[dave design]

Last October I scored this mill on ebay for $1400. As you might guess, it was sold as a "non functional, for parts only" machine. As usual, laziness prevailed and I'm just getting around to looking at it now. Having a part time gig now also helps with a few pennies to throw at it.

This is not a benchtop machine like many other Dynas, but a full size enclosed VMC with box ways, 40 taper spindle and the Dyna Mechtronics 4M pc based control.

After maybe a half hour fudging around with making a power cable and wiring up to the panel, I was stoked to find it seemed to power up and at least look like a halfway functional machine. The display came on and I was able to navigate through the various position and status pages. I was even able to start the spindle and change the speed in MDI with S codes. But... I'm getting a persistent error 378 "DRV: In disable state!(1-X,2-Y,4-Z,8-Z,16-U) (-15)", and no axis will move when trying to jog or when attempting to home the axes. A little further poking around revealed that all the servo amps have their red fault leds on. They are Advanced Motion B40A20I amps. I pulled the drive disable pin on the drive connector to fool it into enabling (normally 5V, pulling low disables the drive), and the led stays red (indicating fault), even after a power cycle. The input power is clean 160V and the halls and encoders seem to be functional, and all the motor phases measure about 1.3 ohms, which seems right for these motors. As far as I can tell, the only status the control gets from the servo amps is the fault out, which is pin 16 on the amp, so I tried yanking that one as well (+5V=fault), and both the control and amps still show fault. There is abundant evidence of prior "trouble shooting", with lots of bits of wire and fastons in the bottom of the mag cabinet. I haven't traced out every cable yet, but the "big control board" is getting +5, +12, -12 and 24V, and no travel limit, oil/air low, or other switches are tripped, verified in the plc ladder and by checking the appropriate pins on the connectors. This is starting to look like it's deep down in that control board somewhere, yeah, the big one in the mag cabinet with all the ic part numbers ground off (crudely by hand, with a Dremel). But also having all the drive amps blown just seems wrong...

I'm not really asking or expecting someone to come up with a miracle solution, that would be like saying "hey, please give me $2000 worth of debugging advice for free". Instead of a cry for help in the dark, I'm hoping to make this more of a rebuild/repair/retrofit thread, with the hope that it might help the 4 other people in the world who own these machines ;+}

One more comment, I have been in contact with Dyna Mechtronics, and have found them to be more than helpful. They have experienced the down economy as everyone else has, but they still seem to be willing to go out of their way to help.

some shaky cell phone images:

[2ferrous]

Wicked good deal.

There are a couple of really knowledgeable guys in this forum section that could point you in the right direction.

I honestly couldn't say either way if it would be better to repair or retrofit.

Timothy

[dave design]

Semi-good news. I had thought all along that it's a really bad idea to have an enable pin on a 6Kw servo board pulled low to disable, but that is the standard way Advanced Motion ships these amps. But on page 41 of the installation guide on AMC's website I found the following (paraphrasing, adobe won't let me copy the text) "...this logic can be reversed, either with a dip switch or by removing a SMT jumper...", which is apparently what they did. Shorted pin 9 to ground and nothin but green.

The semi-bad news is the control is still saying everything is disabled when it shouldn't. Bummer.

[Caprirs]

If there were a problem with an AMC amp, only that amp would show the red LED while the others would show green. Thus, I doubt the issue lies downstream of the amps.

First, verify that the emergency stop circuit is not activated. On the front keyboard is a tiny LED that will flash when the emergency stop button is pressed or one of the axis over travel limit switches is tripped. The e-stop and limit switches are all wired normally closed so an open anywhere in the switches or wiring will cause this. If the emergency LED is not blinking...

Check incoming voltage from the wall. The machines are rated to run on 220VAC +/- 10%. In reality, the closer to the edges of that range, the more likely the machine will have trouble. The only items to use the three phases are the coolant pump(s) and the spindle motor/amp. Everything else runs single phase using the power from the big gray transformer in the bottom of the back cabinet. Use a meter and check each leg/phase to earth ground. If the incoming three phase from the wall has a wild/high leg, don't have that leg feeding the big gray transformer. Moving the phases around will affect the rotation of the coolant pump, but not the spindle.

