Duet 3 6HC - Short to ground issue.

[dc42]

@TDK said in Duet 3 6HC - Short to ground issue.:

Do you know what the issue with these boards is? Wrong current sense resistor? Bad FETs? I would like to debug the board I have but don't want to void the replacement warranty.

It's looking like a bad batch of diodes, which should have been picked up by the ATE, but unfortunately the test procedure was abbreviated without my authorization. We expect to have a definite answer by Monday.

[If Marvin is listening, it looks like we are going to be replacing a lot of diodes!]

[elmoret]

Which diodes, out of curiosity? I'm not seeing anything relevant on the schematic, could be blind though.

[dc42]

@elmoret said in Duet 3 6HC - Short to ground issue.:

Which diodes, out of curiosity? I'm not seeing anything relevant on the schematic, could be blind though.

Looks like the v1.01 MB6HC schematic is not yet on github.

[elmoret]

Didn't even know there was a v1.01!

[dc42]

I've just added the v1.01 schematic, https://github.com/Duet3D/Duet3-Mainboard-6HC/blob/master/Duet3_Mainboard_v1.01/Duet3_MB_Schematic_v1.01.pdf. The diodes concerned are in parallel with the driver output mosfet gate resistors.

[TDK]

I probed the 8 diodes on my non-working driver 1. All 8 show a diode drop of 0.225-0.230 V.

All 8 on the working driver 0 show the exact same reading.

Just a data point.

[A Former User]

Ouch, been there done that .. feeling the pain!

Weller WMRT<3

[dc42]

@TDK said in Duet 3 6HC - Short to ground issue.:

I probed the 8 diodes on my non-working driver 1. All 8 show a diode drop of 0.225-0.230 V.

All 8 on the working driver 0 show the exact same reading.

Just a data point.

Thanks for the data point. Can you try probing the resistors that they are connected in parallel with instead? I have a suspicion that the diodes may not be making contact with some of the PCB pads due to oxidation on the leads. It takes only 1 of 16 pads to not make contact to get the S2G error.

BTW I found that none of my multimeters can measure the diode because of the 150R resistor in parallel. It seems that my multimeters don't supply enough current for this test. So when I receive a faulty board later today I will test them using a 1.5V battery in series with a 100R resistor and a multimeter set to read mA.

[dc42]

I've confirmed that the problem is caused by the soldering between the 6-pin dual diode packages and the board being bad. Here is an example.

The front left pin is not soldered down.

This only seems to be a problem when the diodes have manufacturing code KAXS4 engraved on them. I believe that the pins on one or more reels of diodes with this manufacturing code were oxidised or otherwise contaminated, so that proper soldered joints were not always obtained.

[tony73]

@dc42

Hello everyone ! duet 3 6ch ordered and arrived about 3 days ago to this code (WD41xxx) and diodes (kaxs4) but are all defective or will some be good? i'm only trying movement without printing. what kind of problem should i see? thank you!

[elmoret]

You would see a short to ground error, and the affected motor not moving.

[tony73]

@elmoret
hi in case of future problems will they be under warranty for how long? or do they have to replace them? what will they do?

[elmoret]

Yes of course, Duet's warranty information is publicly posted: https://www.duet3d.com/warranty

[TDK]

@dc42 said in Duet 3 6HC - Short to ground issue.:

I've confirmed that the problem is caused by the soldering between the 6-pin dual diode packages and the board being bad.

I can confirm this issue on my board as well. Probing across the resistor gives different results than across the diode pins.

Interestingly, it looks like one of the diode packages on my board was reworked. Maybe the assembly house spotted the issue during inspection and tried to fix in some cases (but clearly not all).

[dc42]

@elmoret said in Duet 3 6HC - Short to ground issue.:

You would see a short to ground error, and the affected motor not moving.

To clarify: if any of the 4 diode packages associated with each driver is not properly soldered, then if you are using 24V power, you will get a short-to-ground report and the motor concerned will vibrate or not move at all. But if you use 12V power, you will most likely not get a short-to-ground report and the motor will move normally.

At this stage we do not know whether all boards with KAXS4 marked diodes are likely to be affected, however it does appear that the issue affects more than a single reel of diodes.

[tony73]

@dc42

Hello everyone! as already said my duet 3 6ch card arrived a few days ago and one of those interested cod. (WD41xxx) diodes (kaxs4) I have tried all 24 volt drivers for now they all work no errors!

[T3P3Tony]

Hi Tony, thanks for the info. That aligns with what we are seeing, not all diodes soldered badly.

What is your board serial number?

[tony73]

@T3P3Tony

Hello ! and this right?

[Erika de la Cruz]

I received my Duet 3 last week; I had time to install it until Monday; unfortunately, Driver 0 and Driver 1 are not working, motors vibrate, and in the Duet Web Control there is a message saying "Short-to-ground reported by drivers #(number of driver)". I tested the same motors/wires in drivers 2-5, and they work perfectly, my system has dual Z and X-axis, so I really need the 6 driver channels. The serial number is WD41385.

[tony73]

@Erika-de-la-Cruz

hello are you powering the card at 24volt?