PDA

View Full Version : VFD interference problem



eurikain
06-03-2016, 11:06 AM
Hello everyone,

I've been struggling with EMI problems with my VFD. The second my spindle starts spinning, my motors start doing uncommanded moves. Last week it was obvious since I didn't use any shielded cable for the spindle and inside the control box.

However, yesterday I replaced ALL the cables going out of my control box with CY cable. Also, all small uncovered sections (close to the motors for example) are wrapped in aluminium foil and covered by a protective sleeve. In the control box, I replaced most signal wires with CY cables: drivers to connectors, VFD to spindle connector, BOB to VFD (FOR and speed control). I also connected all the screens inside the control box to the earth.

Still getting EMI, and it hasn't changed one bit.

I don't have good pictures of the control cabinet after I changed to CY cable, because I don't have a proper camera with me today. However, I did take a photo with my laptop so the quality is not goo but it'll give you a good idea. Plus, I add older photos of the control cabinet that were made BEFORE I used CY cables.

Before CY:
17850
17851
17852
17853
17854
17855
17856

After CY:

17857
17858
17859
17860

I'm pretty depurate now.. the only solution I see really is to take the VFD out of the control cabinet, power it completely separately from it and just pulling a connection from the control cabinet / BOB to the VFD for Mach3 control.

Can anyone please advise me ? Please don't judge my machine or control cabinet, this is my first build ever, and I am far from being an expert and this is the best I could do with what I know so far.

Thanks a lot !

All the best,
G.

Noplace
06-03-2016, 11:16 AM
Maybe I'm wrong but I dont see any grounding wires for your control cables, so I think you have to do that first. and perhaps a power line filter for the VFD and ferrite beads on all of them as well, also physically separating power lines from control lines.

maybe someone with more experience can comment as well :)

Clive S
06-03-2016, 11:45 AM
I am not surprised that you have emf problems I think I would remove the vfd from the control box plus sort out the rats nest of wires try to keep or the signal wires away from the power wires and avoid running them side by side.

If you have to cross power wires do so at 90' if possible. Have you connected all the screens to one star point? I don't think it helps having the PP ribbon cable I would try and replace that by connecting the PP cable from the PC direct to the BOB without the ribbon cable.

You have the basis of a good control box but you have not made the best use of the cable management in there.

edit: Is that a plastic control box? have you mounted everything on to a backplane (ie metal plate) I think one of your main problems will incorrect grounding

magicniner
06-03-2016, 01:52 PM
Different VFDs generate different levels of interference, I have a .55kW Omron on a small manual mill in the workshop that kills the radio and knocks the benchtop CNC mill about and yet the .75kW Siemens on the CNC mill itself causes no interference from 750rpm through to 7000.
It is fairly standard to have all the kit in one box but if it's a dirty VFD it may be worth relocating it if problems persist after you add an RF filter on it's mains input

- Nick

JAZZCNC
06-03-2016, 03:32 PM
Geoffrey got to be methodical to find this. Make one change then test before making another.

Start with small trivial things first like removing the Tie raps holding cables together around the machine and separate the VFD cable from Stepper or signal cables.
Next make sure the screens from the long cables go to the Star Earth not just the cables inside the box.

I See you have 4 wires going to the spindle and suspect one of them you have connected to the Spindle for Earth.? Remove this at the VFD end and try.
Also Only connect the spindle cable Shield at the VFD Earth not both ends.

After doing these I would remove the VFD from the Box as a test. If solves then fit a Line filter and try again but it still may not work if you have sensitve elelctronics in the control box.?. . Like Cheap nasty BOB.!!!!!

One last thing you haven't earthed the PC case to control box have you.? If so remove.

eurikain
06-03-2016, 06:52 PM
Alright, so after spending the day trying to fix this with Graeme on Skype, we seem to have found the source of the problem.


I have connected all screens to the star point. I've taken the VFD out of the box, power is separate, it's completely diconnected from the control box, and still, no improvements.


Turns out disconnecting the PC solves the issue. So it seems like there is a grounding problem with the BOB. As JAZZ said, it's a cheap chinese BOB, so I might have to change it.




edit: Is that a plastic control box? have you mounted everything on to a backplane (ie metal plate) I think one of your main problems will incorrect grounding
Fiber glass cabinet. Fireproof plate in the back. No metal.



Like Cheap nasty BOB.!!!!!
Jee JAZZ, you don't like them.. :D !



