Heat related issue with homing of Z-axis

[AZieger]

We have a CNC router table made by NEWCNC.  For several years now we have had an issue related to control cabinet temperature and the machines ability to successfully "see" the z-home switch.  We re using Mach 3 and I am viewing the diagnostic window to see whether or not the green light is illuminated on the z-home switch.  The switch itself has a light on it, as does the input on the PLC on the head of the machine, and both of those will light up, yet the light on in the software will flicker or not turn on at all.  This issue goes away completely if the machine is cold.  In fact, I can direct compressed air at the transformer on the PMDX-126 board for a few seconds to cool it down, and the light on the screen will come back on for a couple of minutes, until it warms up again.  So I am thinking that the transformer (part circled in picture) is getting too warm and causing the voltage to drop.  Is this a known issue? 

I need to investigate the cabinet cooling system as I do not believe it is functioning correctly and may be the main reason why this issue began.  But even when the cabinet is ambient temp, on hot days (90F+) we can see the issue happen.  I am planning to add a heat sink to the transformer and direct a CPU fan at it, but is there something else I should be looking into?

[Steve Stallings]

It is doubtful that the power on the PMDX-126 has a problem. The
way to verify this is to measure the voltage at the 5 volt points
on the screw terminal blocks. The minimum acceptable voltage
would be 4.5 volts.

It is more likely that the input circuit of the PMDX-126 is falsing
due to leakage current from the proximity sensor. As the PMDX-126
warms up, its trigger point may shift. Try adding a 1K ohm resistor
from the input signal terminal to the +5V terminal of the same
terminal block. Yes, I said +5V not ground, internally the PMDX-126
measures things against the +5V and looks for signals below +5V
for sensors that pull low (switches and NPN proximity style) or
above +5V for sensors that pull high (PNP sensors powered from
24 volts). The resistor bleeds the leakage current into +5V so that
it disturbs the signal less.

[AZieger]

Thanks for the reply Steve.  I had a little time to troubleshoot today and here is what I found:

The voltage between GND and +5V on J11 and J12 is indeed 5.0 volts.  However the voltage between pin E (where the z-home input is wired) and GND is about 4.2V.  This is the same for the X, Y and B home switches (pins 11, 12, and F, respectively).  However X, Y and B home perfectly every time. 

I tired the 1k resistor between J11 +5V and J11 pin E.  This brought the voltage up to 5.0V, but still the same issue.  I verified that the voltage gets pulled to 0 when the switch closes.

I will also add this: the LED for pin E always works.  It will illuminate and stay steady, while the status light on mach3 will flicker an eventually die out as it heats up.  So it would appear that the PMDX-126 is seeing the switch, but somewhere in the line of communication between it and the PC, the message is lost. 

And again, cooling of the transformer (and the surrounding area, kinda hard not to when you are blowing air at it) will cause the light in Mach3 to eventually flicker and then reappear, until the cooling is stopped at which point it flickers and goes dark again.

Finally, I decided to measure the mains voltage and found it to be 260V.  Is this our issue?  I see that the voltage switch is on 230V, and the board is rated for 220-240V.  What would 260V do?  Is this causing things to overheat?

Thanks!!

[AZieger]

Alright, I have narrowed it down further.  It is the voltage regulator near the secondary status inputs that is overheating (circled in pic).  I am measuring 13.4 volts going into it, and 5.0 volts coming out.  The other regulator has 12v going in and 5v coming out.  Going to get a thermocouple on it tomorrow, but it is definitely the culprit.  Even a tiny amount of airflow on this component will cool it enough for the z-home to be seen.  I ordered another PMDX-126 last week in case all else failed, but first I think I am going to try a heat sink on this little guy and a fairly strong fan pointed at it.