On the Big Board in the back cabinet, find the three screw terminal strip along the bottom edge towards the right side. There will be 120VAC between the two spots with wires attached. It is critical that the voltage at this connection be between ~117-123VAC. This is because the DC power supply for the AMC amps is unregulated. If the voltage is outside that range, change the taps in the big gray transformer in the bottom of the cabinet. When the voltage is outside the specified range, the big board can prevent the AMC amps from enabling.

The machine should have a Mitsubishi spindle motor and amp. When power is applied, the amp should display E7 (I think). When enabled by the Dyna control, the display will read the current rpm of the spindle motor which better be 0 at start up.

Since the history of the machine is unknown, it's possible previous owners have tried some odd things. You'll have to check with Dyna, but it is possible someone has loaded software on the PC hard drive that is not compatible with the firmware in the eproms on the boards. There are a total of five eproms each with a sticker on the chip indicating socket and checksum, three on the big board and two on the keyboard. The software and firmware must all be matched.

Where is the machine located?

[dave design]

I'm seeing 120V exactly at the screw terminals, and the e-stop led is not flashing. I called Roberto at Dyna today, and he said the only thing that could cause the 378 error with no other "precursor" errors was the axis definition being set to all 0s. Sure enough, went to the parameter page and found 000. Set it to 111 and at least the green servo led stays on long enough to jog the y and hear a thump before it faults again and the 378 error comes back. If I don't try to move anything, the machine will sit happily with no drive faults.

Re the AC in, well here's a little confession. Since I realized the machine only needs 3 phase for the spindle, and I only have 1ph in my shop, I just hooked the 2 legs of the tranny up to my 1ph 220, exactly as you suggested. The spindle does not fault, and it does indeed show 0 when powered up. It starts and runs with no problems, and as you state, shows the rpm on it's status display. I have no coolant pumps, unfortunately, so no issues there but their AC wires are safely insulated and tied back. My EE buddies with inverter experience tell me the bus caps and bridge in the spindle drive won't really like 1ph, but it will work temporarily.

Judging by the notes on the manual and the other evidence previously mentioned, the previous owner definitely tried anything and everything. I may not be doing any better, but I have worked around motion control and heavy power stuff for a long time.

The machine is located in San Jose.

Thanks!!!

If there were a problem with an AMC amp, only that amp would show the red LED while the others would show green. Thus, I doubt the issue lies downstream of the amps.

First, verify that the emergency stop circuit is not activated. On the front keyboard is a tiny LED that will flash when the emergency stop button is pressed or one of the axis over travel limit switches is tripped. The e-stop and limit switches are all wired normally closed so an open anywhere in the switches or wiring will cause this. If the emergency LED is not blinking...

Check incoming voltage from the wall. The machines are rated to run on 220VAC +/- 10%. In reality, the closer to the edges of that range, the more likely the machine will have trouble. The only items to use the three phases are the coolant pump(s) and the spindle motor/amp. Everything else runs single phase using the power from the big gray transformer in the bottom of the back cabinet. Use a meter and check each leg/phase to earth ground. If the incoming three phase from the wall has a wild/high leg, don't have that leg feeding the big gray transformer. Moving the phases around will affect the rotation of the coolant pump, but not the spindle.

On the Big Board in the back cabinet, find the three screw terminal strip along the bottom edge towards the right side. There will be 120VAC between the two spots with wires attached. It is critical that the voltage at this connection be between ~117-123VAC. This is because the DC power supply for the AMC amps is unregulated. If the voltage is outside that range, change the taps in the big gray transformer in the bottom of the cabinet. When the voltage is outside the specified range, the big board can prevent the AMC amps from enabling.

The machine should have a Mitsubishi spindle motor and amp. When power is applied, the amp should display E7 (I think). When enabled by the Dyna control, the display will read the current rpm of the spindle motor which better be 0 at start up.

Since the history of the machine is unknown, it's possible previous owners have tried some odd things. You'll have to check with Dyna, but it is possible someone has loaded software on the PC hard drive that is not compatible with the firmware in the eproms on the boards. There are a total of five eproms each with a sticker on the chip indicating socket and checksum, three on the big board and two on the keyboard. The software and firmware must all be matched.

Where is the machine located?

