View Full Version : Another "infant mortality" issue...

04-27-2011, 01:20 PM
Had a limit switch go out on me. First happened a few weeks ago. After running the machine, and having it just sit there afterwards for a couple hours, the machine went in to alarm with the red light on. At the time I was very busy and didn't have time to troubleshoot. I just shut the machine off. I powered it back on the next day, and it was still in alarm. I left town on business with the machine "down". Got back this past weekend, and the machine powered up fine. I used it to machine the splash guard that I posted in the other thread. I started making a video to show the splash guard, and the machine alarmed again - perfect timing to video the issue!

I called Phil to troubleshoot - and he didn't answer. As I was listening to his outgoing voicemail message, he called me back. Fastest call back I have ever had from anyone! We debugged the issue and determined it was the X-- limit switch (the switch on the left side of the table, that triggers when the table is fully travelled to the right). I dried off the switch, but it was still in alarm. Phil told me how to get the machine functional again, with the switch disabled. He is sending out a new switch under warranty.

I will post up if this fixes the issue. I have only machined metal on the machine twice so far, and one of those was aluminum. Phil said a chip could cause a false trigger, but with so little time on the machine with metal, it appears to be a faulty switch.

Happy with the progress so far, and hope this eliminates the problem once and for all...

04-27-2011, 02:00 PM
Wow! I had the exact same issue happen on my machine only it was my X++ limit switch. The one on the far right next to the home switch. Same thing - failed in the closed position. Phil also has a new one in route to me. Prior to that I had been running coolant quite a bit - wonder if it has to do with leakage into the switch.

Other than that machine has been working perfectly.

04-27-2011, 03:18 PM
Both times mine errored out I was running coolant. I have really only cut wood up to then, so it never manifested the error. I suspect it does have a leak in the case, as little as I have run metal I can't believe it was an errant metal chip...

Maybe I will epoxy coat the new one to protect it a bit more before installing...

04-27-2011, 04:50 PM
Those are pricey switches - $80 in single piece quantities. "supposed" to be IP67 rated - dust proof and submersion up to 1 meter under water. Nice switches, surprised they failed.

I may coat mine as well when the new one comes in only I will use a flexible sealant around where cable enters housing.

04-27-2011, 05:00 PM
Caulk would probably be better than epoxy...

04-27-2011, 05:31 PM
Those switches are one of the machines weak points. Homing the machine with them was a mistake. They are not accurate enough.

04-27-2011, 07:00 PM
Have you made any changes? Do you have any suggestions for an upgrade? It's just a standard proximity switch, which is pretty universal for CNC machines - what would you do differently?

04-28-2011, 05:51 PM
Well, I haven't made any changes. I just don't depend on the switches to home accurately. I always make sure I can pick the part up again if there is a power failure or I shut the machine off.

Once everything is picked up you can depend on the machine to repeat.

I like the switches on the Tormach better but I still make sure they are repeating and usually do with in .001".

04-28-2011, 10:41 PM
A few notes that are pertinent to the discussion.

- We use identical switches for Home and limits on all 2009+ 1610L machines. They are Omron industrial proximity switches. They cost about $57 in production quantities, and are solid state, sealed, waterproof IP rated units. They are complicated devices however, and can fail. When they fail, they are designed to fail "safe". These switches are fairly generic in industrial CNC machines.

- As configured on our machines, these switches are intended as Non - precision position references for programming and positioning (not required for operation as detailed below).

- The commonly (in this context/forum, non-industrial) used mechanical switches are far less reliable, and much cheaper ($2). The key word in the below note is "usually". They are not industrially rated (generally), and have a much shorter and less deterministic life. When they fail, they generally fail in an unsafe manner and result in a "crash". Fine for a low power machine that is not used in an industrial setting.

- Neither type of switch should be used for "precision" reference (in our verbiage precision in this application would mean not detracting from machine precision) - this requires a switch in this case good to ~0.0001. There are probing systems available in the industry for this reason.

- Remember that Homing an command position (Open loop) machine is for fixture or code reference only and is not required for operation. It may materially detract from the precision achieved depending on the metrics, devices, and code.

- The reason (generally) for homing a interpreted position (single loop closed) machine is to correct for measurement (accumulated encoder) errors and requires a precision reference for operation.

Precision, reliable, automated reference is not cheap. Yes, these systems save time. No, an automated system will not make any better parts than properly manually indicating a part/tool/fixture.

Here's the take away: For automated, precision, position reference, an probing or precision reference system should be used. Next best is to use a precision manual reference.

If you want to use a non-precision rated switch (reliable or not) for precision reference, do so only when making informed decisions about the performance/cost/liability trade-offs.

Work safe.

Mikini Mechatronics, LLC