Part II: Converting Chinese CNC Router from Open Loop Steppers to DMM DYN4 Servos

[slodat]

Current status: Lots of progress and a few setbacks..

First thing that came up was the Z axis motor's encoder has an issue. I've isolated it to the encoder. DMM is sending a replacement tomorrow. After getting that isolated, I checked out all of the Acorn I/O. New inductive proximity sensors are working well, so I moved on to getting the X and Y axis homing working. I wanted to use the DYN4's ZRI for precision homing. A huge thank you goes to forum member Joe for all of his help.

The idea behind the ZRI precision homing is to use your home switch to get close and then use the DYN4 marker pulse for precision homing. I'm sure there are a lot of ways to do this. So.. here's what I did for homing X and Y (Z will work the same).

Code: Select all

M105 /X P-1 F50  ;Move X minus at 50in/m until input 1(X axis switch) opens
M106 /X P6 F1    ;Move X plus at 1in/m until input 6(ZRI+ input) closes
M26 /X           ;Set X home here
M106 /X P-6 F5   ;Back X off ZRI+ so next axis can use it

M105 /Y P-2 F50  ;Move Y minus at 50in/m until input 2(Y axis switch) opens
M105 /Y P6 F1    ;Move Y plus at 1in/m until input 6(ZRI+ input) closes
M26 /Y           ;Set Y home here
M106 /Y P2 F25   ;Back Y off Switch

Don't copy and past this. It works on my machine because of the relationship between the given axis' switch and where that motor's shaft was in it's rotation when I set it up. This is why on the X the ZRI move is in the opposite direction of homing and on the Y the ZRI move is in the same direction as homing. I have a lot of travel left after the Y axis switch opens, so I just keep going in that direction for the ZRI. On the X my switch is close to the end of the axis travel hard stop, so the ZRI move is in the opposite direction. You have to move off the ZRI before you go to the next axis so that the input can change state for the next axis.

I made the cable for the ZRI input by cutting this double male straight through DB9 cable in half, found pins one and nine, some heatshrink to act as an insulator and strain relief and ferrules. Pin1 goes to COM and Pin9 to the INPUT you are using.

The schematic I used:

DB9 ZRI cable:

Lastly, it seems I have a problem others have had - noise causing Acorn/DYN4 to lose steps. I have asked DMM to weigh in on this. I am using their DB25 cables and have separated AC power from the step direction cables. I'll see what the experts say..

A walkaround of the machine in its current state that shows the losing steps issue:

https://youtu.be/CpK6874EGfE

This video shows the issue on the Y axis, resulting in both Y axis drives faulting out.

https://youtu.be/AkBV7j-1284

[slodat]

Had a few issues to sort out.. One of the servo motors was wrecking havoc on the system. DMM sent a replacement immediately. Got it installed and that fixed it. Had an issue where I was losing steps between Acorn and the DYN4 drives. The solution I went with was to use the Acorn DB25 step and direction signals, remove the 2k resistor from the DMM DB25 cables, and connect the brown wire from the DMM cable to +5vdc. This took care of my dropped steps.

Before:

After:

Got the homing sorted out. My cncm.hom file, using the DYN4 encoder ZRI for precision homing on all three axis.

Code: Select all

; 2019-02-26 Homing file
; cncm.hom
;
N100 ;Check for open home switch
IF [#50001] THEN M225 #100 "X Axis Home Switch Tripped!!!\nPlease Clear Before Homing\nPress Cycle Cancel To End Homing"
IF [#50002] THEN M225 #100 "Y Axis Home Switch Tripped!!!\nPlease Clear Before Homing\nPress Cycle Cancel To End Homing"
IF [#50003] THEN M225 #100 "Z Axis Home Switch Tripped!!!\nPlease Clear Before Homing\nPress Cycle Cancel To End Homing"
IF [!#50006] THEN M225 #100 "ZRI Switch Tripped!!!\nPlease Clear Before Homing\nPress Cycle Cancel To End Homing"
IF !#[50000 + 6] THEN GOTO 100 

