper does not count, but 45 years experience sure as h**L does! ha!

Same adjustment works on Servo Dynamics drives assuming you have the tach pot adjusted close enough to give you the range on the analog input scaling pot.

[meckardt55]

The info I gave was not for (JIM DAWSON) it was for (CNC WAR) with an A2100 control and the BDS4 is not a DC servo. The info I gave WAS based on experience with many machines of that type. It was meant to coincide with the setup information supplied by the OEM for that machine with that control using the BDS4 drives. I can send you the setup info if you need it. Jim, you are correct, we used a scope to setup the old DC drive many years ago but now with the A2100 control showing the gain, following error and many other servo related info on the servo setup screen we don't need to use a scope to setup a newer AC drive with the newer controls.

MIKE CNC Machine Services INC | Cincinnati Milacron MachinesCNC Machine Services

[mike_Kilroy]

If you press DISPLAY on the A2100 front screen then SERVO SET and go to servo 2 (X) or servo 3 (Y) you can see the values and adjust the gain and balance of each servo to match. You should set each servo to 205.4 rpm/volt then adjust the COMMAND SCALE on the BDS4 to get a gain of 2.0 when moving the axis at any speed. This takes 2 people to make this adjustment, one to shout out the gain number and one to adjust the pot on the drive.

MIKE CNC Machine Services INC | Cincinnati Milacron MachinesCNC Machine Services

PERFECT!! Thanks for sending that to OP Mike!!