One last thing you haven't earthed the PC case to control box have you.? If so remove.
Nope, not intentionally, but apparently it is through the BOB.


One thing to note. When there is no power to the control cabinet, but the PC is plugged and turned on, the BOB receives 5V, but the 5VDC PSU led turns on, apparently because it receives power through the BOB ! Weird...

JAZZCNC
06-03-2016, 06:59 PM
Jee JAZZ, you don't like them.. :D !

Nope and now you know why. You have been lucky to find it quickly this isn't always the case, in fact often not the case. Which is exactly why when these troubles come my way first thing I do is rip out the BOB.:grief:

m_c
06-03-2016, 08:08 PM
You're most likely getting a ground loop through the parallel cable shield. As you're using a ribbon cable internally, try unscrewing the parallel connector from the cabinet and see what happens. If it cures the problem, you can either come up with some way of insulating the connector, or just cut the cable shield near the connector.

eurikain
06-03-2016, 08:13 PM
Nope and now you know why. You have been lucky to find it quickly this isn't always the case, in fact often not the case. Which is exactly why when these troubles come my way first thing I do is rip out the BOB.:grief:

Okay but now, where do I go ?

JAZZCNC
06-03-2016, 08:38 PM
If it cures the problem, you can either come up with some way of insulating the connector, or just cut the cable shield near the connector.

It's Fibre Glass cabinet so won't be that.!!


Okay but now, where do I go ?

The good old US of A. . . . . PMDX126.

eurikain
06-03-2016, 09:08 PM
The good old US of A. . . . . PMDX126.

How about NOT USA ? CSLab ? And while I'm at it, why not go with ethernet motion controller instead of a BOB ?

m_c
06-03-2016, 09:40 PM
It's Fibre Glass cabinet so won't be that.!!


Good point. I'll just go back to sitting in the corner and keeping quiet!

JAZZCNC
06-03-2016, 09:53 PM
Good point. I'll just go back to sitting in the corner and keeping quiet!

Nobody puts Baby in the Corner.!! . . . . You keep on Dancing Mc



How about NOT USA ? CSLab ? And while I'm at it, why not go with ethernet motion controller instead of a BOB ?

Well yes that would be my first choice but your talking different animal and cost depending on model.

eurikain
06-03-2016, 09:59 PM
Yes but if I order the PMDX 126, 160 Euros + tax when it gets here, you're looking at the same price than the 230 Euro 4 axis motion controller they've got at CSLabs. Plus, really, I think I can re-use it in the future for another build.

Are there any choices other than CSLab, for a decent controller ?

JAZZCNC
06-03-2016, 10:13 PM
Yes there are but by time you have bought decent BOB to match they are not any cheaper than IP-M. While some like Ethernet Smooth stepper may beat IP-M on features(that you won't probably use) Nothing touches the IP-M on quality for same money.

Just be aware the IP-M isn't ideal if your using Slaved motors.

eurikain
06-03-2016, 10:46 PM
What if I buy an ethernet smooth stepper as you told me a few weeks back ? ESS smooth stepper ? That would prevent the ground loop going back to the PC since it's ethernet based.

komatias
07-03-2016, 12:12 AM
What if I buy an ethernet smooth stepper as you told me a few weeks back ? ESS smooth stepper ? That would prevent the ground loop going back to the PC since it's ethernet based.

The Ess will help some but the ground loop will always cause issues. Have you powered the Bob with the USB cable? If so then have you tried to disconnect the cable and use a different 5v source?

eurikain
07-03-2016, 09:18 AM
Nope, the BOB is powered only using a 5VDC external power supply. Weird that I'm the only person with this issue.. I've got the exact same board as the one you sell on your website. Typical cheap Chinese BOB. But it works for everyone else.. why not for me :'( ?

I have a PCI parallel port expansion board. It might be interesting to install it and test if it improves or fixes the ground loop. It might be the native parallel port of my PC that's not working properly. I'll give it a try on Thursday.

JAZZCNC
07-03-2016, 03:46 PM
The Ess will help some but the ground loop will always cause issues.

It will with the ESS which think pulls Gnd from PC like the BOB but not with IP-M which is isolated from PC and pulls Gnd to Mains Earth. It also uses Differential signals with isolated Inputs which is why it's so very good at not being affected by noise.

That said you still need to be 100% sure it is the BOB and not some other cause.?