M106 /Z P-3 F15  ;Move Z at 5in/m until input 1 opens
M106 /Z P6 F0.5  ;Move Z at 0.5in/m until input 6 closes
M26 /Z           ;Set Z home here
M105 /Z P3 F25   ;Back Z off Switch

M105 /Y P-2 F50  ;Move Y at 50in/m until input 2(Y axis switch) opens
M105 /Y P6 F1    ;Move Y at 1in/m until input 6(ZRI+ input) closes
M26 /Y           ;Set Y home here
M106 /Y P2 F25   ;Back Y off Switch

M105 /X P-1 F50  ;Move X at 50in/m until input 1(X axis switch) opens
M106 /X P6 F1    ;Move X at 1in/m until input 6(ZRI+ input) closes
M26 /X L1276     ;Set X home here
M106 /X P-6 F5   ;Back X off ZRI+ so next axis can use it

g1x0y0f200

Should be able to get the cabinet buttoned up and covers on tomorrow and do some test cuts.

1200ipm rapids at 0.1 second accel/decel.
https://youtu.be/usCS3JnX1L0

Precision homing using DYN4 marker pulse.
https://youtu.be/mp3m75iBG-w

[slodat]

Did some more testing tonight on my X axis, looking to see if I'm losing encoder counts.

The test moves:

Code: Select all

g0x-4
g04 p1
g0x59
g04 p1

Results:

Code: Select all

		Start		After 25	After 50 more
g0x-4		  396092	  396089	  396089
g0x59		-5846136	-5846145	-5846149

After 75 traversals back and forth (63" moves) at 969ipm, largest error was 13 counts over ~9450 inches of travel. Accel/decal is set at 0.2s. This was with my spindle running at 18,000 rpm, varying the speed along the way from the VCP.

My solution was to remove the 2k resistor from the DMM DB25 cable, connect the brown wire of that cable to +5v, and use the Acorn DB25 step and direction outputs.

With the GEAR=4096 my max rapid is 969ipm. Here it is in all it's glory, running like hell!

https://youtu.be/mADw8_DT-p8

[cut2cut]

Awesome. Congrats !

Jake

[slodat]

Keeping all of this together.. This is how I used (modified) DMM's DB25 DYN4-> Acorn cable:

Brown - +5vdc (I bypassed the 2k resistor built in to the cable)
Green - Acorn DB25 step
Blue - Acorn DB25 direction
Orange - Acorn input5
Black - Acorn screw terminal common
Red - Acorn +24vdc screw terminal
Yellow - Acorn screw terminal enable

The DMM DB25 cable has a small pcb soldered to the db25. The 2k resistor is through hole soldered on that board between pins 11/22 and the brown wire. I desoldered the resistor and the brown wire. Then I soldered the brown wire in the former resistor solder pad connected to the pins. Thereby making a straight through connection between the brown wire and the pins. I then connected the brown wire to my +5vdc.

[Dan M]

So you got a bad motor and a bad encoder and the cables they sell don't work without modifications?

[slodat]

It's just one bad motor. I isolated the issue to the encoder on that motor.

Yes, the cables did not work out of the box for me. They are working now.

I am very happy with the DMM servo upgrade!

[Dan M]

What are the steps per revolution with the DMMs? I'm just curious how the setup works vs steppers.

Dan

[slodat]

What are the steps per revolution with the DMMs? I'm just curious how the setup works vs steppers.

Dan

Dan,

I'm at 16,384 steps/rev and 968ipm rapids.

[Dan M]

What are the steps per revolution with the DMMs? I'm just curious how the setup works vs steppers.

Dan

Dan,

I'm at 16,384 steps/rev and 968ipm rapids.

Ok cool, how many oz are the ones you got? I'm thinking about building a new machine from the ground up. I like the idea of AC servo's and the DMM are very reasonably priced. Thanks for the info I always like to hear from people who bought the stuff for personal use.

Thanks,
Dan