The 2DM860H chinese controllers vs the rest?

[Kitwn]

So where's the photograph with the lid off? Is the failed device replaceable with a soldering iron?

[Doddy]

So where's the photograph with the lid off? Is the failed device replaceable with a soldering iron?

Photo?, okay...

General board



Replaced (failed) wire-ended fuse with 20x5 fuse clips and placed a F10A fuse in to test, no motor. Blew fuse



Difficult image to take, but shows at leasts 7 devices soldered to board and bonded to heatsink. At this stage, with two drivers on order, one for delivery tomorrow, my interest waned.

[Kitwn]

OK, take the point. But CMOS power transistors are a few pence each and it would be a shame to trash the whole unit if they're accessible for easy soldering.

[Doddy]

You're right, and maybe a bit of a tear down could be useful for the wider audience.

[Doddy]

Okay, so a bit of a tear-down of the 2DM860H.



Highlighted - 3 black screws that hold the board to the heatsink. 2 silver screws set into plastic top-hat isolators that secure the power electronics firmly to the heatsink.

Bottom left - 2 pin connector that is the AC/DC power input. This feeds through the fuse (normally a wire-ended... but what's there now is a bit of a hack attempt to repair. The big black circular thing is a NTC power thermistor designed to limit inrush current (it starts at high resistance but quickly drops as it warms up - likely used to allow the processor to start up and settle the power system before damaging anything.

Top left - I expect a small SMPS for the logic (untested). Left centre (left edge of board) - you can just see 4 pins for the power rectifier on the rear of the board Smoothing caps along the top and to the right of the board - all the logic. Thar be dragons.

Next pic, underside of the board, and a load of power devices...





The big thing at the top is the bridge rectifier. Four pins, which you can't see because of the thermal washer on it, but these are '-', '~', '~', '+", from left to right as orientated in the image.

The 8 TO220 packages are IRF640N N-Channel MOSFETs.

Parallel, and to the right of the top four of these are the stepper driver connections, labelled (top to bottom) "B-", "B+", "A-", "A+". If you look at the 8 MOSFETS, there's 4 pairs, vertically aligned. The left-most is driving the "B-" stepper; to the right of this drives the "B+" phase, then the "A-" and the "A+"

Each pair of MOSFETs are half-bridge configuration, and two pairs makes a standard H-Bridge for each stepper winding. The four pairs driving the 2 windings on a standard 4-wire stepper.



And a rough-order board layout...



Now, if, like me, you've blown one of the controllers up, chances are the problems are with these two H-Bridges, and the easiest method of diagnosing the problem is just measuring for shorts from pins to ground. Clearly, the bottom four MOSFETs will each have their Source terminals connected to ground (and the top fop have their Drains connected to V+)

I quickly tabulated a few measurements.

    Rg-d  Rg-s  Rg-0  Rd-0  Rs-0
    ----  ----  ----  ----  ----
M1: 1R    1R    4k    600k  2.6M
M2: 20k   20k   20k   600k  17R
M3: 2.6M  20k   3M    600k  6M
M4: 2.6M  20k   3M    600k  5M
M5: 20k   20k   20k   500k  0R
M6: 20k   2R    1R    18R   0R
M7: 2.6M  20k   20k   3.2M  0R
M8: 2.6M  20k   20k   3.2M  0R

Looking at that, I reckon MOSFETs M1 and M6 are blown. M2 is suspect, but I think the low resistance source-to-ground is via the internal short on M6 (M2 source, connected to M6 drain, which is near-short-to-ground).

The MOSFETs are through-hole devices, although I has a solder rework station I'll just crop each leg off the device, apply tension with some tweezers and heat each pin one at a time with a soldering iron to remove each pin. Solder-wick each side of the board clean, then remount and solder in from the reverse side of the board.

I do have one IRF620, but the Rds for this is too high and Id too low (5A); and a bunch of IRF3205's but the Vds for these is too low (50V). So, I'll order a pack of IRF640Ns, and replace at least M1 and M6. And report back on the robustness, or lack therein of the 2DM860H.

[Kitwn]

Doddy,
That's a brilliant job, thank you!

I know most people want to just plug-and-play with modules but when you see how simple it is to fix some of the faults for yourself and realise how cheap some of the components are it saddens me to see people just give up and trash expensive modules that can be repaired for very little. Most faults with this kind of device will be in the power supply or output devices, all of which tend to be the large components that cost very little and are easy to replace as you've demonstrated. RS components will send you a bag of 5 new transistors for less than the cost of a pint, including postage!

[Doddy]

A little unfair there - there's a long way to go before this works again, and I'm carrying cost and risk in repairing it - for some the downtime wouldn't warrant the cost of replacing with new. Also, to be absolutely honest, I don't think it's cost-effective for me to have spent trying to clear some form of metal debris from the fan supply solder pads for the last hour (finally resolved with a 0.4mm drill - which I managed to snap in the hole) - it was clear on inspection that the fan wiring was a bodge job with the assembler unable to poke the wire through the hole - silicone hides a whole multitude of sins on PCB assembly in china.

If it wasn't for this damned lock-down I'd throw the controller in the bin and be done with it, but being forced to spend time either with the wife or with a soldering iron, there's some curiosity value for trying to repair. But I do have some rather expensive tools to support that curiosity. Also for many in this game there's a lifetime of experience of feed rates and chip loadings that I just don't claim to understand, but then I wouldn't expect them to understand the behaviour of a h-bridge and failure modes of a MOSFET.

Horses for courses.

(but I'm determined now to get this sucker working again)

[Kitwn]

My apologies if I've sent you down a rabbit hole you were not intending to follow! On the other hand I think it's worth pointing out that many equipment faults can be repaired with a few tools which, once bought, will last for many years. There are more choices available than bin-and-replace for those willing to learn a new trick and I suspect you, like me, have reincarnated more than one piece of expensive electronics by knowing how to recognise and replace a faulty power supply capacitor costing mere pence.

If it were me I'd be bringing the soldering iron onto the dining table which is surprisingly close to my wife's weaving looms so I don't have to make your difficult domestic choice

[Neale]

The obvious question to ask is why, if there was a fuse, it was not accessible from the outside for replacement purposes? The answer, I think, is for the same reason that I don't stick fuses everywhere in my own control box. The fuse in this case did not protect the mosfets - it protected the rest of the upstream wiring and power supply from a potentially dangerous overcurrent situation causing a fire. Pound to a penny that the mosfets popped first, then the fuse blew to protect the rest of the system. The manufacturer knew this - no point in a replaceable fuse if the driver is dead anyway. Fuses are very important and shouldl not be forgotten - but just remember what they actually achieve and use them accordingly.

On a more positive note it's interesting to see that these drivers use well-marked and easily available replacement mosfets, although I would be a bit hesitant about working on through-hole boards - haven't had to do that myself.

[Doddy]

MOSFETs due in tomorrow from Farnell... along with some other goodies. Neale - that's pretty much the purpose of a fuse - and anyone relying on one to protect semiconductors has a lot more optimism than me. The more I look at these drivers the more I like about the design - the accountants haven't got too close to these, just close enough to the training of the assembly workers :)