Clive S
07-03-2016, 04:07 PM
Have you tried getting rid of the ribbon cable and plugging the pp lead straight in to the BOB

eurikain
07-03-2016, 04:21 PM
Jazz, I'm not sure to understand what you are saying... will the ESS fix the issue IF it is a ground loop going to the PC ?

I have tried getting rid of the ribbon cable. Same problem, though I had disconnected the X axis motors because I didn't want to twist the gantry. But then it was a little more difficult to hear the noise but it was still there. I will do some more tests on Thursday, I'll connect everything back together to get the maximum noise, then disconnect the DB25 cable and see if that stops it. I'll let you guys know ASAP.

komatias
07-03-2016, 04:29 PM
Any ethernet socket is galvanically isolated on both sides. So it will not pass direct currently to and from the computer. The ground loop will then not be present. But if there is an issue in the wiring then it may cause issues nonetheless.

eurikain
07-03-2016, 04:36 PM
Yes, but the ground loop could still flow through the cable from the BOB to the ESS, or not ?

komatias
07-03-2016, 04:38 PM
It should not really unless the optos on the BOB are buggered or there is an issue on the board.

eurikain
07-03-2016, 04:56 PM
It looks like the board only has optos on the input signals though. But as you said, there are so many cases of people using this BOB without problem that we can't just assume it's the BOB that's the problem ^^.

Clive S
07-03-2016, 05:32 PM
But then it was a little more difficult to hear the noise but it was still thereWhat is noise you are hearing? You can't hear EMF

JAZZCNC
07-03-2016, 09:30 PM
Jazz, I'm not sure to understand what you are saying... will the ESS fix the issue IF it is a ground loop going to the PC ?

I have tried getting rid of the ribbon cable. Same problem, though I had disconnected the X axis motors because I didn't want to twist the gantry. But then it was a little more difficult to hear the noise but it was still there. I will do some more tests on Thursday, I'll connect everything back together to get the maximum noise, then disconnect the DB25 cable and see if that stops it. I'll let you guys know ASAP.

In theory yes it should but like been said depends if this really is the source of trouble.? That said I'd still change the BOB because it's clearly not very good because shouldn't happen even with parallel port.

Now like Clive ask's What Noise.?

I think you miss under stand what is meant by Noise. It's not audible so if your hearing any noise that's something else.?
If your using AM Digital drives then I suspect it's because you haven't tuned them. Or what your hearing is the current reduction in the drive working while motors are stood still. This will make an Hissing sound.

Electrical noise or EMF is a Silent Killer.!!

Very difficult to help from distance when it comes to noise. All you can do is be meticulous with wiring and gnding. Ensure separation between signal and power cables.
If still present then fit EMF Filters etc.

If still present then start back tracking.

Graeme
07-03-2016, 10:09 PM
I was helping (well trying) to help Geoffroy over the weekend with his "noise" issue. His current problem is that when he runs the spindle all the steppers start making random moves. It is these rapid random stepper moves that is the noise he is hearing.
We started looking into the shielding of the spindle cable and he managed to improved it, but no change to the random stepper movements. He then removed the VFD unit and all VFD wiring from his control cabinet -still the same issue. So it does not look like it is caused by the cable radiating noise into the BOB.
The last thing we tried was to disconnect the parallel port cable from the BOB. This did seem to stop the steppers from moving when the VFD was run.
The reason I suggested he disconnected his slaved X axis motors was when the spindle was run these motors were sometimes going in opposite directions causing his gantry to twist. Removing them kept his machine safe, but did lower the "noise" from the steppers.
It looks like the BOB is seeing noise on the signals from the PC when the spindle is run, which is then being used to send random STEP and DIR pulses to the stepper drivers. This could be caused by an earth loop between PC and control cabinet or by a poor 0v connection between the two. More testing will be required to find out which.
Hope that clears up the meaning of the "noise", any suggestions welcome as to how to solve this one.

eurikain
07-03-2016, 11:26 PM
Thanks Graeme, you couldn't explain it better :)

cropwell
07-03-2016, 11:57 PM
You may well have a ground loop, however you would do well to relocate your BoB so that the PP cable from the computer plugs directly into it (get rid of the unshielded flat form cable).

A good principle to follow when wiring up a cabinet is to put signal cables on one side of the cabinet and power cables on the other.

Cheers,

Rob

JAZZCNC
08-03-2016, 01:02 AM
The last thing we tried was to disconnect the parallel port cable from the BOB. This did seem to stop the steppers from moving when the VFD was run.

