Need help with unexpected/random stops on basic Parport/5 axis BOB setup

Lezard wrote: My friend suggested that the PC should be powered from inside the CNC cabinet, in particular because the Bob gets its 5v power from the PC USB.
I did not understand exactly why, but I trusted his judgment. What would you say ?

I never power a bob from a usb and not from a usb charger either as the smart ones control the charging current with hand shaking

Best to use a small 5V power supply

>Best to use a small 5V power supply

OK, I'll also try this as well. I should have some sort of small standard PSU used for appliances, could this be worth a try before I get a proper one ?

==============
I need to move now, will continue testing over the week-end and report here. Thank you all for your help )

Your friend is right, but in this case if all is wired properly it should not matter as the BOB is used for exactly that : Isolation.
Give it some more time on the outer socket, it looks like there is an issue with grounding, after all.

Hello again,

I've been able to work with the CNC for a couple hours, and did not see any unexpected stops.

I've left the workshop now, but left the machine ON, so that I'll be able to check it again tomorrow : if all is well, it should still be ON. I've seen (or rather heard the click on the Z axis motor brake) go OFF even when it was completely idle, with no spindle running, no moves etc...

Crossing my fingers at the moment, more tomorrow.

Thanks again, it's great to have experienced people to help !

Hello again,

Went to the shop this morning and was happy to see that the CNC was still ON. We might be on the way to something...

To push things further, I've launched a 2.5 hours .ngc file, with no stock and no mill, so I'll go back later to see whether the machine has completed it or not. Fingers crossed !

I must admit that it all sounds like black magic to me, if anybody can point me to resources that could help improve my understanding of these EMI, GND things, I'd be happy.

Thanks for taking the time to read me

Lezard wrote: I must admit that it all sounds like black magic to me, if anybody can point me to resources that could help improve my understanding of these EMI, GND things, I'd be happy.

It does seem like black magic, but it is not. As with god or love, it is hard to explain things that can not be seen or touched or played with.
Everything that uses electricity uses wires for that electricity to pass through, and while that is happening there are drops in voltage depending on the current flowing through them, that in turn will produce magnetic fields around those cables and that magnetic field will induce current to other nearby wires causing interference. The amount and frequency of those interference's will wary a lot, depending on a lot of factors, way to much to list them here. That is why it is a challenge to find the source of those problems in electronics, especially when having low level signals with high power equipment in the same machine.
Use of shielded cables and proper grounding is imperative for proper functioning of machines.
linuxcnc.org/docs/html/integrator/wiring.html

Hello again,

Just came back from the shop, the machine has gone through its big .ngc file without any hiccup, so it seems OK.

So, bottom line is : PC powered from dedicated socket inside the cabinet => trouble
PC powered from a wall socket outside the cabinet, nothing else changed => OK.
I'm going to run a reverse test tomorrow to confirm this.

I don't understand how/why grounding the PC to another ground (?) makes such a difference. They are all connected to the same electrical panel in the shop... ? :blink:

Hello,

I've tried to power the Bob with a simple 5V power supply, rather than the PC's USB, and power the PC from its dedicated power socket inside the CNC cabinet : same results, the machine goes OFF on its own, after some variable time.
I've noticed also that the PC screen is a bit unstable, with some sort of diagonal ripples going through the image.

With the PC powered from the outside wall socket, all seems well.

At least, I have now a setup that will work, but I'd like to get a better understanding of this. When building the cabinet, we have paid attention and tried to provide proper grounding. All the 'noisy' cables, from the vars to the servos, are double-shielded, with braids connected to the cabinet back plate through an alloy plate and purpose-made cover.

What can I do to try and isolate the source of the problems ?
Thank you for your help,
Laurent

Lezard wrote: I've noticed also that the PC screen is a bit unstable, with some sort of diagonal ripples going through the image.

That means there is something very wrong with the wiring/shielding/grounding, but at least you have something to test with.
Looks like there is something drawing quite some power through the ground on that machine.
Use a DVM and measure the voltage from the ground of the outside socket ( the one that has no issues) and the one inside, i would bet that you will have quite a difference, most probably around 60V AC.
You will have to get that sorted before starting to use that machine as any grounding mishap in the shop can cause some issues with the machine .

Hello,

I've made some measures today :
- voltage between GND of outside socket and cabinet chassis (resistance to cabinet GND arrival = 0) :
- with all equipment powered, _not_ enabled : 0 Vac
- same, enabled : readings between 110 and 440Vac (!!), from a multimeter set to 500Vac scale
- readings depend on which var is powered : I've tried to connect each one in turn to see if one might be the sole bad guy.

- I've also checked the resistance between various points supposed to be grounded inside the cabinet, the chassis, and the GND arrival in the cabinet : everything seems OK, except for a 5 Ohm reading between the common GND of the vars and the GND source itself. These are mounted on the same DIN rail, and are supposed to be connected through the rail.

- I will check again that the resistance between the outside socket GND and the cabinet is close to 0.

Any thoughts ?

Thank you for your help,
Laurent