Please Help: DC brushed Servo Motor Vibration at very low speeds

Thread Starter

jeffbrog

Joined Mar 9, 2014
16
I hope someone can shed some light on a problem I am having with a DC brushed Servo motor in a factory made laser engraver.
The machine is no longer supported by the US manufacturer because it is 19 years old. Talking to them is useless. Same problem with the motor manufacturer, even though they appear to make the same motors for OEM configurations.
I am not an electronics technician, but i have a very good understanding of general electronics, electric and repairs. What I try to describe may not be the correct terminology, but hopefully I'll get the idea across! Here is the problem:
The laser engraver uses (2) 48 volt brushed DC servo motors for the X and Y axis control. The motors do not have any gear reduction or reduction in the way of different sized pulleys on the belts that drive the carriage. The Y axis is generally rotating fairly slow as it moves. Probably in the range of 30-100 rpm for moving to the start/home positions, etc. When the laser is in "vector cutting" mode, as in a CNC type use, it rotates very slowly when the speed is set to the minimum ranges of the driver (in the PC or machine control panel). The machine is supposed to be able to be controlled in speed setting from 1-100% for various engraving/cutting and applications.
When the speed is set to most settings below 10% the Y axis motor will set up a severe vibration that affects the speed and causes "jerky/stalling/erratic" movement. It seems to be fine above that speed. It does not always start vibrating instantly at a given low speed but can start after 4-6 rpms.
One old bulletin from the manufacturer states that "the Y motor is set to low gain on start-up and will not vibrate, but when it is engraving the gain is set high, and the machine will vibrate if the short belt is loose".
They are referring to the timing belt from the motor to the axle shaft that drives the Y axis. I have tightened to their recommendations , and tried a variety of different belt settings, tighter and more loose , but no joy. I have removed the motor from the mount and held it in my hand with no load and get the same vibration, so I don't think it is "machine" related bearings, loading, etc.
Since the motor manufacturer does not offer any parts or motors, I have disassembled the motor to check it over. I thought the ball bearings were too loose, so replaced those with new R4A-ZZ-C3 ball bearings, I removed any end play on the shaft, checked the encoder for dirt, dust, etc. (none), checked the brushes for wear, brushes looked almost new, maybe not quite seated in to the armature/rotor yet, as they were worn only on half of the rectangle surface, they may a have been replaced shortly before I bought the unit. The bars on the armature/rotor are clean, smooth, not worn, the gaps between are void of build-up, and in general look very good.
At this point I'm considering all options.
I probably can't change the frequency or pulses sent by the proprietary controller board.
I can't find a replacement motor yet on e-bay, but don't necessarily think it's the motor at this time.
Could it be the frequency that is being sent to the motor is setting up a resonant frequency within the motor?
Should I remove the motor and run it with a dc power supply to seat the brushes better (for full contact)?
Would a resistor in series to the motor or capacitor across the motor do any good? I don't want to blow the control board.
I have considered changing the ratio on the motor drive to a 2:1 ratio and changing the encoder wheel from 512 cpr to 256 cpr so the motor will rotate at twice the speed and maybe avoid the problem.
Any ideas or comments are appreciated.
Sorry so long!
Thanks,
Jeff
 

MaxHeadRoom

Joined Jul 18, 2013
28,619
You could remove the motor and test with a variable DC supply of some kind, but I am guessing it is either in the Drive, tuning or even could be the encoder feedback.
That is quite low encoder resolution by present day standards, and running it with a 256 count will upset your move distance calibration unless you have access to the parameters.
Tuning would be my first choice, are the drives identical? as you may be able to swap them temporarily in some cases.
Max.
 

Thread Starter

jeffbrog

Joined Mar 9, 2014
16
You could remove the motor and test with a variable DC supply of some kind, but I am guessing it is either in the Drive, tuning or even could be the encoder feedback.
That is quite low encoder resolution by present day standards, and running it with a 256 count will upset your move distance calibration unless you have access to the parameters.
Tuning would be my first choice, are the drives identical? as you may be able to swap them temporarily in some cases.
Max.
Hi Max,
Both motor drivers are built into the motherboard that runs the laser, the X axis is a smaller motor, about half the size of the Y motor. So they probably have a different current rating, etc. but not sure.
After some additional thought the encoder/ gear ratio change, idea probably wouldn't work, without access to the parameter, like you said.
I'll shoot a picture of the motherboard and attach it.
I didn't remember seeing any adjustable pots on it.
The machine has settings of 200.300, 400, 600, and 1200 dpi. I think the 1200 is not entirely accurate and is more likely closer to 1000dpi.
Jeff
 