Does the BOB have Charge pump.? If so then disconnecting from PC will kill the Outputs so you won't get any movement.!!

Going to sound obvious but have you tried another parallel port cable.? Also you are using a proper straight thru cable not some Old printer cable.?

cropwell
08-03-2016, 08:30 AM
Does the BOB have Charge pump.? If so then disconnecting from PC will kill the Outputs so you won't get any movement.!!

Going to sound obvious but have you tried another parallel port cable.? Also you are using a proper straight thru cable not some Old printer cable.?


Hi Jazz,

This BoB does not have a charge pump. I have suggested that the OP uses the cable that came with the BoB (these are generally sold with PP and USB cables) if there was one. Also that flat PP cable inside the box needs to go or at least be shielded.

Cheers,

Rob

komatias
08-03-2016, 09:34 AM
Goffroy,

you did some work on the spindle if I remember from your other thread. You sure that the grounding is still intact in there?

Also I cannot tell from your photos if the shield of your stepper motor cables is grounded. Each of the sockets in the control box should have a solder tang to allow them to be grounded too unless they are designed to be used with a metallic box.

Finally, is your actual house wiring set up correctly with a good ground to real ground?

eurikain
08-03-2016, 11:14 AM
This is how the cabinet is set. All hight voltage power cables are on one side of the cabinet. Signal cables are close to the BOB, there is only 5V power cables that run next to them.

I'll get rid of the ribbon cable, but that doesn't solve my ground loop. Why would I have a ground loop while other don't ? Is it the computer that's the problem ?

eurikain
08-03-2016, 11:18 AM
Going to sound obvious but have you tried another parallel port cable.? Also you are using a proper straight thru cable not some Old printer cable.?
I haven't. I have a PCI PP card that I will try to install and use on Thursday, this might prevent the ground loop, who knows :)

Clive S
08-03-2016, 11:32 AM
I haven't. I have a PCI PP card that I will try to install and use on Thursday, this might prevent the ground loop, who knows :)It is important as others have said to try a different pp cable or at least check the continuity from pin one to pin one etc at both ends of the cable.

komatias
08-03-2016, 01:23 PM
This is how the cabinet is set. All hight voltage power cables are on one side of the cabinet. Signal cables are close to the BOB, there is only 5V power cables that run next to them.

I'll get rid of the ribbon cable, but that doesn't solve my ground loop. Why would I have a ground loop while other don't ? Is it the computer that's the problem ?


I think it is down to the fact that many use metal cabinets to house the electronics and also for the machine frame. Spurious currents then have an easier path to earth.

eurikain
11-03-2016, 02:02 PM
I confirm that it is a ground loop. I changed the parallel port today using a PCI card I have. Same noise issue. Disconnecting the PC fixes the issue.

komatias
11-03-2016, 04:52 PM
I confirm that it is a ground loop. I changed the parallel port today using a PCI card I have. Same noise issue. Disconnecting the PC fixes the issue.

have you got an oscilloscope to probe the parallel port?

eurikain
11-03-2016, 05:33 PM
Nope, sorry ! I have a voltmeter though :) Probably won't be of any help huh :'(

Clive S
11-03-2016, 06:30 PM
Have you tried a new pp cable yet?

Clive S
12-03-2016, 11:06 AM
I confirm that it is a ground loop. I changed the parallel port today using a PCI card I have. Same noise issue. Disconnecting the PC fixes the issue.How can you confirm that it is a ground loop problem!! Have you confirmed that your PP cable is not just a printer cable (not all printer cables are the straight though type) You can check it with a multi meter

magicniner
12-03-2016, 11:41 AM
Is there a reason why you don't have an RFI filter on the VFD power lines close to the VFD?

eurikain
12-03-2016, 04:45 PM
Well, after trying to sort things out again with Graeme, we came to the conclusion that there was no ground loop. The PP cable is straight through, and there is nothing wrong with the 0V connections. So, I decided to take everything out of the control box, and next week I'll wire it step by step and check for interferences at each step.

