[da_Rayman]

I have just upgraded my plasma machine from steppers to proper servos and a USB smooth stepper, which runs great until i start the plasma arc. sometimes mach3 stops immediately, sometimes it takes a second or two. I thought maybe the interference maybe came through the GND wires so i started to disconnect them, everywhere, just to see if there would be any difference (btw nothing changed). As you might have figured out by now, there is a lot of guessing involved here. Another thing i did (might worth mentioning that i did this first) was to bypass my breakout board. good? bad? From the pictures you might find out something that i did extra wrong or so.




Here is that error message i get:
"Error returned from USB function FT_white."
"Error: FTDI USB Driver Error"

"Error = FT_IO_ERROR"

"Bytes requested: 64"
"Bytes written: 0"

To the question. Does anyone have any idea what to do next? or should i start searching for a resting place for the USS under something very very heavy?

[neilw20]

The smoothstepper needs to drive a breakout board.
The smoothstepper looks just like a parallel port on the PC and needs to be treated as such.
FT_UIO error? Garbage is getting into the USB line.
Smoothstepper driving the optos in the drivers will be marginal at best.

[da_Rayman]

anyone got any idea?

[Christian Knüll]

Hi,

USB is very very sensitive to interferences from the mains line.
Just try running the smoothstepper with a notebook in battery mode.

Christian

[da_Rayman]

I tried other ideas today with no go.

i grounded the smooth stepper card directly to the ground-connector in the 230volt wall outlet.
used a shorter usb cable and i removed the metal "cover" at the type B end, the end at the USS.
i reconnected the breakout board.

i will try a laptop but doesn't anyone have any other ideas?

[Christian Knüll]

Hi,

well, a line filter at the plasma cutter and / or a USB isolator is likely to solve the issue - but before you spend money on those I'd first try a laptop to see if the problem is really caused by the mains line.

Shielding is usually no big issue with USB unless the cable is extremely cheap.
The only other thing is cable length - but you already replaced it with a shorter one.


Christian

[rowbare]

My experience with the USB SS is that it is pretty sensitive to noise. Mine would fault when my spindle motor would start to load down a bit. Check the SS forums. There are probably some solutions there.

It looks as though you are powering from the USB, you might want to try powering it from an external supply.

bob

[da_Rayman]

Right now, both the USS and the breakout board is being powered from usb cables. Before i started to fiddle with everything both cards was powered with a small 5V 1A unit, smiliar to this.

but i'm gonna reconnect it and try it again, along with a laptop.

[neilw20]

You have to isolated where the noise is getting to the laptop?
Have you run it on batteries with the ONLY connection to the laptop being the USB cable, as suggested in earlier post.
Until you physically isolate where the noise comes from, and it is USB related because the FDT errors make that clear,
(though there may be other causes - not just one) you have not identified WHERE the noise gets in.

[da_Rayman]

I tried some new things today. I connected the plasma machine to another outlet. That outlet is in another section of the power grid and is only connected to the same switchgear, maybe 150 meters away. I also tried running the computer and the servo drivers on the separate outlet (with the plasma on the regular outlet) with little difference. Now i can light up the arc for about 5-7 seconds before the error message comes up on the screen. I tried the usb cable that i have removed the shielding from but with that one, the error message came up instantly when i lit the arc.
I'm still gonna try with a laptop, as soon as i can get one but this is starting to get ridiculous.

[Hog]

I spend weeks trying to solve this problem spending money on every solution suggested on the forums - and there are a lot! $600 later and still no go.

In the end the only thing that solved it was getting an ESS. The ESS is so much more resistant to the effects of noise and the USS so sensitive! You might want to get a ESS straight away and save yourself a lot of pain.

Hamish

[da_Rayman]

That's a bummer. kind of annoying that shops like cnc4pc.com doesn't write that as a warning under the item description.

All this wasted time cost alot of money so now i cant afford a new card, so all i can do now is run my servos on the parallel port at 1/10th encoder resolution and 2/3 max speed...

Anyways thanx for the help everyone.

[Hog]

Hi da_Rayman,

Yes, I was incredibly dissapointed also when I discovered it was the board itself causing issues - I got my motion control box from an experienced, reputable seller.

The thing is, there are lots of USS implementations around (1000's I think) that work perfectly well. Why it is that a few people like you and I just seem to have issues in terms of our setups is a bit of a mystery.

I am in a relatively low EMI environment and implmented the most stringent EMI protection (shielding EVERYWHERE, short leads, eqi-grounding, faraday cages, ferrite cores, seperate power supplies, and all the other board-specific suggestions too!) with no joy.

I hope you can get something sorted out in time.

All the best,
Hamish

[Larken]

Da_ray, Are you just testing it with it all over your bench, or is it properly mounted in a grounded cabinet. Because all those wires are just a big antenna !

You have to fully opto isolate any USB controller.

[davidtye23]

So far, so good.

[neilw20]

2" of wire in the wrong place is an antenna. My friend was rather surprised when I made him move one wire that far.
ANY current going through a shield can cause havoc, with the supposedly shielded wires.
And shields won't cancel induced voltage from magnetic fields. You need to cancel magnetically induced voltages.
Any wiring over 3m (10') is likely to need magnetic cancellation techniques like RS244 balanced lines.
Fiber optic wins, hands down.

[Al_The_Man]

Good reading. With CNC in mind.
Al.