I live in East Texas where out humidity if usually around 80%. However today a cold front came through and dropped the humidity to the point that I am getting shocked opening the door. The colder it gets the worse static is. After the freezing point the air is very dry and conditions for static perfect. So there are a lot of posts on the forum about grounding your dust collection, router and so on so look that up and if you are not grounded then get things grounded.
When you have a moving spinning router bit cutting and air moving from the dust collection static is inevitable. So depending on where you live it becomes increasingly important if you live where humidity is low. If you live at the beach it is less important but not fool proof that you wont get static.
Try running an ‘air-job’ both with and without your dust collection running.
If it happens more with the dust collection on, then almost certainly the problem is static. A good first step to fixing that is to use a better dust collection hose - either one specifically designed to be anti-static, or one with a wire support that can be grounded.
My thoughts as everyone has an opinion on what works best for them This is just some thoughts of things to look at. I am not saying you have to do it all. You can pick what works best in your environment. So I am going to lay out a whole bunch of stuff, Every system solution can be unique and sometime it take a couple of things.
The noise is coming from some where and getting into the wrong place, your job for your system is to find it and reduce it.
The Diagnoses:
In your case it sounds like the CPU is getting wacked by EMI. I suggest this as resetting the controller rests the problem. If you had reset the computer and it reconnects, this would indicate the controller is still running and listening to the USB port. but it is not.
The CPUs job in the controller is to follow a list of instructions that it reads out of memory. Due to an electrical spike, glitch or EMI noise, the cpu gets lost completely or it gets a bad read of data and does something something strange. This is very common. It has sensitive components working in a noisy EMI environment.
So how do we control it.
We find the source/sources and reduce or eliminate the noise. It can come from the Router, the Vacuum or any thing else in your shop that makes sparky noise, including some lights, bad motor brushes… First and easiest, Make sure the AC lines are away from the sensitive DC lines. My Vacuum’s 120vac line was running by the Power Supply of the cnc machine and the AC was being coupled into the P/S and then up the power line to the cnc controller. This glitches the memory in the machines controller.
Most vacuums make a lot of EMI noise. Make sure sources of noise (other machines) are grounded good so the EMI noise goes down that ground and doesn’t get out Try to contain the noise at the source. Some will always get out.
We can also choke / reduce the noise from getting into the cnc controller. I added ferrite cores on data and power lines. For the controller’s P/S line, take the excess cable and loop it around in a coil as close as you can to the controller, this creates a inductive choke that attempts to block noise (higher resistance to noise passing) . Better yet, If you have a larger split ferrite core, loop the P/S cable through it 3 or 4 turns. It makes a better choke If you do this, I suggest you do item 3 below as you also choked the machines ground. We don’t want noise to flow into the controller however if it does get in, we also want it to go out somewhere.
Give the EMI noise some where better to go. The controller has lots of wire coming out of it, we need the noise to find a better path then these. Ground as much of the machine as possible. The X gantry is not grounded. Run a 16 AWG stranded ground wire from the X gantry through the drag chain and down to the Y rail where it is grounded. Make sure the Y rails are grounded properly. I suggest adding a separate Ground to the Y rail to the same ground you have the P/S plugs into, as we don’t want to create a bad ground loop. The little wire in the P/S cable may not be thick enough to shunt out a lot of EMI noise. EMI likes the path of least resistance. A nice thick ish ground wire provides a better chance of EMI going where we want and not where we don’t.
If applicable, ground your vacuum hose to shunt static build up to ground. Do not ground it to the CNC machine. Take it back to where the vacuum is plugged in. Do not attach it to the router or any where else on the cnc machine, we want the static to go way from the cnc machine and we don’t want to create a potential ground loop.
I have heard different theories on Ground Loops. Yes Ground loops can be bad. And yes we do everything to avoid them. Basically this is ground lines ran in such a way that it can get looped back on its self. Over longer distance grounds can build up separate ground potential differences. Joining these together causes unwanted current to flow on the ground lines.
Some say plug your controller into a separate isolated receptacle. however … the router is mounted to the machine and has a separate ground in its power cord. Mind you, it is held in place with oiled bearings so is the ground connected to the machines chassis ground ??? If it is, will it create a ground loop with the controller’s ground that is also connected to the chassis . Hmm. Grounds should run like a tree, always splitting outwards and never joining back to the tree.
However this is a system with multiple power plugs and multiple grounds
My advice is if you have multiple grounds on a device, such as the Router and the Controller P/S, or they have a potential to come in contact; Keep the paths as short as possible. IE Share the same receptacle. This reduces the the chances to produce large ground potential build up. All other grounds should be kept separate and taken back to the common branch as to not pass back through other devices. Give the noise a nice easy path to be shunted to ground. The Vacuum is potentially separate and can plugin to another circuit, spreading the load. I question if the USB of the CNC controller is optically separated, therefore I plug my Computer is plugged it the same receptacle so both ends of the USB are on the same ground potential.