Clive S
12-03-2016, 05:44 PM
Well, after trying to sort things out again with Graeme, we came to the conclusion that there was no ground loop. The PP cable is straight through, and there is nothing wrong with the 0V connections. So, I decided to take everything out of the control box, and next week I'll wire it step by step and check for interferences at each step.Put in all in a metal box if you can

eurikain
12-03-2016, 08:57 PM
I need to find one first. Probably going to get a bigger cabinet so I have more room to work with. But beforehand I need to make things work.

cropwell
13-03-2016, 12:50 PM
I would power your VFD from a RFI filter such as this one http://www.ebay.co.uk/itm/RASMI-14A-Single-Phase-RFI-Filter-for-Inverter-/182030644340?hash=item2a61df3474:g:SGYAAOSw~OVWyKN V and put ferrites on your PP cable at each end - like these http://www.ebay.co.uk/itm/20-Pcs-Gray-UF70A-Clip-On-EMI-RFI-Noise-Ferrite-Core-Filter-for-7mm-Cable-/121453826500?hash=item1c473651c4:g:mBoAAOSwCypWpz6 J but you won't need 20 of them and you have to get ones that fit the cable.

Anyway - general question - Anybody know ? - do the better quality VFDs have RFI on the mains input ? and do the Chinese ones not have it ?

Hope this helps,

Cheers,

Rob

komatias
14-03-2016, 10:12 AM
My Teco did.

It would trip the RCD in the 32amp wall socket though. It had an option to disengage it by snipping a wire.

routercnc
14-03-2016, 02:30 PM
Not related to interference but for safety - I was just looking at the 4th photo in post#1 ('dsc_0004.jpg') and noticed that the blue and brown cables going to the toroidal power supply appeared to be connected to incoming cables using male/female spade connectors. There looked like there was exposed metal on at least on of them.

If they are indeed the mains supply cables I don't think this is a good idea. Either put some heatshrink over each joint, or better still solder the cables together, heat shrink each individually, then heat shrink over both.

JAZZCNC
14-03-2016, 02:40 PM
My ABB has one but like George it tripped the RCD. Thou mine was easier to disengage it was just remove a screw.

Clive S
14-03-2016, 03:08 PM
Not related to interference but for safety - I was just looking at the 4th photo in post#1 ('dsc_0004.jpg') and noticed that the blue and brown cables going to the toroidal power supply appeared to be connected to incoming cables using male/female spade connectors. There looked like there was exposed metal on at least on of them.

As they say, eyes like a sh#thouse rat. well spotted:thumsup:

cropwell
14-03-2016, 03:34 PM
As they say, eyes like a sh#thouse rat. well spotted:thumsup:

I think the connectors are not fully pushed together. There are other mains connectors fully exposed though, like on the SMPS's for 12v and 5v.

eurikain
14-03-2016, 11:24 PM
So I unwired everything in the control box. Rewired separating power lines from control lines, and reducing enormously the amount of cables in there by chaining +5V and enable between the BOB and drivers. I now use 2 core shielded cable for the PUL/DIR signals. That made the wiring a heck more simple. Limits are not yet wired, homes neither, and the 24V PSU is not powered. Guess what ? Works like a charm !

So now I'm going to move forward step by step, making sure that each increment does not cause noise to appear.

Things look muuuuuuch neater now :) I'll keep you posted. As for the ferrites, I'll order some of them just as a safety.

Clive S
14-03-2016, 11:41 PM
Ok sounds good don't forget the pics:encouragement:

JAZZCNC
15-03-2016, 12:19 AM
and reducing enormously the amount of cables in there by chaining +5V and enable between the BOB and drivers.

Expalin or show this please because if what sounds like this is not good.?

cropwell
15-03-2016, 01:55 AM
Hi Jazz,

I hope he means that the enable (-) are being daisy chained from the enable terminal on the BoB and that separately he is daisychaining +5v on the BoB to step(+) dir(+) and enable(+) on each driver (chained on to the next driver and the next and the next). It'll work (Hmmmm ?) but I don't think I would do exactly that.

Cheers

Rob

eurikain
15-03-2016, 08:42 AM
Rob is right. No worries, I meant "daisy chained". They each get 5V and the correct amount of amps. Same thing with the enable.

Bad news, I introduced noise again, then rolled back to the state I was in when I posted my first message, and now the noise is not going away. Same thing, though much better than it used it be: in the first second of the spindle starting to turn, as well as the last second before it stops, the steppers move a few steps.

