DIY CNC Milling Machine From Scratch

**totts** · 23-08-2016

Originally Posted by A_Camera

Disable jogging and fix the stalling issues first. You must be able to run rapids repeatedly without stalling. Once you can do that then you can start checking for the rest of the problems.

I set up the machine and had no problems with short code but long code I found I had missed steps (after machining that part) so I then ran code in rapid

G00X10Y10Z10
G00X0Y0Z0

Repeated this about 1000 times and it went back to 0

After running a few test in jog mode the motor stalls as per video.

I'm hooking up a new computer tonight to determine weather it's the pc or the Bob.

Sent from my iPhone using Tapatalk

**A_Camera** · 23-08-2016

Originally Posted by totts

I set up the machine and had no problems with short code but long code I found I had missed steps (after machining that part) so I then ran code in rapid

G00X10Y10Z10
G00X0Y0Z0

Repeated this about 1000 times and it went back to 0

After running a few test in jog mode the motor stalls as per video.

I'm hooking up a new computer tonight to determine weather it's the pc or the Bob.

Sent from my iPhone using Tapatalk

So you have no stalling at full speed? Run this a 1000 times:

G00X(maximum X)
G00X(minimum X)

...to check X and likewise for Y and Z separately to check stalling. Running as short as to 10 units (inches or mm?) can be too short to test for stalling, but it is enough if missing/extra steps are needed to be checked. Also it is better to separately test the axes because running all three at the same time may result in less speed per axis than the maximum.

**totts** · 23-08-2016

I had previously ran the x axis separately. And the y,

Some progress!

I have tried a new computer tonight and all the jerky ness has gone so I'm going to get a new desktop pc sorted and run my tests all over again and hopefully get some parts made!

Sent from my iPhone using Tapatalk

**totts** · 18-09-2016

Okay right after a new pic, running loads of test programs and the program itself in fresh air and then coming back down to a dti, and it returns to zero, I decided to run a test part.

Run the test part and I had it skip again argh!

I ran the program in fresh air but this time, I ran it with the spindle on!

And it skipped, so I've come down to interference.

Problem is, all my cable is cy cable, all grounded at one end, checked the earth continuity all is well.

But I have a cy cable on the vfd to motor cable which is earthed at one end and wondered if this could put interference down the earth? I've placed vfd at one end of my garage and the control on the other and currently running the program again so we shall see if I get the same problem.

Sent from my iPhone using Tapatalk

**routercnc** · 18-09-2016

After a few mystery e-stops I tried putting a filter onto the mains side of the VFD but no improvement. I then ran an extension lead from the other end of the workshop and plugged the VFD into that, instead of the row of sockets for all the CNC stuff (PC, monitor, control box power mains supply). Not a single problem since and has been good for months. Note that the VFD and motor cables themselves are still in the control box, next to all the other electronics. It's just the power lead that is connected remotely and that has solved it.

**totts** · 18-09-2016

Interesting, my vfd is on a different socket to the rest of the control, but it is on the same ring main just two sockets apart.

Sent from my iPhone using Tapatalk

**totts** · 18-09-2016

Well the program has finished and it still moved so I'm going to take the shield off the earth tomorrow

Sent from my iPhone using Tapatalk

**JAZZCNC** · 18-09-2016

You haven't earthed the PC to the control box have you.?

Also do you have Earth on the Frame.? I'm presuming all your Earths go to one Star point.?

**Boyan Silyavski** · 18-09-2016

Nothing can cure a crappy board with poor filtering on the signals / i don't know what that is exactly but i know some boards are good

and have it/ .

I had last week's similar problems, fortunately not with the motor output signals, just probes, switches and PWM output. Neiter filtering of line input, grounding, shielding or whatsoever helped. One of said boards worked well in other machine. All worked fine untill i switched my servos.

But in you case a shielded cable from board to motor drive might help. I have resolved a friends machine similar problem once, changing said cable with a shielded one. Problem solved. In fact his cable was shielded before but wires at end were very large, So i cut them at minimum so all is shielded. Changed cable just to be sure though, as i had a better one at that moment.

Are your cables between this and that shielded?

And by that mean that the shield must be grounded to case

**totts** · 19-09-2016

I've tried to make all my signal wires shielded, there are just over two inches long at either end.

Machine frame is earthed via the spindle from one point and all shields are all going to one point. Pc control cabinet is earthed too.

But my pc has a shielded USB cable which means the pc and the control box is actually earthed together and I have had a few false errors at some point saying it was disconnected.

Sent from my iPhone using Tapatalk