Thread Starter

jeffbrog

Joined Mar 9, 2014
16
On second (and third) thought, if I did replace the the encoder wheel with the 256 cpr, would the motor spin 2 rpms and only think it was making 1 rpm?
I was thinking since the driver is counting the pulses from the outer divisions on the encoder it would be looking for and running till is sees 512.??
But I think the encoders and driver keep track of the number of revolutions by the inner divisions and "holes" in the encoder wheel and this would screw things up.
I haven't played with servo motors and encoders at all, so I'm still learning!
Jeff
 

MaxHeadRoom

Joined Jul 18, 2013
28,619
That is what I meant by move distance, when a machine is configured, it has to know the feedback scaling for a input increment, IOW when a 1 meter move is commanded, then it knows from the scaling parameter and by monitoring the encoder how many pulse it has to see for 1 meter, to put it simply.
Incidentally most systems take advantage of using the four edges of the encoder pulses in order to multiply the basic encoder count x4, so if you system does this, then the controller sees 2048p/rev.
Most encoders now use 2048 as a low end and end up with 8092p/rev a minimum.
Max.
 

Thread Starter

jeffbrog

Joined Mar 9, 2014
16
I think I have found the problem.
I removed the motherboard that has the motor drivers are built into it.
On the back of the board it was obvious that a small surface mounted component was blown.
C-30 is the board's label for it. It is tanish/brown in color, connected across pin 6 and pin 8 of the LMD 18201T -H Bridge driver.
Now all I need to know is what value the component is and if it's a capacitor or ?
I can only guess as to why it blew in the first place, and that guess would be that the motor brushes look very new, and not yet seated in. Maybe the old brushes were worn out and caused a problem. Maybe it blew from old age deterioration. Any other ideas. See the picture for a look.
There is another LMD 18201T driver for the smaller X axis motor with a good component C-29 in the same position still on the board, but markings are not real legible.
I am including a photo, if someone can help.
Pin 6 appears to be Vs power supply and pin 8 appears to be signal ground.
Motor is 48 vdc.
Please let me know if you need more info.
What do you thin Max?
Thanks,
Jeff
 

Attachments

cmartinez

Joined Jan 17, 2007
8,220
I think I have found the problem.
I removed the motherboard that has the motor drivers are built into it.
On the back of the board it was obvious that a small surface mounted component was blown.
C-30 is the board's label for it. It is tanish/brown in color, connected across pin 6 and pin 8 of the LMD 18201T -H Bridge driver.
Now all I need to know is what value the component is and if it's a capacitor or ?
I can only guess as to why it blew in the first place, and that guess would be that the motor brushes look very new, and not yet seated in. Maybe the old brushes were worn out and caused a problem. Maybe it blew from old age deterioration. Any other ideas. See the picture for a look.
There is another LMD 18201T driver for the smaller X axis motor with a good component C-29 in the same position still on the board, but markings are not real legible.
I am including a photo, if someone can help.
Pin 6 appears to be Vs power supply and pin 8 appears to be signal ground.
Motor is 48 vdc.
Please let me know if you need more info.
What do you thin Max?
Thanks,
Jeff
That definitely looks like a capacitor... as a wild guess, I'm noticing there's symmetry in the board... perhaps the value of that cap is the same as C29? You could dismount C29, measure it, and replace the blown cap with a similar one. Also, make sure that no other traces or components have been affected... especially chips, transistors and diodes.
 

Glenn Holland

Joined Dec 26, 2014
703
Just glancing through a lot of details here.

However, is the servo motor actually a stepper motor? Seems if there's no gear reduction and it's used in a motion sensitive application, it might actually be a stepper.
 