I'll order a CSLab 4 axis motion controller today. Can anyone explain the following to me:
- Zapp Automation sells it for 335 Euros (shipping included) => 0 in stock, have to wait 7-10 more days
- CSLab sells it for 311 Euros (shipping included)

So, obviously I'd go for CSLab official's website.. or does anyone have a spare one to sell ?

cropwell
15-03-2016, 09:51 AM
Bad news, I introduced noise again, then rolled back to the state I was in when I posted my first message, and now the noise is not going away.
You obviously have not rolled back fully. The questions are :-
1. Is the noise coming from your VFD - which it very much seems to be.
2. Is the propagation back along the mains cable or through the air

The probability is a bit of both. Filters and shielding before spending money on a motion controller.

eurikain
15-03-2016, 10:43 AM
I have, really, there's only the drivers, 5V and 70V PSU wired. The VFD is powered completely separately from the control cabinet. It's a different wall socket. Yes, it could be the mains. I need to call my local electronics store for help, I'm really lost now ^^. Can't deal with this myself.

Clive S
15-03-2016, 11:15 AM
I have, really, there's only the drivers, 5V and 70V PSU wired. The VFD is powered completely separately from the control cabinet. It's a different wall socket. Yes, it could be the mains. I need to call my local electronics store for help, I'm really lost now ^^. Can't deal with this myself.Does your building have an earth supplied by the electric company or do you use earth rods in the ground. Are you still using the plastic control box? Pictures would help.

When you say the motors move a few steps when you turn on the vfd are they enabled and the motors locked when this happens?

cropwell
15-03-2016, 01:08 PM
Does your building have an earth supplied by the electric company or do you use earth rods in the ground. Are you still using the plastic control box? Pictures would help.

When you say the motors move a few steps when you turn on the vfd are they enabled and the motors locked when this happens?

I don't know the electrical safety regulations in Belgium, but I would guess they are much the same as in UK. I would expect the earth to be strapped to the plumbing although this is not necessarily a good ground contact with plastic pipes :stupid:.

JAZZCNC
15-03-2016, 04:18 PM
Rob is correct that you shoudln't be spending money without sorting the problem. While the Cslabs are good with noise this doesn't mean it will cure your problem. The noise could be being picked up in number of places, the signal cables to drives for instance which are between Cslabs and Drives.

You NEED to find and remove the noise at the SOURCE. Buy mains Filter for both the Control box and the VFD before buying anything else.

eurikain
15-03-2016, 09:12 PM
Yes, the building earth goes to the electric company I believe. Standard setup here. I am actually not using any box anymore, it's all out of the box, screwed on a fireproof plate (not metal). I am sorry I haven't posted pictures. Here it is:

The X motor plates have been unmounted to prevent the gantry from twisting. The fireproof board is out of the box for easy access and testing.
17947

All relays are not wired, there is no e-stop, no reset button, no contactor, not fans, no switches (only a switch on the power cord extension), no 24V power supply (Earth and Neutral are daisy chained with the 5V PSU, but there is no Load going to it as you can see).
17933179341793517936179371793817939179401794117942 17943179441794517946

So, here is what happens:
1. drivers powered, Mach3 in reset mode, motors disabled, VFD powered, spindle off
2. click reset in mach3, motors are powered, enabled and locked
3. I can jog fine on each axis
4. press RUN on the VFD, spindle starts spinning right away. half a second after pressing RUN, I can feel (with my hand) and hear the motors move a few steps (ike 2-5 I would judge by feeling). Obviously I can't put my hands on all motors simultaneously, so I try one by one (the VFD is wired separately so I can carry it around with me).
5. press STOP on the VFD, same thing, right before the spindle stops spinning (so I would say in the lower frequency, around 500/1000 RPM), the motors move a few steps.

There are two parameters to the noise:
1. the time offset between the moment I press RUN or STOP. I guess this depends on the time it takes to get to the problematic frequency.
2. the number of uncommanded steps the motors move

If I press RUN, then wait until the spindle reaches the desired speed (100Hz), then press STOP, wait until it stops, and repeat this process a few times, I would say the problem occurs 7 times out of 10. The number of uncommanded steps and the time offset are always random.

Hope you can see things better now :)

Cheers,
G.

cropwell
15-03-2016, 09:30 PM
I can see one problem straight away. You have no main power to the BoB. In the (invisible) picture below the two middle connectors on the right are the main power (12 - 24v) for the BoB electronics. Without that you will get all sorts of random effects.

I tried to copy and paste from the pdf of the BoB manual, but it didnt work - anyway OP has BoB manual.

routercnc
15-03-2016, 09:37 PM
Hi G,

Here are some thoughts on interference in general. My machine was playing up recently with 'external e-stop requests' stopping the machine at random. Strangely the e-stop circuit (24V relay) had not tripped so all the drives were still powered etc. Basically either the BoB or PC/Mach3 had decided it had seen an e-stop signal from somewhere, or some noise, which it decided was an e-stop. This was a pain if you were cutting as it meant homing and 'start from here' messing about.

