duet ethernet blew both fuses, replacing hasnt helped

[A Former User]

@imrj said in duet ethernet blew both fuses, replacing hasnt helped:

@imrj i will try the idea of using low voltage first starting at 3.3v upwards and see how it goes

just to be pedantic, low current was the idea; but never the less I think you have the answer of what is wrong - but I'll leave the rest to Ian and/or David to decide what to do for now.

edit: mabe a closer look at the area between the daughterboard connector and the expansion connector just below and to the left of the cpu in your picture? were you using a thermocouple/pt-100 board?

[imrj]

@bearer no PT100 board, is a Duet Ethernet, thats the module u see there, but ethernet works fine when using 5V USB VIN...i looked there and all over but man just cant see anything wrong at all, i mean the board looks like brand new

[A Former User]

that part i meant? maybe its just a reflection or something, but as I can't see anything else

[imrj]

@bearer nah is just bad reflection, is clean as new there

[A Former User]

Mkay, well, I still think your first and best choice is to check if the board is still under warranty; and then see if the failure would be covered by that.

[droftarts]

@imrj Is it possible that you have applied 24V power with VIN and GND swapped? This would generally have the effect of destroying the stepper drivers; see https://duet3d.dozuki.com/Wiki/How_to_destroy_your_Duet_2#Section_Methods_that_might_work_or_only_impact_some_functions_of_your_Duet
Then applying power the correct way around would cause the shorted stepper drivers to blow the fuse. Not sure why the fan fuse would blow too. 5V is protected by a diode, which is why it works with the voltage the correct way around.

@dc42?

Ian

[arhi]

This type of issue is why I purchased a thermal camera. If you have access to one, get a power source that can go into the constant current mode, slowly increase the current and look at the board trough thermal camera. You will quickly find out where the short is.

If you can't get access to a thermal camera, you can try using a milliohm meter and try to find a short using it.

If you don't even have a milliohm meter you can try making one but fixing of the board is most probably outside your capacity and you should just find a way to get it to someone who can fix it for you.

[A Former User]

@droftarts said in duet ethernet blew both fuses, replacing hasnt helped:

Not sure why the fan fuse would blow too

and presumably causing both TR1 and D21 to fail?

imrj said in duet ethernet blew both fuses, replacing hasnt helped:

.if I short the botton block the PWM FAN2 gets got.....

[dc42]

@droftarts said in duet ethernet blew both fuses, replacing hasnt helped:

@imrj Is it possible that you have applied 24V power with VIN and GND swapped? This would generally have the effect of destroying the stepper drivers; see https://duet3d.dozuki.com/Wiki/How_to_destroy_your_Duet_2#Section_Methods_that_might_work_or_only_impact_some_functions_of_your_Duet
Then applying power the correct way around would cause the shorted stepper drivers to blow the fuse. Not sure why the fan fuse would blow too. 5V is protected by a diode, which is why it works with the voltage the correct way around.

@dc42?

Ian

I agree, I suspect that either you are applying VIN and ground the wrong way round now, or you have done so in the past.

btw applying reverse polarity doesn't always blow the drivers, it depends on the current limit of the PSU. I have got away with it in the past.

[A Former User]

without more information on the what happened prior to the failure I'm hesitant to suggest what to replace, but its pretty clear at a minimum you'll need to replace the parts you identified as getting warm TR1, D21, U5, U6, U7, U8 and U9 + U3 and D1 i'd recon. Almost $40 worth of parts, and a bit of work.

if this is caused by wrong polarity that may well be it. if its caused by over voltage/transient there could be more damage. having the 5v regulator and fan2 output fail is a bit concerning in that respect.

what was the printer doing before it failed?

[arhi]

@dc42 said in duet ethernet blew both fuses, replacing hasnt helped:

I agree, I suspect that either you are applying VIN and ground the wrong way round now, or you have done so in the past.

I find a beefy fast diode across Vdd and Vss immediately after the main fuse tends to protect the board in most cases. Adding a low drop / schottky trough diode after that for the low current side of things tends to increase the chance of everything surviving reverse polarity sky high as high current stuff usually can survive reverse polarity for the short period till fuse is blown due to current flowing through the fast diode across Vdd/Vss and the low drop will completely protect the low current parts from negative voltage.

