[amishx64]

I am testing the electronics for my new machine and am having quite an issue with my stepper motor and/or driver and/or parallel port. I can't figure out what the issue is.

I have a large nema34 motor powered at 50V 7A by a Centent CN0142 Anti-Resonance Microstep Drive via 14awg wire hooked directly to a parallel port via shielded wire.

Mach 3 is set at 100kHz and axis x is at 20mm/s^2 accel and ~2500mm/s velocity. See attached picture.

The motor runs great up to about 1500mm/s but starts having issues at 1600. The motor is fine after a brief time in that range when it tops out at 3000mm/s. Please see the video:

[ame=http://www.youtube.com/watch?v=iuAtGrrvM9I]Strange Stepper Motor Mid-range Dropout - YouTube[/ame]

[Eurisko]

I think that your 'steps per' in the X axis movement profile is set incorrectly.
Note that the 'steps per' is not steps per revolution of the motor, but the number of steps needed to move the X axis 1 mm.

'steps per' = (#steps per revolution of screw) / (screw pitch in mm)

For example, if you have a 5mm pitch screw, the 'steps per' would be 400.

[amishx64]

I think that your 'steps per' in the X axis movement profile is set incorrectly.
Note that the 'steps per' is not steps per revolution of the motor, but the number of steps needed to move the X axis 1 mm.

'steps per' = (#steps per revolution of screw) / (screw pitch in mm)

For example, if you have a 5mm pitch screw, the 'steps per' would be 400.

It is not connected to any screw drive or linear system as of yet. The 'steps per' does not make a difference in this case. Thanks though.

[WoodSpinner]

It is not connected to any screw drive or linear system as of yet. The 'steps per' does not make a difference in this case. Thanks though.

absolutely steps per matters, you have it set at 2000 steps per/with a max velocity of 2500, so thats 5000000 steps per minute going out, you didnt mention what your micro stepping rate was, but you need to consider how fast you trying to spin that stepper, you most likely over reving it given your current settings.without microstepping thats 200 steps per rotation on most steppers, so divide 5000000, by 200 your trying to get 25000 rpm, get real.

I'll shut up now

John

[amishx64]

absolutely steps per matters, you have it set at 2000 steps per/with a max velocity of 2500, so thats 5000000 steps per minute going out, you didnt mention what your micro stepping rate was, but you need to consider how fast you trying to spin that stepper, you most likely over reving it given your current settings.without microstepping thats 200 steps per rotation on most steppers, so divide 5000000, by 200 your trying to get 25000 rpm, get real.
John

John, I understand your concerns about me trying to run the motor too fast. I don't necessarily think this is the problem. My motor is indeed 200 steps/rev and the controller is doing 10 microsteps.

Remember from the first post that the motor starts failing at 1600. It runs fine at 2500.

The fact that the motor runs fine at speeds 0 - 1600 and 2000 - 3000 but not that 1600 - 2000 middle band. This is what I am concerned with. It makes no sense that the motor can run fine at higher speeds and stall at lower speeds.

I'll shut up now

John

Please don't. If you have corrections, I wish to learn. Thanks

[WoodSpinner]

John, I understand your concerns about me trying to run the motor too fast. I don't necessarily think this is the problem. My motor is indeed 200 steps/rev and the controller is doing 10 microsteps.

Remember from the first post that the motor starts failing at 1600. It runs fine at 2500.

The fact that the motor runs fine at speeds 0 - 1600 and 2000 - 3000 but not that 1600 - 2000 middle band. This is what I am concerned with. It makes no sense that the motor can run fine at higher speeds and stall at lower speeds.



Please don't. If you have corrections, I wish to learn. Thanks

Sounds like what I've heard refered to as mid band resonance, usually the comments suggest using a damper to smooth it out, I've never had that particular problem, hopefully someone else will speak out, there are threads dedicated to this problem here, maybe with a little searching..
John

[amishx64]

Sounds like what I've heard refered to as mid band resonance, usually the comments suggest using a damper to smooth it out, I've never had that particular problem, hopefully someone else will speak out, there are threads dedicated to this problem here, maybe with a little searching..
John

This seems to be exactly the problem I am having. Thank you!

