New Blackbox freezes during job. Need Assistance - Please

To Enable USB Selective Suspend in Windows 10, do the following. Open Control Panel. Go to Control Panel\Hardware and Sound\Power Options. On the right, click the link Change plan settings. On the next page, click the link "Change advanced power settings". In the next window, expand USB Settings -> USB Selective suspend.



Alex.

 

Have you tried using another G-code sender? Try your original UGS, install bCNC, or something similar. CONTROL is just another g-code sender to the (Arduino/grbl-based) BlackBox, so you're not limited to it.

The USB tech specs limit high speed devices to 16ft and low speed devices (which I imagine the 115k baud of grbl to be) to 9ft. There's a real chance a "?" or a "ok" was lost somewhere along the line and things just... Stopped.

Consider a powered repeater USB hub halfway along your cable run, if it's possible.

 

This is the biggie. VFD noise is kinda more of a physics problem, less an electrical engineering problem. Common sense goes out the window when you start getting into high frequencies, you have to know how it's going to behave. Transmission reflections and power transfer, skin effect, all that stuff. The winding-case capacitive coupling one was an interesting one I hadn't thought of with @Batcrave's build though- hard to mitigate the dielectric constant of a fixed air gap, unless maybe you fill it with a high-dielectric gas, maybe.

 

Addendum! Earlier I was thinking "y'know, just putting a low pass filter on the VFD output super close to the unit where it can be inside the shielding cabinet, would eliminate like 80% of the problems with a VFD by dropping the crazy HF noise down to like, regular hum". Turns out... Someone already invented that. They're called "sine wave filters", because reasons. Not sure about pricing, but figured a heads up would be in order.

 

That output file looks like your sender is losing the "ok"s and instead picking up part of the garbled "?" responses. So far, still looking like it's the USB cable.

An EMI filter for AC is, I would believe, a bandpass filter set at 50-60Hz. That protects the power input side from supply issues and the supply from noise injection. I'm describing something that converts the extremely high speed PWM output (with its associated conversion switching noise) to something very close to a true sine wave. Looks like they're in the $500+ range: Siemens | 6SL3202-0AE21-4SA0 | Used & Repaired | Sine-Wave Filter - not something anyone on this forum has, as far as I'm aware, just something I found whilst thinking about it (@Batcrave, I might have found something to use those microwave capacitors for...)

With the issue persisting without the VFD turned on, I'm still currently convinced it's the USB cable. This length of run is more suited to an Ethernet-based control system like a (non-grbl) SmoothStepper.

 

What that output file looks like is exactly what I spent the past couple months battling. The description in the initial post about the not-quite-pausing matches up, too. The only thing that doesn't match up is that "without the VFD turned on" bit - since the VFD on/spindle running was the only thing that was constant in my case.

One kinda out-there suggestion, though... is there any chance you have an LED ring (or anything else, for that matter) plugged into the same power supply as your BlackBox? While I never managed to pin it down as the cause in my case (and it certainly wasn't the only cause), it was definitely a significant contributing factor (not because there's anything wrong with the LED ring, mind you - it just seemed to be acting as an antenna for the noisy VFD/spindle, through the largely unprotected power input). When I'd unplug the ring, the problems would either go away or drastically diminish, and I eventually got things more-or-less-sorta-kinda under control with a whole mess of ferrites on the cable (well, a whole mess of ferrites and a separate routing, an EMI filter, shielded spindle cables, grounded steel enclosure, loads of profanity, and a fair bit of threatening with whatever power tools were on hand)

Again, though, it doesn't exactly fit when the spindle isn't running (Peter had speculated that the way the ring wrapped around the spindle played a large part in its antennaness).

One other thing I might be inclined to suggest is the AM radio test. Wandering around waving a battery-powered radio tuned to a dead AM station (I had the most luck down in the low end of the spectrum) can help identify any big emitters of RF noise in the environment that your system might be picking up. The VFD/spindle is usually the biggest source, and you'll hear some from the steppers too, but maybe there's something else that's been overlooked - maybe the rats in the walls have a microwave, or your neighbor's been mowing the lawn with an unlicensed nuclear accelerator.

Of course, finding the source of the source of the noise isn't the same thing as blocking it from interfering.

USB cables can definitely play the antenna role, too. As I've mentioned in a couple other threads, on a previous system the USB cable on the gamepad/jog controller would, despite ferrites on the not-even-especially-long cable, pick up enough VFD noise that at certain speeds it would cause the PC to bluescreen. I'm with the others in thinking the 12ft cable should at least temporarily be taken out of the picture, long enough to see if it makes a difference.

At $219 refurbed, I probably would've considered it, if I'd thought it would actually make the problems go away.

Great! if only I had some microwave capacitors. I thought the only major components I'd cannibalized were transformers and a magnetron or two (and a whole mess of powder-coated sheet metal that I'm sure there has to be some sort of use for)... but I do seem to have a couple caps sitting in the salvage bin.

-Bats
(think you could inventory the rest of my junk collection for me like that, too?)

 

This could actually be a problem too. I've been doing some reading and apparently double-grounded shielding ground loops are something of an electrical engineering old wives' tale and all shielded cables should absolutely be terminated at both ends.

Right now I think you're the foremost expert on dealing with VFD noise! Scary, eh?

As long as both it and the VFD are fully enclosed in a thick, grounded enclosure, with appropriate filtration on mains/signal wires and all the rest of the guff for direct interfacing, I don't see why it wouldn't work. 100kHz RF square wave + harmonic noise vs 50-400Hz sine wave? I think it'd be night and day.

 

Just don't say that too loudly or we'll suddenly have a dozen old engineer's wives in here screaming about how you're wrong.

Terrifying, frankly. The foremost expert on Things That Won't Help. Probably not something likely to inspire much hope in Rod, either.

Sure... but no one could see why any of a dozen other things wouldn't have worked either. And while it would've seemed almost affordable if it was guaranteed to solve the problem, it'd still look pretty steep for something that would probably help, after so many other probablies (and at least a few definitelies) had failed.

As frustrating as it is not to have some sort of solid closure, for the moment I'm just going to hope that the problem stays in remission.


-Bats
(..or that it can be beaten back into submission with another dozen ferrites. per wire.)