Fried my duet

I finally found out what was causing my problems printing last few days. it turned out you guys had right and xbelt got trapped on end bearing causing x aksis to loose steps/tooth on belt.

I did like Erik and made brass bushings , all fine again

After a couple prints i uploaded a buildplate with 4 parts to run for a few hours then the 1 piece didnt stick to bed, for some reason printer missed on z height.
I did a pause, removed the first warped layer and lowered the z a little bit, then resume.
printer continued and it looked still bad so i did pause again then reset.
I did jog z up and remover parts and cleaned bed.
then power off/power on and home axes again restart print and first layer ok on all 4 parts, then all the sudden it stopped then i saw smoke from behind the printer and my duet died 21.05.2014 19:32, suddenly with no warning

I did check all wiring earlier today when i reassembled the ormerod again and i did it now and cant find any shortcut or anything wrong at all anywhere.
Ive been working as automation engineer last 20 years so im pretty sure i would find a fault in wiring if there is any.

Now duet is fried and if i connect usb only, prosessor is getting so warm i cant touch it after 2-3 sec.

I said warning in topic in case there is something wrong with reset function or print/ resume in 59b-DC42 but i higly doubt it. must be my duet that didnt want anymore, it almost printed 3 kg of filament so its tired i guess. lol


Sorry about the bad quality pic, taken on my phone since my stepdaughter ran off with my camera

Regards from Norway
Christian Jansen
Ormerod nr 446

Edited 2 time(s). Last edit at 05/21/2014 04:13PM by CJansen.

Apparently my duet diied yesterday and I've been trying to troubleshoot. Did not expect someone else to have similar symptoms. Well, mine did not have smoke but apparently just a few layers after a print started it froze and i did a reset on it, and it suddenly stopped responding while I was homing the axes again.

I went to take a look at the board and all the wiring (which i highly doubt was shorting or anything along those lines) and I'm pretty sure nothing could have caused it hardware wise.

When ONLY usb cable connected with everything including sd card removed, the Sipex 2525A-2E IC seems to rise rapidly in temp to the point which it is painful to the touch. I took an IR thermo which scanned the IC at 72c just 10 secs after micro USB connected.

Now your post makes me wonder if the newest firmware might be the cause, Im gonna stay off trying to troubleshoot just in case..

Quote
CJansen
...then power off/power on and home axes again restart print and first layer ok on all 4 parts, then all the sudden it stopped then i saw smoke from behind the printer and my duet died 21.05.2014 19:32, suddenly with no warning ...

Oh so sad so sad, only 3kg old and then this, its always the young ones innit...(sniff) are you sure it's quite dead?...could be stunned!...or pining for the fjords??

...oh draaat!, you are in Norway, wall to wall fjords! - Waaaa....(sniff) I'll light a candle - no I'll print a bloody candle, with 59b-DC42 firmware and respectfully pause 5 minutes half way through the print and restart - anything for a fellow brother in ormerod, anything!

...hope you don't take offense, only trying to cheer you up mate :-), lets see what the forensic might show

Erik (sniff)

Quote
ormerod168

Oh so sad so sad, only 3kg old and then this, its always the young ones innit...(sniff) are you sure it's quite dead?...could be stunned!...or pining for the fjords??

...oh draaat!, you are in Norway, wall to wall fjords! - Waaaa....(sniff) I'll light a candle - no I'll print a bloody candle, with 59b-DC42 firmware and respectfully pause 5 minutes half way through the print and restart - anything for a fellow brother in ormerod, anything!

...hope you don't take offense, only trying to cheer you up mate :-), lets see what the forensic might show

Erik (sniff)

5 min break half way in respect. rofl good one m8 And absolutely no offence taken

But seriously speaking.. im pretty sure its fried and dead unfortunelatel cause its getting so hot i cant touch it on usb power with nothing attached , and when i look at it i can actually also see the burning point, right in the middle between M & E in ATMEL

anyways, im pretty sure ian and the guys did make a few spare duets so it wouldnt surprice me if i can get my hands on a new one
also, i did copy your idea with the brass bushings for x and y axis great idea m8, cant believe i didnt came up with that one myselves

Regards
Christian

Christian, I'm sorry to hear about the demise of your Duet. I presume you know about the issue with the heat spreader shorting against the heated bed terminals if ti is installed the wrong way round, and you have made sure yours is the right way round.

