[woodturn]

I had the same I idea, but could not figure it out, but maybe you can. I think that motors on this machine is under power, new motors are nice

[Alex_Cole]

I am not sure on that one oldkitster. You can contact Dyna and buy the wiring diagrams for the driver cards. They will probly charge you $50 or so dollars for it. This would be the safest way to keep from smoking somthing.

AC

[keebler303]

Kitster

I am starting to do the same as you propose. I called Dyna and asked if they could tel me the pinouts for the drives. They said they can't tell me because they still service these machines and so the pinouts are proprietary still. He told me that when the machine becomes obsolete and they no longer service it, they will release the schematics. He did say that they are step/dir type drives but would not comment further. I will be digging into it in the next few weeks.

Matt

[mike944]

Kitster I am starting to do the same as you propose. I called Dyna and asked if they could tel me the pinouts for the drives. They said they can't tell me because they still service these machines and so the pinouts are proprietary still. He told me that when the machine becomes obsolete and they no longer service it, they will release the schematics. He did say that they are step/dir type drives but would not comment further. I will be digging into it in the next few weeks.

Matt

I reverse engineered out all the pinouts that you require. (posted below) I was attempting to develop a Plug & Play PC interface, but i have just about
given up on it. Posted is the results of a LOT of work on my part during the development, that i'm now freely giving out to everybody.


I spent a lot of hours experimenting, and i can not seem to make the original Dyna stwpper drivers work properly. I was trying to develop my interface to work with any software, however, software/hardware limitations seem to prevent this.

The short answer is: Yes, the dyna drives are step/direction drives, however, they do not seem to be able to use the short-duration "pulses" of PC-based CNC controller software. The original Dyna controller feeds them a variable-frequency square wave signal into the "step" input, NOT a relatively short pulse, and a variable delay like modern stepper drives seem to be able to use.

The Dyna drives are full-step drives, but they do not seem to hold the phases "on". It seems like they just turn them on momentarily for long enough to move the motor, and then turn them back off again. This belief is compounded by the spec sheet of the motors, and current rating of the 24v power supply, which is not really large enough to handle all the motors at once. My latest belief is that perhaps the "on-time" of each coil is related to the pulsewidth. I can't check this, because I do not have a real oscilloscope, just a jerry-rigged device that plugs into my PC sound card that works with 5v only. Perhaps someone with an oscilloscope can check this out.


Using various software, i have had mixed results. Nothing acceptable yet though. The best luck i have had is with TurboCNC, which is a DOS-based software. Personally, i think this controller is fine, however, as i said, i was trying to develop a product that would work with all PC-based controller software. TurboCNC works using a pulsewidth of 50ms (TurboCNC documentation says this is in microseconds)

EMC2 and Mach3 can't seem to get pulsewidths this long, and seem to drop steps during the accel/decel portions of the move. EMC2 does not seem to have adjustable pulsewidth (maybe i just don't know how to set it properly) according to their documentation "step signals are turned on during one cycle, and turned off the following cycle" I can not seem to get the speed slow enough. mach3's "motor tuning" screen says step pulse range is 1-5 microseconds, however, it accepts a value up to 25, which unfortunately, is still not quite slow enough.


Now, the pinout that you have all been waiting for.

The stepper diver cards: these have 2 connectors on one side, a 2-pin connector, and a 6-pin connector. I'll start at the 2-pin connector on top (for those who have their drives mounted in the cabinet) farthest away from the 47mf electrolytic capacitor and go in-order.

1: Ground
2: "active axis" (explanation below)

1: Limit switch output (explanation below)
2: Direction input
3: Step input
4: +5 logic supply to drive
5: Ground
6: +24v motor power to drive

Now, for the 2 explanations.
"active axis" (my name) means, that during the initilization cycle, this pin is used by the controller to ask if there is an axis present. It's really only important for the optional U-axis, which can be present sometimes, and sometimes not. I didn't write down exactly how the communication is done, because for my purposes this communication wasn't needed.

The limit switch output is normally held low, but when the switch makes contact, this pin goes high. This is buffered by the drive, because the switch operates differently than this. however, it is a very convienent behavior for connection to a PC.



Now, i'll tell everybody the real secret of easy connection to this machine. Remove the "machine distribution board" from the left-hand side of the machine (looking from rear) Notice the 2 large chips in sockets (INS8243N) The lower one controls everything. You can remove it, and replace it with a IDC DIP header, and run the ribbon cable to whatever interface you want to make. Everything is available at this socket, and are logic-level inputs/outputs at low current.