Once in the past few years, I had a rash of disconnect issues and traced them to intermittent USB detection/connection with the PC. Make sure your power options do not allow USB controllers to disconnect when the system is idle. Try a different USB port. If your control PC is mounted to a common structure, vibration can cause issues as well.
That is great advice. I agree the problems can be any where So a good logical flow of trouble shooting is required.
For disconnects
I would reboot the computer. Did it “reset” the problem. if yes it may be a computer issue, and it indicates the controller is still alive and well.
If it does not reset the problem, reboot the controller now. Did that reset it. if yes then it is “starting” to look like we need to look around the controller, however this still may not be the case as the Controller can help re-establish the USB connection. we are not sure yet.
So far we reset the Computer and then the Controller. The next time it happens, I would reboot the controller first. If this resets the problem, We now have a good idea of what side we need to look at. If it does not and we have to reset both, it may be pointing somewhere else.
The more this can be made repeatable, the more confident we are in narrowing it down.
We are trying to establish if the controller is hung up.
Side note: I know we have lights on the board. I wounded if there is a heart beat light. Mostly there is just a Power light and that does not tell us much. The heart beat is turned on and off by the CPU. If it not beating, then the CPU is hung. Just thinking or suggesting out loud
We still may need to check other things like cables. There are a lot of crappy USB cables out there. Simply unplugging and plugging these in or changing them live might suggest the problem. If this fixes it and you didn’t reset the computer or the controller, well at least we know the controller is not hung.
IMHO, a hung controller is a very common thing in a noisy environment, the question is why did it hang. Most of the time that is an EMI glitch that hit the controller.
While EMI is a very common cause, I’d agree with Shawn (Selectronic) with regard to checking the USB connector - both ends. If there is movement/vibration at those USB-B connectors, they are very prone to intermittent connection problems. I’ve had multiple encounters with this on motorized telescope mounts: even though there is no significant stress on the wires the slight flexion of the plug in the connector gradually worsens the connection. Obviously, dirt or corrosion can too.
Checking with a new USB cable is a simple and fairly cheap exercise.
Okay I have taken a moment to look over the entire machine and cannot find any obvious grounding issues.
The USB cable has ferrites, none of the data cables are running next to air hoses or AC lines, also the circuit is properly grounded to the wall outlet.
Can anyone please take a look at the pictures of the setup and recommend and further action to mitigate this freezing possibly caused by grounding or EMI?
Kevin
I am embarrassed to show these pictures of my dirty shop. Tuesday is cleaning day…
I ran grounds to everything. Case in point,the “Z” is powder coated, no way to get ground properly to dissipate static. I grounded the “Y” also as you will see.
Anywhere there was powder coating I cleaned a bit of it off to ensure a good ground. Oh and it is the grey wire…
I bought an electrical plug, some cheap ( maybe 24 Ga) uncoated copper wire., and a grounding bar. I ran the wire inside my vacuum hose, out, and back to the grounding bar. I ran a strand of wire to the ground on the plug and plugged it into the same circuit as my machine. I then used a multimeter to ascertain other parts of the machine were grounded. They were not. I ran wire to ground everything except for the trim router. I used to not be able to cut plastic. Now I can. The ground bar is the key to avoiding ground loops. Try it. It’s a very cheap solution.
I agree here. ground your gantry with a stranded wire 18 awg ish, down the drag chain, back down to the ground point on the Y rail. (that green wire coming out of the controller and attaching to the Y rail frame).
However I would try this first as it is easy. Move the controller’s P/S away from all that AC electrical. ( Mine was a single Vacuum AC wire passing by the Controllers Power Supply ) Yours appear to be right on top of a bunch of AC power bars next to the very EMI noisy vacuum. Simply move it and its cable way from all that AC stuff.
Back to item 1 above. @Zman has done a great job of grounding everything. I would add that any noise that is picked up by all this grounding has to go somewhere, Right now it goes to that small green wire. back to the controller and then down a small power cable that has a choke on it, through the P/S and finally electrical ground . That is not a great path back to electrical ground. So you could add a better ground path to the same point on the Y rail. Redirecting the noise around the controller to a better ground. A 16 AWG wire running from the Y rail ground point to the same electrical ground that the P/S is plugged into.
You are beefing up the ground so the EMI noise has a nice easily path around the controller back to electrical ground, not through the controller.
Also the job of the choke on the DC cable is to be a high resistance for high frequency current changes ( Noise ). We use chokes to prevent noise from passing. But this is the only ground path back to electrical ground. We need noise to have a easy path back to electrical ground. not through a choke and a thin wire. So just provided a better one.
The reason for the choke is to try to reduce noise coming from the power supply to the controller. So if we do this then we need to add a non restrictive ground path back to electrical ground.
I did not show everything. It all goes to ground at the Quad outlet. I also dropped a few ground rods, one at every corner of my shop. Have a single lead from the machine to a rod. Overkill, maybe.
This is just some thoughts of things to look at. I am not saying you have to do it all. You can pick what works best in your environment. So I am going to lay out a whole bunch of stuff, Every system solution can be unique and sometime it take a couple of things.
So a good logical flow of trouble shooting is required.
