I figured I’d make a message chain for the various things we’ve done to troubleshoot the CNC’s USB disconnection issue to reduce duplication of effort and replication of failure. Please @ anyone else who has been thinking on this.
Right now, the problem is when the controller is disconnect (digitally) from the computer, we have to unplug and replug the USB cord to the CNC. That instantly fixes the problem. I have replicated this lots of times. The main way it does this is if you have to close Carbide Motion. When you restart the program, the computer can’t find the controller. Doug tells me that the same problem occurs when he uses his personal laptop, which indicates the problem is on the controller side.
Things to try:
I thought it might be related to USB power-saving, but I finally borrowed admin rights to the machine and could not find anything to support that theory.
It was suggested by multiple people that it could be related to EMI. Seems far-fetched to me, but easy enough to address. VFDs are famous for creating lots of EMI. Per the internets, the easiest way to address this is to add a filter and/or ferrite choke to the power lines. I don’t have the right size filter or choke for this, but I do have a bunch of smaller beads (the little clip on bumps). I am going to apply them liberally on both sides of the VFD’s power setup. We should also consider moving it further away from the computer and the controller. While I’m at it, I will also choke the hell out of all the damn wires. It can’t hurt. If I can find a proper sized toroid I’ll try that too.
If anyone is feeling really EE, we could borrow the oscilloscope or signal generator and try to prove or disprove this. We might need to borrow someone who knows how to use them, though. I get the theory, but I’ve got no experience with the machines.
The Rigols well do. Recommend the BK sig gens.
Not sure what the sig gen will do for you.
It will not generate a usb signal.
To bad you can’t modify the code in the pgm
Comm fail, reset usb driver, send query to device, etc.
Has anyone sent a detailed report to Carbide? Only us w/ this problem?
Is the comm fail at the pc end or the device end? Given that this problem has occurred w/ 2 pcs, I suspect it’s on the device end. Something is getting reset when you unplug / replug the usb line.
Thanks for working on this. Sorry to be dense, but which woodshop CNC machine are you talking about? I saw a mention of Carbide Motion so I could hazard a guess …. But just curious.
We’ve seen an uptick in crashes reported in Carbide Motion due to video drivers. In the past, this was due to outdated drivers, but now we’re seeing the problem even with up-to-date drivers. The real problem for us is that the crashes are occurring in third party libraries that we depend on, but didn’t write. The only way out of this seems to be updating those libraries, which is a big change.
We spent some time on this, and the initial testing has been more trouble-free than we expected. So, we’ve decided to push out the update to our beta download page to get more eyes on it. If you’d like to try it out, you can grab it here:
As far as I know, it has not been reported to carbide.
So far we have two computers experiencing the issue. I also can’t say the EMI is likely the cause, but it can do weird things and is relatively easy to address compared to changing out the controller board, so I figure why not.
Someone should contact tech support, but I’ve go no idea who that should be.
Doug, the video drivers issue might be related, but it also raises the question about whether Carbide Motion is being updated? These days I assume it is automatic, but is it?
I expect Carbide is experiencing the cost of success. The more equipment they sell, the more unique systems will need to be supported.
If go back to the basics, we need to profile the different video libraries in use and determine which reference chip sets they support on paper. We will need to determine the exact computer chip set that actually works in the real world.
Cheap computers are cheap for a reason, Omit a chip here and use a chip that barely hits spec there. It all adds up.
We all remember DLL hell back in the 90’s. Welcome home.
Welcome to the world of the system integrator. I expect Dell will be in the running.
The controller as it sits on the tote under the table is plugged into the UPS, in a Surge receptacle and not the Battery Backup receptacle. There is a reset button on the controller, and also on the UPS as well.
As to the other components of the machine that require 120v line voltage: The chiller is plugged into the Surge side of the UPS; the computer and monitor are plugged into the Battery Backup side.
I appreciate the new ferrite thingys on both ends of the cable from the computer to the controller, so thanks to whoever put them on. However, I had to unplug the cable several times from both ends to establish the connection. I’d say the ferrite thingys aren’t pointing to the issue(s)