The pinout of the DIP socket is as follows. (using standard chip pinout notation, down one side, and up the other)

1: U-step
2: U-dir
3: Z-dir
4: Y-dir
5: X-dir
6-11 are used for communication with the other 8243 chip, which interfaces with the controller
12: Gnd


13: Buzzer out
14: tool probe in
15: spindle on
16: Not connected
17: X-limit
18: Y-limit
19: Z-limit
20: U-limit
21: X-step
22: Y-step
23: Z-step
24: +5v

[mike944]

Oh, and by the way. Yes, the original dyna motors are kind of on the small-side power-wise, but they do run through a 10x1 reduction gearing system, which means they can produce plenty of torque on the screw for any reasonable cut you might want to ask of this machine. The 10x1 reduction means that the max table speed is kind of low, but hey, it's a pretty small machine, and i don't think most of us are using these as production machines anyway, so personally, i think the speed is fine as well.

[keebler303]

Thanks a lot Mike!

That will help save some time, however it may be bad news with the step pulse width as I was planning for a plug and play with mach. I guess I will play around with it a bit and maybe upgrade if necessary.

We will see what I can come up with.

Matt

[mike944]

I'm not 100% sure exactly why the drives don't step properly, i'm just giving everybody the results of my testing. I haven't figured it out completely, it's just a hypothesis, based on the fact that the dyna controller DOES output a square wave, and the drives don't seem to work reliably with anything less than 50ms pulsewidth.

If you do figure something out, please let me know.

If anybody has the hardware to read the program off the EPROM on each drive board, that knowledge might help a lot to understanding the problem. Maybe some new EPROM software would fix the problem. If someone has a burner that can read/write an 8748H (40 pin old-style EPROM) let me know.

Maybe someone who is good with software can make a mod to EMC2 to output square waves as an option.

[mike944]

Ok, so posting this prompted me to do a little more experimenting.

Looks like the dyna drives use a fixed "power-save" period of approx .08 - .11 seconds after the last pulse, where they shutdown the motor coils. This is in the EPROM program (actually, it's a single-chip microcontroller with EPROM memory) This means it has nothing to do with the step pulse width.

Using TurboCNC, i messed with pulsewidths and made some measurements using my hack-oscilloscope (see prior post). 25 Microseconds seems to indeed miss steps. It didn't seem to matter what the jog speed was, it dropped random steps.

The missed steps are measured directly from the microcontroller output, before the power even gets to the motor power drivers. i'd say about 1 in 1000 steps are missed. Not a lot, but since stepper systems have no feedback loops ANY missed steps is a problem.

I didn't try any times in-between 25 and 50 ms, because as far as i know, no software, other than turbocnc, can go above 25ms. There doesn't seem to be any gain in consistency above about 50ms

I think i figured it out. The answer doesn't help your problem, or mine. I think it's related to clock speed of the 8748 microcontroller. if the step pulse isn't on for at least one complete clock cycle of the microcontroller in the stepper drive, then it can get missed. Remembering this machine is built with 1980's technology, the cycle speed to execute one program cycle of the microcontroller is probably pretty slow by modern standards. Probably just above 25ms. depending on the exact start point of the pulse, a pulse that starts just nanoseconds after the instruction to read the step pin, can possibly be back off before the next instruction to read the step input pin in the next program cycle.

I guess the only hope of re-using the dyna stepper drivers will definitely require software capable of making step pulses longer than 25ms. any other options besides turboCNC?


Matt - Don't let this detract you from doing experimentation on your own, it would be nice to have someone else come up with the same, or even completely different conclusions.

[keebler303]

Mike

I have access to a proper oscilloscope so that should help somewhat. What you are saying with the clock speed makes sense for sure. Do you know what the duty cycle of the square wave the dyna controller outputs? I think a solution to the problem could be a simple circuit which "stretches" out the pulse width, like a delay off of sorts. Basically increasing the duty cycle of step pulse outside the computer. I will do some research into a circuit which does this.

It will be the second week on January before I get a chance to start working on the machine, but I'll definitely post any progress.

Matt

[keebler303]

Mike

I think this may do the trick:

http://focus.ti.com/lit/ds/symlink/sn74121.pdf

Its a monostable multivibrator with an adjustable delay. The delay can be set from nsec to seconds so it is well within the range. When I get some time I will work on a circuit layout to incorporate 3 of these chips and will interface to a BOB and to a header to plug into the dip socket you kindly listed the pinouts for. This board will pass the dir signals directly through, run the step pulses through the sn74121 chips, and pass the limit switch signals back to the BOB.

These parts are in stock at mouser if you are interested.

