The mcu operates at 3.3V, the gate of power mosfet (the control input) operates up to 20V, so the mcu cant drive it because it doesnt have the voltage to do so. Also the gate is capacitive, so when you switch it you can have brief but high current flow beyond what the mcu can provide and emi spikes that can trip up the mcu or clocking to the mcu. So there is another switching circuit (q6 and maybe q5) that is capable of driving the mosfet. The input side of the switching circuit is driveable by the mcu, the output is sufficient to drive the power mosfet. So if you find the power mosfet is doing what it's supposed to (it's being told to be on at the gate of the mosfet), then you want to check whether the switching circuit that is driving it to be on is function correctly or not. If the switching circuit is functioning correctly then likely the mcu is shot, and if the switching circuit is not function correctly then it's time to debug that (likely one of the transistors in the switching circuit at that point).
You wont like this... but you might consider maybe sending the whole machine in; reason why, uncontrolled heating of the extruder is a fire hazard and if not 100% confident in the repair you can't be confident about not having a future fire. I'd spend time on the chassis first, to try and locate the fault blowing out the motherboards (since you have other machines, you can use them for reference when probing around, likely doing resistive measurements is enough). If not able to find the fault, I'd RMA the machine. If the fault is found, then maybe diy repair on the mb, or send mb in to not worry about future.
So I opened it up today to carry out CPS's checks (and see if I could shoehorn a Smoothieboard or something in there), but I also unplugged and reconnected all the plugs in the top little pcb.
The LED mentioned wasn't on, so I flipped it back up and it wasn't getting hot.
I replaced the bottom cover so the board had its fan on it, and started it all up.
Normal. I'm printing now with it.
WTF??
It absolutely run off on me straight after changing the ribbon cable, cooled it down and retested several times, kept doing it.
And I've already tried 2 complete hot ends with all wiring.
I'm really thinking this is a simple wiring/earth/short something like that, that's just enough to wreak havoc.
I'm thinking of going through and re crimping the little plug pins that are on the hot end wiring, does anyone know what they are and where to get the crimping tool? I understand they can't be soldered?
A connection/wiring fault would be something that could make it through final inspection as it wouldn't show at the time of testing...the machine worked or it didn't at the time.
When this print finishes I'll have a look and check that the pins are pushed right into the plug housings, but I'd already checked that.
Could it really be something as simple as a pin crimped over a bit of insulation or something like that?
You could try and use d5 to help locate the problem, just systematically jiggle the wires and see if you can get it to flash doing that. Also the pcb, put a little bit of pressure (not a ton) on it at different points to see if there is something mechanically wrong with it. Connectors too.
You could try and use d5 to help locate the problem, just systematically jiggle the wires and see if you can get it to flash doing that. Also the pcb, put a little bit of pressure (not a ton) on it at different points to see if there is something mechanically wrong with it. Connectors too.
Thank you very much.
That looks to be a wire stripper only though. What sort of plugs are these?
And yes now that I know about the LED on the board I guess I can chase the fault that way.
Sorry posted the wrong link, I meant this https://www.pololu.com/product/1929 Might be able to find a cheaper on, make sure you get one for the correct gauge
thank you for your reply. I will check the cables first.
One thing I recognized is, that the blue led at the bottom of the heatbed is almost steady flashing. If I remember correctly it was sometimes off during printing in the past.
Unfortunately I have no machine to double check that.