1.5KW JD220V to GRBL

[Avalonnw]

Hi,
I've got this 3040 CNC with a water-cooled spindle (not sure if it's 800W or 1.5kW) and it looks like it's got a quite unpopular JD220V VFD. I'm trying to convert the machine to grbl and my temporary solution was to run connect arduino to machine's parallel port anв control it like that. It worked very well but now I want to rid of internal parallel port board completely. And here is the problem. While it's easy enough to connect a bunch of TB6600 drivers and endstops to arduino directly, I can't figure out how to connect my VFD safely so spindle could be controlled as well.
I believe this parallel port breakout board has a little power supply built in (it receives 24v from main supply) and steps it down to 5V. For some reason, however, the spindle control part has an opamp (no idea what that thing is for here) and an optocoupler.
So, I guess, the questions are:
How to connect this VFD to arduino safely?
Do I need optocoupler and/or opamp for it?
Would 5V PWM from arduino be enough for this VFD (in manual it has this weird 0-10v/0-5v line and I can't understand which voltage it uses)?

Similar Threads:

• GRBL PCB
• Need Help!- any one that knows Audrino and grbl shield with grbl pannel?....
• GRBL jitters

[Eldon_Joh]

the vfd should have an isolated power supply, if so it should be safe to connect the arduino directly to the ground of the "vfd, 0-10v or 0-5v" input.

check to see that you have continuity between the "ground" on the input terminals of the vfd and the electrical ground of the unit, which should be bonded to the heatsink. edit: note that this ground is Not the negative dc bus

[Avalonnw]

VFD uses COM, FWD, REV and AVI (for PWM) as inputs. COM is NOT connected to chassis or ground. Also, to operate, corresponding pin should be pulled to COM. And it's 12V between any of those pins and COM. So, arduino can't handle it directly. Any suggestions how to control all that with minimal part count?

[Eldon_Joh]

If your vfd has a 10v output, you can use a pull up resistor and a mosfet or npn transistor so that the 5v output of the arduino can pull the 10v down to ground through the transistor, you can then filter that with an rc filter to get your 0-10v input to drive the vfd. note that this will invert the pwm.

anyhow the com pin should be safe to connect to the arduino, since its allegedly floating relative to the rest of the vfd. anyhow, you still need to check that you do not have conductivity between the com pin and the (-) dc bus

[Avalonnw]

I tried using an NPN transistor to pull FWD to COM with COM connected to GND. Does not work - grbl goes into "hard limit triggered" mode. I'm currently thinking about 3 4N25 optocouplers: 2 of them to switch FWD to COM and REV to COM and another one to pass PWM... Not sure how PWM would work though.

[Eldon_Joh]

what else was connected to com and gnd. i'm assuming by gnd you mean chassis ground.

[Avalonnw]

GND is my arduino ground and chassis ground.
I sort of localized the problem... I think. If I have limit switches disconnected and just short-circuit corresponding pins to GND on arduino or on a breadboard, switching spindle on/off is not a problem. As soon as I connect proper switches, switching spindle triggers them.
Maybe it's some sort of interference? I have switches cable shielded... Also I tried to use a set of capacitors and resistors to filter switches a bit; it allows me to switch spindle on/off but as soon as I try to speed up (manually for now) hard limit triggers again.

[Avalonnw]

Ok. Basically... FILTERING, FILTERING and FILTERING. RC filters on limit switches, ferrite core on all cables, and the problem goes away.

[Rubenvdm]

Have a look at this, does the filtering and isolation from the Arduino to the drivers
Worked for me!

Arduino CNC Isolation Shield – YR-CNC