Proto Trak MX3 Retrofit - Which Boards Mesa?

I am starting to look seriously at converting my Proto Trak MX3 to linuxCNC. I am very familiar with Linux CNC but from the standpoint of using with Step/Direct Stepper systems. This machine is using servos with encoders.

Unfortunately, I cannot find much information on the servo amplifiers. An older thread on this board suggested the amplifier is using differential PWM.

In any case, I am comfortable using meters and Oscilloscopes but am not 100% sure where to start. I am linking to some pictures of the amplifier board, motor and encoder(well a jumbled shot of the encoder housing).

I am hopeful that somebody may have already done one of these SWI Proto Trak machines that can help out. If not I am at least hopeful that somebody can share some of the necessary logic for determining connections and appropriate hardware to look at for interfacing with Linux CNC.

Servo Amp Overview


Servo Amp Connector Closeup


It appears as though the connector on the amplifier has
110 AC Input
5V Input
PWM
NPWM?

And several other pins that I cannot read the labels on because the connector covers the silkscreen.

Motor Encoder Housing


It looks as though I will need to take the time to unplug some wires from the encoder in order to better understand what is going on.

Motor NamePlate


I look forward to any input on the path of least resistance to getting this machine up on Linux CNC

I have researched what is going on at the servo amp a little more thoroughly and labeled the connections on the servo amp with my best terminology to describe what I am observing at each pin. (oops just noticed the typo on Neutral)




Interestingly, the input to the servo amp appears to not be +/-10v rather it appears to be controlled by 4 channels of PWM signal. This is definitely different from the encoder signal which I have identified as well and will document further shortly.

If anyone is familiar with 4 PWM lines as input to the servo amp, I will be interested to learn more about this method as well as what hardware could be employed to control these sevos with Linux CNC.

Thanks!

Its probably differential PWM in locked antiphase mode (50% = 0 drive),
and I think these drives are disabled by shutting off the PWM

Probably drivable with a 5I25 and a 7I85S with maybe a 7I84 added for GPIO

But if you have a working system, a scope trace of the PWM signals would be valuable

I am assuming (but do not know) that the PWM signals are isolated from power so be careful
when poking around/measuring. First measuring the ACV of a PWM lead from ground
with a voltmeter on the 300VAC or so scale might not be a bad idea

Thanks for the response! I have done some research on locked antiphase mode PWM and it certainly seems possible that this is the case. I had done some basic scoping around with a USB Scope plugged into my laptop running on battery which at least makes me feel as though I am isolating the scope from unexpected interactions with mains wiring.

I was seeing a 5v PWM signal on the servo amp input lines. The voltage was in reference to chassis ground which I am not sure is the best in this case as I was finding the signals to be noisier than expected but clearly a 5v square wave in any case.

Another interesting observation that I made was that I was getting a signal on the Xaxis amplifier input even when jogging the Y Axis. One possibility that I can think is that when jogging the Y axis the X is being sent the 50% PWM=0 Drive as suggested above. This would also be consistent with the drives being disabled (No PWM signal) when not jogging. I was also thinking that I should try referencing the +5V with the scope as opposed to referencing the chassis.

In any case, the next chance I get, I will capture the PWM signals with the scope and try to verify exactly what is going on with the signal.

Thanks again for your input!

PCW wrote:

Its probably differential PWM in locked antiphase mode (50% = 0 drive),
and I think these drives are disabled by shutting off the PWM

Probably drivable with a 5I25 and a 7I85S with maybe a 7I84 added for GPIO

But if you have a working system, a scope trace of the PWM signals would be valuable

So I was able to pull some traces off of the PWM leads on the Mill over the last few days (See below). The noise still leads me to believe that using the chassis ground as reference is a mistake but will have to go back and poke around to get a cleaner signal. In any case, I have 4 screenshots below. 2 jogging the machine at 10 inches per minute and 2 jogging the machine at 70 inches per minute. At each speed I have a screen shot for each direction.


^^^^10 Inches Per Minute in Positive Direction^^^^


^^^^ 10 Inches per Minute in Negative Direction


^^^^ 70 Inches per Minute in Positive Direction ^^^^


^^^^ 70 Inches per Minute in Negatove Direction ^^^^

These signals seem consistent (From what I have learned from a small amount of research) with the suggested differential PWM in locked antiphase mode

• Does this appear to be correct?
• Also do these plots show enough to conclude that the 5I25 and a 7I85S Combination will be a good choice to drive these amps?
• If the 7i85 does seem to be a good choice is the signal output to drive the amplifiers something that is programed into the firmware of the 7i85 or is it something that is configurable through LinuxCNC? It probably doesn't matter which but I am curious..

Yeah locked-antiphase differential PWM as suspected (at about 15.5 KHz)

You can see why they used differential since the signals are quite
noisy when measured singled ended

Should be easily drivable with 5i25/7I85S with the
5i25 programmed with a 7i85spx2 config

Have you had any success with the Mesa boards techrat?

I have an older version of the servo drives but they have the same pinout, and the driver board of the M3 pendant stopped working so I'm looking for alternatives.

Hi There JR,

I have not taken the plunge into purchasing the Mesa Boards yet. From my research it does appear to be the way to go. I just need to dive in and get it done. I know it has been a while since your post. Have you taken any steps to get your machine up and running under alternative control? I will remember to come back and post my results once I dive into this project.

Jeff

No, I have not yet purchased any boards.
I am waiting on a response from Mesa regarding a few questions I have, as I would like to use an ethernet based board instead of a PCI or parallel port based board.

Thanks for putting up the signals for the servo board by the way. The only thing SWI could tell me is that they used a 2.5VDC PWM signal.

If the drives take differential PWM, the standard differential drivers on a 7I85S should work OK
these will give between 3-4 V differential into a 100 Ohm termination