It could occur during cutting or even just jogging with the spindle on. I fitted an EMI filter in the invertor mains supply, but no difference. I then fitted a MeanWell mains powered DIN rail 5V supply to supply the BoB instead of a 5V feed from the PC, but no difference. I then used an extension cable reel to plug the invertor into a socket on the other end of the garage and then cut out a part without a problem. All the sockets in the garage are on the same ring back to the consumer unit but having some distance between the plugs seems to have helped. There were some unintended heavy cuts in there (including an accidental ~8mm DOC which snapped the 6mm carbide) and it still did not trip.

Going back to your BoB these are about 8 from China. I had one a year ago and it would not give out step/dir signals so in the end I swapped it out for a different make of board and all was well. You could buy one or two more cheap BoBs and give them a go before a major upgrade. They seem to be variable based on reading around and at that price you can see why.

Looking at your signal cables between the drives and step/dir on the BoB - you've used shielded cable, but I can't see if they are earthed?

eurikain
15-03-2016, 09:40 PM
I can see one problem straight away. You have no main power to the BoB. In the (invisible) picture below the two middle connectors are the main power (12 - 24v) for the BoB electronics. Without that you will get all sorts of random effects.

I know the pins you are talking about Rob :) Even when the BOB is powered by the 24V PSU, there is still noise :(

routercnc
15-03-2016, 09:45 PM
I seem to remember that you can also power it through the USB socket. You could try this from the PC, if this works then longer term use a USB plug and lead but cut the cables, identify the power supply cable and wire it to an external 5V supply.

eurikain
15-03-2016, 09:50 PM
Hello routercnc,

Thanks for your message. It is very interesting what you are explaining. It's something I haven't tried yet, plugging the VFD into a complete different socket out of my room. As you can see on the pictures, the room is pretty small, so there might be something there.

You are right about the BOB's signal cables. STEP and DIR are CY-2 0.5mm^2. The shield is not earthed. I know I should, this was just really me trying to get minimal connections.

I think you're right with the process, as others have suggested. So here's what I think I'll do:
1. try power the VFD from a socket located in another "block" of my house (one that's on a different breaker circuit).
2. If the above doesn't work, I'll put an EMI filter on the mains of the VFD (or inverter as you call it)
3. If that doesn't help, I'll buy spare BOBs from China.

eurikain
15-03-2016, 09:56 PM
I seem to remember that you can also power it through the USB socket. You could try this from the PC, if this works then longer term use a USB plug and lead but cut the cables, identify the power supply cable and wire it to an external 5V supply.

Yes, that's right, but there are two terminals made to replace the USB port, and I'm currently using them:
17948

cropwell
15-03-2016, 10:20 PM
I seem to remember that you can also power it through the USB socket. You could try this from the PC, if this works then longer term use a USB plug and lead but cut the cables, identify the power supply cable and wire it to an external 5V supply.
On the BoB OP is using the USB socket is connected by tracks on the board to the PC 5v and PC GND, so the USB socket is not needed.
But those shielded cables won't do much good if they are not earthed at one end.

JAZZCNC
15-03-2016, 11:03 PM
Yes, that's right, but there are two terminals made to replace the USB port, and I'm currently using them:
17948

This could be your problem. The BOB is tying PC ground thru the PP cable to GNd on the 5V supply. For test remove the 5V and power bob using USB.

Regards Spare cheap BOB's that's like simple adding 2+2=4.???. . . Crap + Crap = Crap so why bother.!

eurikain
16-03-2016, 01:06 AM
This could be your problem. The BOB is tying PC ground thru the PP cable to GNd on the 5V supply. For test remove the 5V and power bob using USB.
Okay this one also is a great idea ! That would explain why the noise does not appear when the PP cable is not connected ! I add this as first entry of my checklist.

Regards Spare cheap BOB's that's like simple adding 2+2=4.???. . . Crap + Crap = Crap so why bother.!
Always knew you were good at math :D

eurikain
19-03-2016, 01:09 PM
Well, guys, problem solved ! Guess who was right ??????????
......
.....
...
JAZZZZZZZZZZ !!!!!