Matt

[oldkitster]

Mike and Matt: This is great info, but way above my though process and is way I asked the guestion. My mill is a dyna 2800 so hope what you are working on will help with mine. If either of you come up with fix I will be in the market to purchace several as I also have 2- 3000 lathes which appear to use the same cards as the 2800. (I just hope the 2400 uses same card) Thanks for the help and I will be watching for more updates, I also want to use Mach 3.

Bob

[mike944]

We had some off-line discussion and Matt has prompted me to get working on this again.

My original plan was to offer an interface card that replaced the RS232 connector on the side of the cabinet with a small board mounted inside the cabinet, and a cable that plugged into the chip socket described above on the the machine distribution board.

I'm still developing this interface. I have a working prototype, however it suffers from the pulsewidth problems described above. I have redesigned it based on discussion, and should have a prototype built and in the testing phases within a few weeks. This is only a side-project for me, i have a day-job as a mechanical engineer, so i don't work on this regularly.

However, as discussed, the interface is designed for the 2200/2400 machines, because that's what i have available to me for testing. In order to fit it to other machines, I would need help from someone with one of those machines.

I suspect that the axis drivers operate the same, perhaps they are rated to drive more motor power though. If dyna was smart, they would have made the controller, and machine distribution boards the same. Any chance you can take some photos inside your 2800 and 3000 electronics cabinets? A quick visual comparison will tell us if they're at least close enough to warrant more research into compatability.

[Rico55]

Thanks to all who have posted in this thread, I am finally getting my Dyna2400 running with new steppers, Xylotex boards and Mach 3. It would save me a little head-scratching if someone could tell me what they have for settings in the bottom line of the Motor Tuning dialog (# of steps, etc.).
Also, is there a small Jacobs chuck available to fit this machine, without losing too much Z height?

[mike944]

in order to get all that info, we need way more information about your setup.

Are you using the Dyna gear reduction system with your new steppers, or did you connect them straight to the screws?

How many steps per rev are the motors? Are you using microstepping?

The screw ALONE is 10.16 rev/inch. (2.5mm pitch, 25.4mm=1 inch)

Multiply that 10.16 by the number of full steps per rev of your motors, multiply by number of microsteps, and multiply by any gear reduction.

For example, in the original dyna setup.
10.16 x 100 (steps per rev of the motor) = 1016

1016 x 1 (full steps only, no microstepping) = 1016

1016 x 10 (10:1 gear reduction) = 10160


Fill in your numbers here.

[Rico55]

Thanks, Mike. That probably gives me enough info to tweak my settings- I am not sure what the microsteps are set to, but I will check and see. I connected directly to the lead screw, and the number of steps is 200. I am still having a problem getting the machine to home properly; I know the switches are working properly, and can get them to stop the program when set as a limit switch, but when I try to home the machine, it will travel part way toward the switch, stop and then zero out the axis setting. Could someone share their settings in the Home/limit tab? I am still not sure what some of these settings are supposed to do, or how to set them; "Slow Zone"?, "Soft Max/Min"?, "Home Off."? I don't see a very clear explanation of these settings in the Mach 3 manual, can anyone clarify these for me? There is also supposed to be a way to jog off the limit switch, after it stops the program, but I am not seeing this button anywhere- where exactly is it?
Thanks again for everyone's help.

[woodturn]

Check out the video tutorial on the Mach 3 web page. Very good. It explains a lot. Good luck
Dave

[oldkitster]

Matt944: I got some time to check the Dyna 2800 driver board and it also has the two 8243 chips on it so would expect the plug and play board you are working on to work with it. Question about your set up, I assume that you reconnect the chip after you install the header in its place, does this keep all of the functions of the Dyna controller keyboard working? Do all of the axis switches then function when under PC G-code control?

Oldkitster

[keebler303]

kitster

It sounds like the fix we are working on will work for you. You do not reinstall the chip and the stock keyboard is no longer used at all. All of the functions of the controller are taken over by the computer. Here is a picture of the board I made for a 2400. It is not fully functioning yet so I have not shared much info about it.



Matt

[mike944]

We'll let you know when we figure it out.

I have information to indicate that the 2800 is controlled almost identically to the 2400. the same interface should work.

[ckm]

However, as discussed, the interface is designed for the 2200/2400 machines, because that's what i have available to me for testing. In order to fit it to other machines, I would need help from someone with one of those machines.

I have a 4400 with tool changer and all I'm about to hack into. I have an oscilloscope and other diagnostic tools. I can do some basic checks if someone wants.

My first task is to get whatever wiring diagrams I can (and I have all the current machine manuals) and then trace wiring to fill in the gap. Luckily, I have tracing equipment...

BTW, the information given so far is really, really good. I had hoped to re-use the existing stepper drives, but that seems like it may not be possible. I might actually go the servo route as I have 2 Gecko 320's sitting on my bench...

Chris.