[dave design]

Today reminded me of why I prefer mechanical problems. It's not like I haven't spent a huge portion of my career building, testing and troubleshooting electromechanical stuff from laser printers to Lasik machines, it's just that I prefer things that I can see and touch and not have to worry about letting the magic smoke out.

After swapping X and Y servo amps, removing the X amp completely and reflowing the solder connections on the power, I at least got the X axis working. Still nothing on the Y. The light on the servo amp showed green but when I tried to jog, I got a thump and nothing. I tried manually turning the Y screw (with power off!), uh oh, not gonna move it with these girlie hands. Tried prying on the coupling with a screwdriver, and was barely able to move it. As I kept walking it along, it seemed to get easier to turn and when I finally moved the slide far enough to get to the ballscrew and get a decent grip, it seemed only a little stiffer than X. Hooked everything back up, powered up the machine and it went into immediate drive fault when jogging, just as before. Cycled power one more time and magically it just started working.

[nomedia="http://www.youtube.com/watch?v=1RKR3c0sO80"]YouTube- Dyna Myte DM2016 resurrection[/nomedia]

Now you would think I would have a working machine now, right? Well, you might have noticed I haven't said much about Z. That's because the machine as delivered had *ahem* no Z axis servo motor or amp. As I've discovered with X and Y, these machines really have no way of knowing anything is amiss with any of the axes until you try to move something, like homing an axis. So I simply tied the fault line to ground on the Z amp cable connection. X and Y home and come up "ok" in the status display, but until Z is homed, nothing of interest can be done with the machine like entering data in mdi or running a test program (though I was somehow able to start the spindle and change the speed in mdi). I had a hongo Advanced Motion servo amp kicking around that was leftover from an EV project (ya, it's 24Kw peak), and scored an appropriately matched Parker motor with the same bolt circle as the original SEM from ebay for $400. Unfortunately, the amp uses D connectors instead of the 16pin header on the oem amps. I'll also have to find the jumper that sets the polarity of the enable line. Back to cables, crimpers and soldering irons...

[dave design]

Stopped by the shop after work yesterday to goof with the machine a bit. Still no Z axis cables done. I remembered Roberto at Dyna mentioning the "controlled axes" user parameter. When I got the machine it was set to 000 instead of 111, disabling all axes, and was the source of the "drive disabled" error. I wondered what would happen if I disabled only the Z. I set the parameter to 011, and sure enough, after X and Y homed, the position display changed to zeros and I was able to enter mdi, set S codes, and get the 2 working axes to do a G3. This is the most encouraging thing that has happened since I saw the lcd light up for the first time!

I'm thinking if I had an emergency part to make, I could put a hand crank on the Z ;+}

[Caprirs]

That's encouraging. Almost there.

[dave design]

After encouraging results, it's time for the reverser. The X amp now appears to be completely dead. When I was first playing with the machine, my buddy was standing at the back of the machine looking at the drive amps, he swears he saw both the green and red led flicker several times on the X amplifier, though I never witnessed this behaviour. That was just before everything suddenly started working, and has been working... until this morning. Absolutely nothing has changed, no cables moved, no parameters changed, nothing.

Determined to find the cause of this intermittent problem, I swapped the X and Y amps again, and this time, the error follows the X amp. The formerly working Y amp still works when driving the X motor, but the (recently) not working X amp is still dead when driving Y motor. I yanked it and reflowed thru hole parts like a madman, but to no avail. Then I knocked the silly little loop gain trimpot off the edge of the board. No good.

I know about the ebay guy who sells used AMC amps $100 less than full boat retail, but anyone know of a better source for new or used? I see at least Servo2Go has pricing online, but they don't list the B40A20I, only the non-isolated B40A20.

I was so close...

[Caprirs]

Again, working from old memory on this. I think the AMC amps are off-the-shelf units with a model number specific for Dyna. For a while, AMC would not repair or replace the amps unless the customer went through Dyna to get service. I do not know if that has changed since Dyna has not used the AMCs for a while.

[dave design]

While waiting for a check from my client so I can buy a new servo amp, I decided to install the Z motor and amp, and debug that and maybe mess with the toolchanger.