There is no known issue with Pause/Resume in 059b, other than that gcodes already in the queue have to be completed before the pause takes effect (as has always been the case).

Quote
Prefinity
When ONLY usb cable connected with everything including sd card removed, the Sipex 2525A-2E IC seems to rise rapidly in temp to the point which it is painful to the touch. I took an IR thermo which scanned the IC at 72c just 10 secs after micro USB connected.

It sounds like then 2525A has failed, and since that is connected to 2 pins of the SAM3X, it may in turn have fed 5V into those pins of the SAM3X, killing it.

Various things might cause the 2525A to fail:

* If the switching regulator fails short circuit (temporarily or permanently), it will feed 12V to the 5V supply of the 2525A

* With the USB connected, a mains ground transient could easily cause a spike on the 5V supply to the 2525A. We already know that ground transients can cause the Duet to lock up, so it's almost certain that large ones can destroy it. I'm careful to connect the PC and the Ormerod to the same surge-protected distribution block, and to have nothing else that is likely to generate transients connected to the same block. I also have a ferrite bead on the USB cable. Most of the time, I don't even have the USB cable connected.

I don't think that the 2525A is useful. I think it would be better removed, and ferrite beads put on all 5 connections to the USB connector (or a series resistor in the case of the ID pin).

Quote

Now your post makes me wonder if the newest firmware might be the cause, Im gonna stay off trying to troubleshoot just in case..

I understand your caution; but I don't see any way that this sort of failure could be provoked by firmware.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

DC42, i was also 100 % sure firmware has absolutely nothing to do with it, just wanted to alert just in case. and your confirmation about the firmware makes me sure its not a softwareissue.

I do know about the heatspreader shorting and i can confirm it didnt short cause i was watching the print and it was printing in the middle of bed on second layer when it just died .
i also have isolated the heatingbed connections to prevent that to happen.

I also dont have usb cable attached when printing.

But this aint first time ive had hardware died / burnt for absolutely no reason, ( last week we had a dust monitor node frying cause it didnt tolerate 24vdc on the 24vdc terminal ) so i know some components fails just out of the blue, so i guess i have had bad luck and got a component on my duet from a bad lot

Ill talk to ian to see if i can order a new one from them or if rs can supply me one.

Regards
Christian

PS. so far your firmware has worked like a charm for me. continue the good work

Edited 1 time(s). Last edit at 05/21/2014 04:12PM by CJansen.

Sorry to hear the News of Dying Duets.... I hope that their souls rest in peace.... They do have Souls..... When a print goes wrong I normally call mine an A$$ Soul.....
And as a sign of respect I will be going to the Pub, and holding a 2minute silence while drinking my first pint in two whole weeks......
It may be followed by a number of others, pints and silences, as I hate trying to talk while drinking a pint, as beer goes every where...


If you disconnect everything, but leave the power connected, and then power it up, does it still get hot?
If not then replace one cable at a time to see if it's that...

---

Please send me a PM if you have suggestions, or problems with Big Blue 360.
I won't see comments in threads, as I move around to much.
Working Link to Big Blue 360 Complete

Very strange, frying in the middle of a print like that with no USB connected. Does your 2525A chip get hot too, like Prefinity's? Or does the SAM3X smoke so fast that you can't tell?

EDIT: also worth trying with just the USB connected, like Prefinity. If the SAM3X doesn't heat up so fast, that could indicate that 12V is getting somewhere it shouldn't.

Edited 1 time(s). Last edit at 05/21/2014 04:37PM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42
Very strange, frying in the middle of a print like that with no USB connected. Does your 2525A chip get hot too, like Prefinity's? Or does the SAM3X smoke so fast that you can't tell?

EDIT: also worth trying with just the USB connected, like Prefinity. If the SAM3X doesn't heat up so fast, that could indicate that 12V is getting somewhere it shouldn't.

Very strange indeed!

i added i picture of the duet to my original post , take a look at it. it goes hot laying on my desk with only usb connected.

the 2525A i dont know, the SAM3x gets so hot i dont dare to look more, you can see on the picture that the SAM3X desided to call it in
But atleast now i have time to finnish up the rest of the alu parts to my ormerod


EDIT: putting in closeup of SAM3X



Regards
Christian

Edited 3 time(s). Last edit at 05/21/2014 04:57PM by CJansen.