[dc42]

There is already a diode protecting the 5V regulator.

If you apply reverse polarity then the body diodes in the output mosfets of the stepper drivers will conduct. Each driver has 2 pairs of series connected mosfets. So that's 10 pairs sharing the short circuit current of the PSU.

[arhi]

@dc42 great, so the fuse should blow before the big damage

[dc42]

@arhi said in duet ethernet blew both fuses, replacing hasnt helped:

@dc42 great, so the fuse should blow before the big damage

It depends on the short circuit current of the PSU. If the PSU supplies 20A continuously into the Duet, then I think it's likely that the drivers will blow, because that's 2A * 2 * about 2V per driver, which is about 8W per driver, and the 15A fuse will probably take a minute or two to blow. If the PSU has short circuit protection then it will probably go into hiccup mode, resulting in much less average current and the fuse remaining intact. That's what happened when I connected VIN the wrong way round.

If the fuse blows instantly when you connect VIN then it's likely that your PSU is providing a lot more than 20A.

[arhi]

@dc42 interesting, I don't have much experience with these "blade" type fuses. I never used them in my projects. Minutes sounds too long. I like to use polyfuses and the glass ones for "lower" current setup and efuse for high current ones. For low current stuff, I never measured response time but it's kinda "instant". E-fuse will trip faster then millisecond depending on the schematic. I never considered that fuse will take "minutes".

Polluting the topic, but maybe useful, some military stuff I designed some years back had interesting power protection. We controlled a lot of heavy stuff (over 40kW in total) and the guy who designed the power input part of the controller (remember, this can and will be, "serviced" by a 19yo soldier in the mud somewhere far from a service station) used a delayed circuit with a relay (big chunky ones) so when you connect power, the system does a "diagnostic" checking if power is ok (voltage levels, polarity, stability) and after a second or two if everything is ok it will turn on the relay and power the rest of the board. Normally we'd put there a beefy rectifier so input polarity gets irrelevant but we are pulling over 100A from those 48V sources

[A Former User]

@arhi said in duet ethernet blew both fuses, replacing hasnt helped:

I never considered that fuse will take "minutes".

Can take hours even. Depends on the over current and fuse rating ofc, even your 16A residential fuse/breaker would support 20A for anywhere between 30 and 60 minutes or more unless specifically installed as a fast breaking fuse which is uncommon.

[arhi]

@bearer said in duet ethernet blew both fuses, replacing hasnt helped:

Can take hours even. Depends on the over current and fuse rating ofc, even your 16A residential fuse/breaker would support 20A for anywhere between 30 and 60 minutes or more unless specifically installed as a fast breaking fuse which is uncommon.

I'm talking about electronics, not mains For mains I have 3 "slow" ones just after the meter (one for each phase) and then I have "normal" ones on the each floor of the appartment and then my room / workshop has it's own breaker box with "ultra fast" breakers + the workbench is 5kW 1:1 galvanic separated from the rest of the world with 3 separated variaks after the 1:1 with some super fast "settable" breakers (you can vary the current that will trip them). I'm moving this year into a new place (yeeeeeeeeeah) where I'll have a whole small 50m2 house just for my workshop, doing installation "these days".

So those big copper ones and the heat ones and the bimetal ones I get, but the blade ones to me fall in same category as those small glass ones that have very low "holding" strength and fail rather quickly and due to the difference of low current and short circuit they break inside milliseconds as even if your psu can deliver 100mA only just an output cap will be enough to break the glass one or trigger a poly when you short it out. That's why I mentally just wrongly looked at the blade fuses while it's totally normal as while on "micro electronics board" it's actually energetic part handling 10+ amps :). It's a common issue when you don't use something you look at it with completely different, often very distorted, eye sight.

[imrj]

ok i took care of the problem......

[A Former User]

that is one way to fix it.

I'd still recommend trying to find the cause of the failure before hooking up an even more expensive board though.

[imrj]

@bearer power was checked before firing it up, all checks out, is just this dam new RRF3 thats driving me nutz now....