First problem was the wiring, I knew the new motor's pinout from the Parker website, but since it uses an MS connector, and there is no document that shows the function of pins on the Dyna pcb, I had to trace the wires from the encoder on another axis. Removing the cap on the X axis motor reveals a little pc board with an 8 pin connector conveniently labeled A+, A-, B+, B-, Index+, Index-, +5V and ground, great, now I can at least get the encoder wired up correctly and only have the guessing game on the 3 motor phase wires and hall connections. Wired up the amp, hooked up the phases and halls with my best guess, powered the amp and hit home, and got a thunk. Went through the hall permutations a few times and it seemed the axis was trying to go down, not up, so reversed the motor phase wires and eventually found a combination that allowed the motor to move. But there was another problem. When I cranked up the current on the amp to about 50%, the motor tried to run away, l mean fast! It was acting like the control wasn't seeing any encoder pulses. If I gave it some running room. it would shut down with a position tolerance error - after moving about 10" in one second! I don't have a scope any more, so I downloaded a software scope that uses the pc's soundcard. Only 20kHz max, but plenty of speed for seeing encoder pulses when turning slowly by hand. I Checked both channels and sure enough, nice clean square wave pulses, with 90 degrees between the 2 channels one way and 270 the other direction. I swapped the probe leads from Z to X and immediately realized the problem, pins 2 and 3 on the control board were index + and index -, not A+ and A- as I had been led down the garden path to believe. There were a couple intermediate cables, so I thought maybe I didn't trace it out properly, so I did it again. This time, I traced all 8 wires from the connector on the board with the silkscreen labels directly to the Dyna main pcb, and it definitely indicated that the labeling on that pcb was wrong. Oh well, not that uncommon really. Swapped the wires on the 9 pin connector at the Dyna control board and finally got the position loop closed. Motion was pretty jerky, but after some messing around with PID settings, everything calmed down and seemed to run pretty smoothly. Wrote a little program with a G84 tapping cycle and it ran, cool! Unlike the Haas at my client's shop, my machine seems to be able to do rigid tapping (the Haas has a 30K rpm spindle which didn't have rigid tapping available as an option). If they give me a Soldiworks file with 467 tapped holes, I'm doing it here.

I tried a T code in mdi, the Z went home then sent the carousel in and just stopped, saying something about the "home dog" not being engaged. What part of the hood are these guys from? A check of the user parameters showed the upper and lower toolchange positions were both set to 0. Made some crude measurements with a scale, adjusted the parms, and it was able to pick up a tool and put it back. Toolchange [nomedia="http://www.youtube.com/watch?v=KE_4EiXN9hI"]YouTube- Dyna toolchanger working[/nomedia] (warning, loud compressor running in backround!).

On to servo2go to buy the amp...

[dave design]

[nomedia="http://www.youtube.com/watch?v=QhSo7uy9-CI"]YouTube- Dyna Myte DM2016 cutting.3gp[/nomedia]

Not much to see, but it is really moving all 3 axes. Now I just need a coolant pump and a tool shank for the Renishaw probe, and an MPG, and a bigger compressor, and...

Regarding the Dyna 4M control. It seems to have a lot of interesting features, and is a great effort considering the time it was designed. I am getting used to it, but I wish they had spent another 10 bucks and added some hard keys for mdi, edit, and other frequently used functions. And how the heck do you get the main dro display to show the work offset position instead of machine position?

[Caprirs]

When looking at the machine position, the soft keys at the bottom of the screen should show, from left to right: machine position, path trace, work position, blank, solid, blank

The machine screen shows the coordinates relative to home with work coordinates displayed as smaller numbers to the right.

Path trace has options in the upper left to rotate the view and zoom in/out. A small icon shows the current views axes.

Position has the DRO relative to the current work offset with machine coordinates and distance to go displayed in smaller numbers to the right.

The unlabeled F4 soft key will bring up a virtual keypad for functions such as single block, essentially duplicating some of the keys on the keypad.

The solid screen is capable of showing a cut part, but not in real time. The virtual solid is defined in the user parameter page and all tool data must be defined in the tool parameter page. After running a program in simulation mode, immediately pressing the solid soft key will show a rendered cut part that can be sectioned and rotated. Also, only on this solid screen, there is a report button that appears along the right margin which will display the estimated time to run the program.

The unlabeled F7 soft key will show a screen used for diagnostics only. It shows the encoder position registers as well as other info not typically used by most users.