Anyone got spare shielded four core cable?

[dharmic]

Interesting - and a relatively easy one to try.

Watching how my rubbish digital calipers go totally ape anywhere near the spindle gives a pretty good indication of the strength of the EMI being generated and, thinking about it, it will be just as strong in the box at the moment too.

[RCaffin]

Try putting your VFD outside the box your control system is in.

Absolutely!
I did this, and it helped enormously.

I also put a LARGE dual-winding ferrite filter at the output terminals of my DC spindle PS. Now, a VFD may have more wires coming out (4?) so you will need to put all 4 through a ferrite core of some sort. That blocks the RFI from going down the cable and radiating. VERY effective. Not effective enough? Try two ferrites.

Do NOT try putting suppression capacitors on the power leads. In some cases the cap just blows (been there, done that), while in other cases the cap stops the power supply from starting up (been there, done that).

The root of the problem lies in the modern fast-switching MOSFETs used in all these supplies. 'Fast-switching' means fast transitions, which means mega-RFI. But no-one is going to go in the other direction (slow-switching), because that way lies glowing transistors and sudden smoke.

Cheers
Roger

[dharmic]

Not sure I have a DC spindle PS - AC single phase in, gets split off the panic button and through a separate power switch then into the VFD. Three phase output. Control cable (what I'd guess are ground, VRef and VSignal) in from the breakout board.

There's a big old single ferrite ring between the mushroom and the mains plug. Mushroom also wired to a second power switch for the servo power supply. Breakout powered from USB only and is kept isolated from the case.

So first thing to try is loading it up with the 4th axis not dangling in a pile of chips where it can short out. Just in case that's all it is.

Second thing is to pull the VFD out of the enclosure.

Third thing is to run the control cable through a ferrite ring too. I note that OmioCNC have clamped not one, not two, but FOUR extra ferrite beads onto the USB cable along with the single one moulded into the original cable. Bandaid much? But if the board is fritzing out due to interference, it suggests to me that these clamps are probably in the wrong place.

Fourth thing I'm thinking is to nail the ground of the USB breakout down to a single star point along with the VFD, stepper supply, stepper drives and the mains earth in.

I can do all of these things without lashing out on cable which may not be needed.

Fifth thing is to re-cable with shielded wire.

I'm starting to think they've managed to bandaid it to a working state with a floating breakout and a couple of ferrites in the box but my extending the spindle cable has upset the bandaids just enough for it to get nasty occasionally. We'll see.

Once again thanks for the input folks, really nice to have help thinking all this stuff through!

[RCaffin]

No, you have a variable frequency 3-phase output. BUT, it is still driven by switching MOSFETS, and there will be still LOTS of RFI on the output wires.

From your description it does not sound as though any of the ferrites are where they really matter: on the 3-phase wires coming out of the VFD. If you can stop the RFI from running down the output wires, then (and probably only then) you will achieve something.

All the other ferrites you mention are no substitute. The ferrites on the USB cable are indeed bandaids, and barely adequate ones at that. Me, I switched to an ethernet interface because it is far more noise-resistant. I never looked back. Went from one crash a day to none in 12 months.

The ferrite on the mains lead is another matter. That is to stop RFI from going back up the mains into the rest of your house, There are SAA Regulations about that (requiring that). Without filtering there you could be giving your neighbours hell. You must keep them in.

nail the ground of the USB breakout down to a single star point along with the VFD, stepper supply, stepper drives and the mains earth in.

That will ensure safety at 50 Hz (and is a good thing), but sadly it will not stop all the RFI.

I'm starting to think they've managed to bandaid it to a working state with a floating breakout and a couple of ferrites in the box but my extending the spindle cable has upset the bandaids just enough for it to get nasty occasionally.

That is exactly my opinion and my experience too.

Cheers
Roger
PS: which State?

[dharmic]

No, you have a variable frequency 3-phase output. BUT, it is still driven by switching MOSFETS, and there will be still LOTS of RFI on the output wires.

From your description it does not sound as though any of the ferrites are where they really matter: on the 3-phase wires coming out of the VFD. If you can stop the RFI from running down the output wires, then (and probably only then) you will achieve something.

Good point, might see if I can dig up some ferrite toroid cores and run 'em on the spindle lines.

All the other ferrites you mention are no substitute. The ferrites on the USB cable are indeed bandaids, and barely adequate ones at that. Me, I switched to an ethernet interface because it is far more noise-resistant. I never looked back. Went from one crash a day to none in 12 months.

The USB BOB I have now won't run a fifth axis. Longer term I'd like to add a trunnion table and extend the gantry height a little to cope, that's about the point at which I'll be looking at upgrades and it'll be ethernet for sure. The only questions remaining are which one (smoothstepper, kflop etc) and whether any has the control code onboard (like Repetier etc on the 3d printers) instead of needing a host PC without costing both arms and legs.

The ferrite on the mains lead is another matter. That is to stop RFI from going back up the mains into the rest of your house, There are SAA Regulations about that (requiring that). Without filtering there you could be giving your neighbours hell. You must keep them in.

SAA? I thought it was the AMA here when things came to EMC compliance and OoE for electrical safety. Regardless, EMI is a bear of a thing and I don't want to be one of "those guys" inflicting it on the neighbours.

PS: which State?

Perth, WA for me.

[InMesh]

Oh yes I did the ferrite on all the leads as well as the Line filter

[RCaffin]

The USB BOB I have now won't run a fifth axis.

The Ethernet Smooth Stepper gives THREE full parallel ports of I/O. So far I have not filled the second port.
PC, WXP, Mach3, ESS, 2xMB-06 BoB from Homann, then custom.

SAA? I thought it was the AMA here when things came to EMC compliance and OoE for electrical safety.

Oh, you might be right. We went off the SAA handbook, but that was long ago. The pollies fiddle all the time.

> Perth
Too far away. I'm in Sydney.

Cheers
Roger

[dharmic]

Nah, only three days' drive

Come over for a coffee sometime!

And I made booboo with my correction, should be ACMA. They deal with all the telco rules, spectrum management, C-tick, A-tick compliance and all that crap.

Sorry - the smooth stepper is a parallel breakout only? Good to know. Hoping for something with a little more smarts. We'll see how that road goes when the time comes.

[RCaffin]

Um - yeah, I am not sure who manages that in oz. SAA handle mains installations and safety, but you are right: it may be ACMA for radiated interfence.
Oh well, must go, chips are flying: test run for making GT2 pulleys. Slow business as gentle cut strategy.

Cheers
Roger

[RCaffin]

the smooth stepper is a parallel breakout only?

NO way!
It is a full timing engine. Mach3 does zero timing when the ESS is there. The ESS does ALL the timing for pulses, and much faster than a PC could. It takes ALL the real-time load off the PC, so the CPU sits at about 5% utilisation when Mach3/ESS are running. Yes, it does drive the equivalent of 3 parallel ports, but Mach does not need to know about addresses if I remember correctly. just 'port 1', 'port 2' and 'port 3'. Powerful beast.

cheers
Roger