Bridgeport/Romi EZPath S Retrofit Planning

More
23 Dec 2021 00:13 #229780 by TurboStreetCar
Good evening everyone, New to the forum and ive held off on making this post so i could do a fair bit of research into what i would need to make this happen. I BELIEVE i have it all mapped out, but before i go spending money on stuff id just like to double check my thought process.

Backround: The EZPath S i have is a 2 axis toolroom CNC lathe with a 2 speed gearbox and is CSS capable. Its currently, actually, fully functional. The reason im considering the retrofit is because the EZPath control is very dated and is missing some VERY basic functionality. The spindle control is a manual switch, and it can only feed in IPR. It also has zero M code user outputs, so interfacing a collet open/close solenoid, bar feeder, or parts catcher is nearly impossible to automate.

Electronics: Currently it is equipped with 200V Yaskawa SGDB-05ADG Servo drives and matching Servos with 8192 PPR Encoders (However manual claims 3000PPR, Yaskawa Model# shows 8196PPR). It has a 7.5 HP Motor with a Yaskawa CIMR-P5U25P5-75CV VFD.  The servo drives are position mode, and accept direction/step pulse inputs. The VFD accepts analog 0-10V speed control. 

The Plan: I believe from my research, i would use a Mesa Ethernet 7i95 board with a Mesa PWM->Analog converter, and modify the 7i95 firmware to convert one of the step/dir outputs to PWM for spindle control. The 7i95 has 6 step/dir outputs and 6 encoder inputs, which is plenty for my setup. I BELIEVE there is enough I/O on that board for my purposes, however adding an additional I/O board seems simple, so im not so worried about that. 

Ive already sourced a Dell Optiplex 3040M (i3-6500T) which ive installed the newest Debain Buster ISO on and ran the latency test, results seem to be in the 70K range for the servo thread, which from my research, seems to be plenty good enough for use with hardware step generators. 

Do i seem to be headed in the right direction with my hardware selection and how ive wrapped my head around the process?

Appreciate any help on this endeavor. 

Thanks so much,
Dan

Please Log in or Create an account to join the conversation.

More
24 Dec 2021 00:08 #229838 by andypugh
It might be simpler to use a 7i76 and 5i25/6i25 or a 7i76E for this. That gives you all the stepgens and an analogue spindle control built in (with spindle encoder support)

The 7i76E is cheaper than the 7i95 too.

Please Log in or Create an account to join the conversation.

More
24 Dec 2021 00:34 #229842 by TurboStreetCar
I had considered it, But the 7i95 has 6 encoder inputs.

So i could use 3 of the encoder inputs for the spindle, X , and Z.

The lathe also has manual handwheels, that are basicly just manual encoders. They command a change in position through the control, no mechanical connection to the axis.

I could use two more of the encoder inputs for those handwheels, for a total of 5 encoder inputs.

I also read, that i could convert one of the unused step generators (id only be using 2) to output PWM, which i could pass through a converter to run the spindle VFD.

Am i on the right track with that?

Please Log in or Create an account to join the conversation.

More
24 Dec 2021 01:19 #229846 by PCW
if you use 7i95_1pwmd.bit firmware, stepgen 5 is replaced by a pwmgen.

Please Log in or Create an account to join the conversation.

More
24 Dec 2021 03:13 #229853 by TurboStreetCar
Yep! Thats what i read about.

When i configure the board, i would only configure it with 2 step generators, 1 PWM output and 6 encoder inputs.

This should give me extra outputs from the unused step generators, and the 24 inputs should be enough I/O for what i want to do.

I know linuxcnc is highly customizable, but Does this all sound reasonable? Just trying to make sure i have my head on straight before i start buying the the wrong cards.

Thanks so much!

Please Log in or Create an account to join the conversation.

More
29 Dec 2021 21:37 #230281 by andypugh
If you feel that the encoder feedback is valuable, then you should probably use the 7i95. However if you are using step-dir for position control then it is likely that the encoders are mainly used by the drives, and LinuxCNC might not need to know about them.

Please Log in or Create an account to join the conversation.

More
30 Dec 2021 01:43 #230306 by TurboStreetCar
That was a question i had.

The original configuration, used a DOS based computer control. Encoder feedback is indeed used by the SGDB drives (i believe for PID loop feedback to control acceleration rates and position accuracy/verification), but it is also passed through to the control for further position control/verification.

I would need at LEAST two encoder inputs for the handwheels for Z and X.

The original control WILL throw following errors if you crank the handwheels too quickly. They're electronic handwheels (i believe the term is MPG, but they're smooth, no detents), so if you flick them really fast it will throw a following error as the servo cant jog the axis quickly enough to keep up with an instant position change.

My thought process was to send the servo encoder pulses to the control to maintain the position feedback loop like the original configuration

It seemed the 7i95 made this simple, but id be open to other better opinions.

Can one of the 7i95 encoder inputs be used for spindle encoder? I would think so, but im not sure if there is a difference between encoder inputs for axis vs spindle.

What would be used other then step/dir with servos? I figured that was standard?

Thanks,
Dan

Please Log in or Create an account to join the conversation.

More
30 Dec 2021 01:53 #230307 by andypugh

I would need at LEAST two encoder inputs for the handwheels for Z and X.


You could look at a 7i73 for user-interface control, or possibly consider software counters (1kHz max count rate, but that can be OK for MPGs).

The original control WILL throw following errors if you crank the handwheels too quickly.

LinuxCNC won't. It will either run as fast as it can, and stop when you stop cranking, or run as fast as it can, and stop when it gets there. This is configurable.

Can one of the 7i95 encoder inputs be used for spindle encoder?

Of course. You can connect anything anywhere in HAL.

What would be used other then step/dir with servos? I figured that was standard?

Traditionally a servo system would run a position PID loop in the controler, based on encoder feedback, and use an analogue +/-10V velocity command to the drive.

Please Log in or Create an account to join the conversation.

More
30 Dec 2021 02:23 #230309 by TurboStreetCar
Hmm, the 7i95 encoder inputs wouldnt work for MPG inputs? I figured i could port the encoder input to an MPG input somehow.

I was under the impression LinuxCNC had a following tolerance, which if given feedback, would be monitored and throw and error if outside tolerance?

Funny you say that about the analog control loop. I was just looking back through the wiring diagrams, and it seems the original control, connects to pins 5 and 6 of the SGDB servopack. IT appears, that is the input pins for speed/torque control, which uses an analog voltage.

However, these SGDB servopacks CAN accept step/dir inputs aswell, i would just need to change a parameter or two in the drive.

Is there any benefit to the analog control over step/dir?

The input is MUCH appreciated. Thankyou!

Please Log in or Create an account to join the conversation.

More
30 Dec 2021 02:59 #230310 by andypugh

Hmm, the 7i95 encoder inputs wouldnt work for MPG inputs?


They will definitely work. But may be overkill with three channels and a 10MHz count rate.

I was under the impression LinuxCNC had a following tolerance, which if given feedback, would be monitored and throw and error if outside tolerance?


Yes, absolutely. But it makes no sense to allow that to happen with frantic MPG twiddling, so LinuxCNC simply limits the axis speed if you jog too fast.

Is there any benefit to the analog control over step/dir?

Not really. Both work, and work well.

Please Log in or Create an account to join the conversation.

Moderators: cncbasher
Time to create page: 0.160 seconds
Powered by Kunena Forum