Thread Starter

jeffbrog

Joined Mar 9, 2014
16
Hello all, Thanks for all the input given.
I got to the point where I didn't have time to look into this any further, but now I just got back into it a little bit more.
The motor seems to be fine, so I have ruled that out for now.
I believe the problem exists on the control board, since that small surface mounted capacitor ("AA5" printed on the cap) was blown on the backside.
It is/was mounted across two pins (6 and 8) of a LMD 18201T H-Bridge. Picture above is of the backside. There are two of these H bridges, one for the x and z axis.
I replaced the blown cap with a .1uf cap to see if that would fix the low speed pausing/slowing problem, but no go, still the same.
The cap I used is not a surface mounted cap, but one with leads on it, so I could solder it to the pins of the H bridge.
I'm wondering if there could be a problem in the LMD18201T bridge internally? But why it would run great at higher speeds but not the slowest of speeds though. The speeds I am talking about are very slow, in the neighborhood of less than 5-10 rpms. The motor kind of pauses or stalls or slows for a second or two and then resumes, and can do this several times a minute or so.
I have downloaded the data sheet for the H bridge, and looked at all the info and diagrams, but there are a few things about the way it is connected in my circuit board that don't make sense compared to the data sheet. But since this came from commercial USA $15k-20k (circa1993) laser engraver, I have to think they knew what they were doing! Also, the other motor works just fine with their circuitry!
I've been thinking about replacing the H bridge, but hate to do that if it is good, and the problem is elsewhere.
Anybody here familiar with the H bridge? I can elaborate more about the differences to the data sheet.
Thanks again.
 

Alec_t

Joined Sep 17, 2013
14,280
To run at slow speed (<30rpm) the motor presumably is being pulsed with DC? If the pulse timing is done by discrete components rather than by a microcontroller then a fat capacitor might be involved and could have gone bad, leading to oscillation?
 

Thread Starter

jeffbrog

Joined Mar 9, 2014
16
To run at slow speed (<30rpm) the motor presumably is being pulsed with DC? If the pulse timing is done by discrete components rather than by a microcontroller then a fat capacitor might be involved and could have gone bad, leading to oscillation?
Yes this is a 48v DC servo type motor with disc encoder on the backside. There are only 2 large capacitors on the whole board, right between the 2 heat sinks for the H bridges. I just checked both of them yesterday with a cap checker, both are in spec. at 470uf. I'm attaching a picture of the front of the board. On the back of each heat sink for the bridges is another component one has 2 legs the other 5 (visible in photo). I can get the numbers off them or any other component if it helps.
photo.JPG
 

Alec_t

Joined Sep 17, 2013
14,280
Since there's an umpteen-pin micro evident on the board I guess that looks after the motor pulse timing. The caps then would be just power-supply reservoirs.
 

Johann

Joined Nov 27, 2006
190
It is not functioning properly anyway, and the board damage is on the H-bridge pin(s), so perhaps the pcb tracks on the component side of the board got damaged or the H-bridge itself. Perhaps it is worth a try to replace the bridge and at the same time inspect the tracks of the pcb very well. It seems to be a power fault on the power component (bridge).
 

RichardO

Joined May 4, 2013
2,270
Yes this is a 48v DC servo type motor with disc encoder on the backside.
Just a thought... Maybe you have a bad encoder or connection to it.

I would expect that missing pulses would not cause the problem you are seeing but the reason for faults can often be deceiving until they are fixed.
 

MaxHeadRoom

Joined Jul 18, 2013
28,619
The whole system which as you say is proprietary appears to be the controller, HMI and servo drives all in one, it may come to the crunch of doing a retrofit.
There are quite a few options out there from Camsoft to a linux based system and also Dynomotion etc.
Max.
 

Thread Starter

jeffbrog

Joined Mar 9, 2014
16
Thanks for the input, ideas, and suggestions.
I'm posting a pic. of the rear of the board, shows my Frankenstein cap that I added to replace the blown SM cap. Maybe this is not the right type of cap in the way it should work?
I haven't seen any burned traces, from what is visible. I may take off the heat sinks and check under them.
One of the many things I find ODD about this board is the large traces that come from the output pins of the H-bridges to the motor connectors for the DC current (which I would expect), but only minuscule traces for the input pins (and all other traces on the board). I would think the needed power IN for the bridges would be comparable to the power out to the motors and thus require larger traces elsewhere.
Before I replace the H-bridge, Can someone tell me anymore about what cap I might try in case the one I've used (.1uf) is insufficient?
Should I try another value?
It is connected across pin 6 and pin 8 of the LMD 18201T -H Bridge. The data sheet does not discuss any cap across those pins (it does for other pins). The other bridge has the cap across pins 4 and 6!!! It is facing the other direction on the board (180°).
Epilog circuit board rear.JPG
 

RichardO

Joined May 4, 2013
2,270
I can't tell for sure but it may be a 4-layer PCB. That would explain why you can't see some of the large traces to the motor driver.
 
Top