Now to solve it...

[Eurisko]

amishx64,

I'm curious how you are calculating linear speed from rotational speed with no linear reference. Also, unless you're using a tachometer to measure the rpms, you can't be sure that it isn't losing steps at the higher rpms.

With the settings that you're using, you're assuming a linear movement of 1 mm per revolution of the motor. So, linear speed=rotational speed.

2500 rpms won't produce much torque, and the numbers you have are meaningless if there is no load on the motor.

[WoodSpinner]

John,

I'm curious how you are calculating linear speed from rotational speed with no linear reference. Also, unless you're using a tachometer to measure the rpms, you can't be sure that it isn't losing steps at the higher rpms.

With the settings that you're using, you're assuming a linear movement of 1 mm per revolution of the motor. So, linear speed=rotational speed.

2500 rpms won't produce much torque, and the numbers you have are meaningless if there is no load on the motor.

actually, I took his settings which refer to linear motion, and converted it to rotational speed
John

[Eurisko]

actually, I took his settings which refer to linear motion, and converted it to rotational speed
John

Sorry, John.

My question was directed to amishx64, I don't know why I used your name...

[amishx64]

Eurisko,

The numbers are really completely arbitrary. I don't have a system currently using these units. I was simply using the Mach defaults and using them to provide a (fake) velocity range of where the motor would and would not work.

I don't know what rotational speed the motor was moving at. I simply wanted to know why there was a stall / droupout at a certain speed and not speeds above or below the problematic speed range.

I am sorry for any confusion this has caused all of you. :stickpoke

[Mariss Freimanis]

I listened to your video. The drive is having problems reading step pulses at high rates of acceleration. The drive also has problems with step pulses that have a lot of phase jitter such as what Mach3 puts out. Finally, the CN0142 is a 3.5A drive and not a 7A drive.

I know because I'm the guy who designed it in the early '80s.

Mariss

[amishx64]

I listened to your video. The drive is having problems reading step pulses at high rates of acceleration. The drive also has problems with step pulses that have a lot of phase jitter such as what Mach3 puts out. Finally, the CN0142 is a 3.5A drive and not a 7A drive.

I know because I'm the guy who designed it in the early '80s.

Mariss

YOU helped design it?? Now that is cool!

You are right. The chart says 7A series 3.5A parallel. My mistake. Edit: I can't read. Fixed

Any hints on what to do as far as phase jitter or acceleration is concerned? Even when I have the accel set to the lowest possible point it still has the same issue. Using higher accel rates enables me to get through that resonance band but it is my guess that steps are lost and I know there is not sufficient torque in that range.

Am I out of luck with the driver? I guess I can run it at slower speeds...

[Mariss Freimanis]

I didn't "help" design it, I was the designer so I'm to blame.:-)

You don't have a resonance problem, what you have is a poor design problem in the step pulse phase modulator circuit of the drive or a corrosion problem.

1) The feedback circuit in the phase modulator couldn't adapt quickly enough to step pulses that have a marked difference in period from pulse to pulse. The consequence is an overly abrupt change in period causes that step to be ignored, (as in missed).
This was a known issue and the CN0142/0143 and it carried a +/- 20% specification on how much the step pulse period could vary from the previous one.

When you are accelerating or decelerating, the period of each subsequent step will necessarily be different than the preceding one. Mach3 doesn't do you any favors here either because the way it generates step pulses results in period to period time variations (phase jitter). Combine the two (rapid acceleration and Mach3) and you get the result you are seeing.

2) As you notice, the CN0142 is encapsulated in epoxy. When it was manufactured, it was washed in an aqueous alcohol solution for flux removal and that solution found its way into every microscopic nook and cranny of the drive. What moisture was left after drying was forever sealed in place by this epoxy.

