Sieg KX3 Controller Board Dying?

Hello all! I've had my KX3 since 2011. I had it installed by the late John Stevenson. I've been using it off and on all this time.

BTW, I switched to using a uc300eth in 2018, and now use a Surface tablet, instead of a museum piece PC. Much nicer!

I've just started another project that requires some CNC milling. However, the machine has started behaving erratically.

Input pin 10 is now permanently active. This is connected to the hardware estop, Disconnecting the button still shows an active pin 10. I've had to reverse the active low trigger to get out of the stop state, and I'm now running without a hardware estop. Not good. To note: for some reason, the actual wiring for the estop did not follow the circuit diagram in the manual. The two wires were plugged into P10 and Estop respectively. I would have thought the ground wire would have been plugged into AGND, not Estop?

The Z limits trigger randomly sometimes. Is there a Z limit switch hidden somewhere? I can't seem to locate it.

Regarding input pin 10, I'm wondering if the opto-pcoupler is at fault and whether it's worth replacing that. Or if I can get hold of a new Mach3 board, identified as SM11 in the circuit diagram? Anyone know if this is still available? Or maybe something of a similar spec, even.

Thanks!

Nikki

I can't help much with the E-stop problem apart from suggesting you check the actual e-stop knob if functioning correctly. I take it the UC300eth board is the one you think is playing up?

There are Home and 2 Limit switches on the Z axis but they are all optical proximity rather than the microswitches found on X&Y so could just be a build up of fine dust has found it's way in. You need to take the back cover off to get at them.

Edited By JasonB on 13/10/2020 08:00:28

Those don't look like optical switches Jason. They look like standard inductive proximity switches. Triggered by a something metal passing in fron of them. You can see the operating flag just below the front one in your photo. Light on the back is an indicator to show if switch is detecting something or not.

Thanks Matt, yes light goes out as they pass the bolt heads

Thanks for the replies. The location of the Z limit switch is very helpful. I'll have to do some gymnastics to get to the back of the machine. The machine is actually backed up against a garage wall with little room for manoeuvre. UK garages are shoes boxes!

Currently, there's nothing attached to the P10 screw terminal, yet it shows a high signal. I don't think that's normal. Probe is also showing a partially lit led, despite the wire connected to its terminal dangling free unconnected. Could be stray EMI. I just need to unscrew the wire to double check. I've got some 817 optocouplers on order now. I think I'll desolder it and compare it against a new one. It'll have to be desoldered in-situ, as I don't fancy disconnecting the board.

Hi Nicola,

Before you climb in the back of the machine with a soldering iron a couple of things to consider.

First is the long green connector really a plug and socket? On some of them the top part with the wires and screw terminals will unplug from the bottom half which is soldered to the board. The long black line on the connector in the photo looks promising.

Secondly do you have any spare opto inputs on the board? If so, can you move the e-stop to one of the spares (need to tell mach3 where you moved it).

Keith

Omg, it's actually a plug and socket! Plug comes away easily. That's going to make it a damn sight simpler to diagnose.

There's pin 15, which I'm currently using as the index pin. It's not critical. I sometimes like checking the real RPM. I can use a handheld tachometer for now. E-stop needs to take priority.

Regarding the UC300eth board. It doesn't replace the KX3 bob controller board. It takes over the responsibility of the PC to send the parallel port stepper control signals to the KX3 bob in a timely fashion. That means the PC can lazily send gcode in its own time, over ethernet, and the timing of that is not critical, only that it needs to keep the gcode buffer topped up on the UC300eth. If my understanding is correct.

Edited By Nicola Casali on 13/10/2020 23:36:30

The optocoupler appears to be fine. I tried a replacement anyway. P10 remains high and the LED is lit. Removing the presence of an optocoupler completely, extinguishes the LED. I'm not sure what's going on here. Using the Index pin, P15, as the estop seems to work.

I think I'll contact ArcEuro. This is where I purchased the machine in 2011. Maybe they can still obtain spares.

Thanks for your help.

Edited By Nicola Casali on 17/10/2020 15:22:26

Edited By Nicola Casali on 17/10/2020 15:23:02

Is there a pull up or pull down resistor on the board that has failed? If you have not tried it yet check the voltage on the inputs to the opto couplers when there is nothing connected, see if one is different from the others.

Martin C

Edited By Martin Connelly on 17/10/2020 15:29:37

This is the board without the optocoupler. I've checked the two closest resistors. They show correct resistance. The two pads of the left side of the optocoupler location show 4V. I could experiment and tack on a pulldown resistor of some high value.

Switching number 7 on the DIP switch to off seems to disable a pull-down resistor on pin 10, which I measured to be about 1.5M - 1.9M Ohms (can't remember exactly). The active light goes out, unless I press e-stop. Sorted? Seems to be behaving properly now.

This I discovered after I felt compelled to do some further investigations, upon realising a replacement board would cost £300.. In all 9 years I've had this machine, I have never touched the DIP switches. Not sure what's going on.

One guy posting on mycncuk in 2018 seemed to have an identical issue to mine on his KX1. Not sure how he resolved it. I'll update if I discover anything else or the fix stops working. Need to put this machine through a test and see if the z stops are still temperamental.

Thanks!