The 5V signal to the BOB had two routes to 0V... getting the 5V from the PC solves the problem :) This is why disconnecting the PP cable was causing the noise to go away, since it would suppress one of the two 0V possible routes.

Thanks for your help !

I do have one more tiny issue:
Mach3 controls the RUN/STOP signal and spindle speed fine, but the BOB seems to output an unstable signal for the spindle speed control. When running, the frequency can't stay still and keeps going up and down by 1-2 Hz, causing the spindle to run faster/slower (you can hear a slight increase/decrease in speed).

I can bypass this issue by disabling Mach3 speed control and only do the RUN/STOP, but while I'm at it, let's fix this issue if possible :)

JAZZCNC
19-03-2016, 08:06 PM
Well, guys, problem solved ! Guess who was right ??????????

Always helps with solving problems if you give us all the facts.!!. . . . . It's was a rather important bit of info you didn't make available, esp considering the BOB and been accused prevously.!


I do have one more tiny issue:
Mach3 controls the RUN/STOP signal and spindle speed fine, but the BOB seems to output an unstable signal for the spindle speed control. When running, the frequency can't stay still and keeps going up and down by 1-2 Hz, causing the spindle to run faster/slower (you can hear a slight increase/decrease in speed)

It's called Crap Bob Syndrome and I doubt there's much you can do for it.!!

Clive S
19-03-2016, 08:13 PM
It's called Crap Bob Syndrome and I doubt there's much you can do for it.!!You could try giving it an Asprin :whistle:

JAZZCNC
19-03-2016, 08:16 PM
You could try giving it an Asprin :whistle:

Nah Bullet straight between the eyes works every time.:dejection:

eurikain
19-03-2016, 08:27 PM
You dorks :D

Well, a clue could have been that when the 5V PSU was off, its power LED was slightly on (not fully, but you could see some light), indicating that power was flowing from THE PC through to BOB to the 5V PSU (receiving -5V I guess).

Anyways, problem fixed, let's now move on to cutting !!!!

cropwell
20-03-2016, 10:07 AM
You dorks :D

Well, a clue could have been that when the 5V PSU was off, its power LED was slightly on (not fully, but you could see some light), indicating that power was flowing from THE PC through to BOB to the 5V PSU (receiving -5V I guess).

Anyways, problem fixed, let's now move on to cutting !!!!


Had you originally connected V- on your 5v PSU to GND ?

On your BoB the terminal marked PC GND is actually the current return path and should not be connected to GND but only to V- on the power supply. The BoB itself has a GND plane connection next to the 12-24v terminal, which is the current return path AND GND for the BoB electronics.

eurikain
20-03-2016, 07:07 PM
Yes that's what I had done. PC GND to 0V of the 5V PSU, and PC +5V to +5V of the 5VPSU.

Can you provide a quick diagram on how to do what you're explaining ?

cropwell
20-03-2016, 07:36 PM
Yes that's what I had done. PC GND to +5V of the 5V PSU, and PC +5V to +5V of the 5VPSU.

Can you provide a quick diagram on how to do what you're explaining ?


I trust you mean PC GND to -5v on the power supply.

On your power supply check with your meter for continuity between -5v and the earth terminal. It should not be connected internally in the power supply. If it is you will have a ground loop. These power supplies are generally not connected between the -5v terminal and the Earth terminal. Also make sure that if you use the 5v power supply that -5v is connected to PC GND on the BoB and nowhere else (I am assuming that you do not have any other requirements for 5v in your control box).

The Earth connection on the power supply should be connected to the star earthing point to provide shielding for the power supply.

In a conversation with Graeme, he thought that the power supply, if under loaded, may react badly to interference from the VFD and cause all manner of effects, in fact it could go unstable by itself. So even if you get the earth and current return paths right, you may still have problems as you are only driving the optos on the BoB.

Jazz hit on a solution, and the PC USB supplying 5v to the BoB is workable, but I do not like using USB ports to provide power as they have limitations.

If it is working now don't muck about with it, but I still don't believe that we have got to the root cause of the problem.

Cheers,
Rob

Clive S
20-03-2016, 09:01 PM
Rob I agree with you re not loading the power supply enough it would help if you put some sort of load on it even a 12v bulb would do it.

cropwell
20-03-2016, 11:20 PM
Rob I agree with you re not loading the power supply enough it would help if you put some sort of load on it even a 12v bulb would do it.


Well, it was Graemes idea, and I thought that a load like a bulb would shed some light on the problem.