Over a long period of time, (months or years) corrosive action from this trapped moisture could "poison" components inside the drive. By far the components most prone to this poisoning were hermetically sealed glass-metal packaged diodes (DO-7 package). This package was particularly sensitive because capillary action trapped moisture at the glass to metal interface and only overnight baking would drive it out.

My partners insisted on encapsulation for anti-piracy reasons. I hated it because it made the drive unrepairable and impossible to trouble-shoot. On top of that, about 5% of the CN0142/0143 drives developed this delayed corrosion problem.

On the other hand, the CN0142/0143 was the great-grandfather of the G201 drive which I designed in 1999, after I sold my interest in Centent Co. It of course doesn't have any of the above mentioned issues.

Mariss

[amishx64]

Mariss,

THE designer. Very awesome. I understand you do Gecko designs now too. I am very impressed.

Thank you for explaining phase jitter and the conflicts with Mach3. Very informative. I appreciate that now because I was going to look into getting a damper for the stepper motor. (flame2) It is a shame that Mach3 has messy signaling. Moving to EMC is an option but not one I want to take. Besides, I don't know if it would help any.

It is funny you mention the epoxy. I was wondering why in the world that was put into place. The only thing I could think of is dirt and grime. One of the screw terminals on my drive stripped out (motor ground) so I can no longer tighten the cable into place on that one terminal. If there was no epoxy, I would've just soldered directly to the board or gotten another terminal. I solved this issue by soldering the wire to the inside of the screw terminal. Not the cleanest job, but it was the best I could do.

Do you know what a common symptom of the alcohol corrosion would be? Just so I know what to watch for... I have 4 of these drives.

I looked at getting Gecko drives but the cost of just one exceeded the cost of me purchasing all 4 of the CN0142's (used via $eBay). I think I will see what the performance of the CN0142 is like before I consider a Gecko replacement. I think I should note at this point that cost is an issue for me considering that I am in college.

Any word on the G215 release date? (I looked at some of your threads.)

[Eurisko]

I listened to your video. The drive is having problems reading step pulses at high rates of acceleration. The drive also has problems with step pulses that have a lot of phase jitter such as what Mach3 puts out. Finally, the CN0142 is a 3.5A drive and not a 7A drive.

I know because I'm the guy who designed it in the early '80s.

Mariss

Mariss,

That is exactly what I was going to say. Geez Mariss, if you keep jumping in with real-world experience and advice that actually works, you're going to spoil it for the rest of us...

[Mariss Freimanis]

Eurisko,

I hope I'm not doing that. I'm grateful for all the people who do reply to questions regarding our drives, past or present. It is impossible for me to reply to every question so I'm thankful for the good advice and help people like you selflessly offer.

Mariss

[Mariss Freimanis]

amishx64,

The G215 hardware is finished and tested. The Revision 0 firmware for its FPGA is finished and tested as well. What I'm working on right now is the embedded PIC assembly language firmware. The PIC will allow the G215 to use its internal non-volatile memory to store from 128 to 1,024 axis coordinates (depending on memory size used). As usual, it will be done when it's done.

Mariss

[amishx64]

Mariss,

Good to know. Can you say with confidence that a gecko drive would fix the issues that I am experiencing? Thanks.

[TarHeelTom]

Mariss,

Good to know. Can you say with confidence that a gecko drive would fix the issues that I am experiencing? Thanks.

I can't answer for Mariss, but I personally consider the Gecko to be the gold standard.

I looked at getting Gecko drives but the cost of just one exceeded the cost of me purchasing all 4 of the CN0142's (used via $eBay). I think I will see what the performance of the CN0142 is like before I consider a Gecko replacement. I think I should note at this point that cost is an issue for me considering that I am in college.

The big difference here is that the Gecko just WORKS. And if it doesn't, Mariss fixes it and sends it back quickly. With the Gecko, you get your money's worth. With the CN0142, you don't get your money's worth, unless you got it for free.

Tom