I've been thinking about this and I think I may have found a possible cause (it's only that, the actual cause could be something entirely different).

Here is a photo of the voltage on the drain of the extruder heater mosfet, before I added a gate resistor to slow down the turn-off:



The voltage appears to rise by 50V in about 20ns. That's an average slew rate of 2.5V/ns.

In the hot end wiring loom, there are 2 wires connected to this mosfet drain, one wire connected to a SAM3X ADC input, and one wire connected to the SAM3X VSSA (analog ground) pin. There will be capacitance between the mosfet drain wire and the ADC and VSSA wires. The loom is about 900mm long and I'm guessing that this capacitance will be of the order of 50pF each. A slew rate of 2.5V/ns across a 50pF capacitor draws a current of 125mA. So there will be about 125mA flowing into each of the Duet pins that are connected to the ADC input and AGND, for 20ns once every PWM cycle.

The ADC input has a 1uF capacitor connected between it and VSSA (analog ground). But it's a polarized capacitor, and it's ESR for a 20ns pulse may be quite high. At best, it shifts the current pulse to the VSSA pin. At worst, there is a 125mA 20ns pulse flowing into the pin protection diode of the ADC input.

The AGND connection is connected to ground through inductor L7. So as the current tries to flow to ground through this inductor, there will be a voltage induced across it. The cable capacitance and this inductor will form a series-tuned circuit, which raises the possibility of getting very large voltages between VSSA and ground. Unfortunately I don't know the value of the inductor, the schematic labels it unhelpfully as "100MHz".

There is probably an internal diode in the SAM3X between the VSSA and GND pins. But we are asking it to carry perhaps as much 250mA for 20ns.

It is well-known that excessive current into the pin protection diodes of MCUs can cause parasitic transistors to turn on and destroy the device. Atmel does not give a current rating for the pin protection diodes, but one of their application notes suggests that for the atmega processors, the safe limit is around 1mA. I normally design for no more than 100uA current into pin protection diodes. A current of the order of 100mA is way too high, even if it is only for 20ns. The SAM3X can be clocked at 84MHz, which implies that it's capable of responding to things in less than 20ns.

The bed heater mosfet and cable will produce a similar effect, however the slew rate and capacitance are both lower than for the hot end, and it turns off much less frequently.

I know we have some electronic engineers on this forum, so would any of them care to comment on this theory? I can think of a few ways to test it:

1. Look for pulses on the ADC input and on VSSA with an oscilloscope, on a board that has not been modified to slow down the slew rates on the mosfet drains.

2. If there is a large enough population of boards that have been modified to slow down the slew rates at the mosfet drains, then eventually they should show lower failure rates than unmodified boards.

EDIT: thinking about it some more, I may have under-estimated the capacitance, I think it may be nearer 100pF.

Edited 2 time(s). Last edit at 05/21/2014 05:50PM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

BTW, Ian made a tread for data collecting of Duet failures:

Data needed on Ormerod failures
[forums.reprap.org]

Quote

Hi all

I really need your help, to collect some data, to try and work out what is going on with a spate of recent Duet failures. I am not apportioning blame, and it will not affect the state of any warranty claims. I just need to understand what's going wrong, so we can fix it. This relates to a number of emails I have received recently, but I have also seen in comments in the forum of other, similar problems. Examples of these failures have been cited in this thread [forums.reprap.org] by PhilipTheMobster, tat88 (maybe), robbycar, emmander, TONYR, OttoES and possibly quite a few others, and some who don't post here but have contacted me by email.......

Ian
RepRapPro tech support

Erik

Quote
dc42
I've been thinking about this and I think I may have found a possible cause (it's only that, the actual cause could be something entirely different). ...

Yes, it sounds plausible, but I am not quite convinced. I've made some pretty heath-robinson lash-ups with mixed signal and heavy current switching via long multi-core cables going directly to CPU and other types of chip input, and whilst they frequently cause firmware crashes and other erroneous operations, I have never experienced a case where an induced spike actually fried a chip - though static discharges have sometimes done so, (which I suppose could be a possibility), and I once blew a circuit with induced RF from a 100W HF transmitter aerial inches away from a signal cable. I'm not saying that it cannot happen, but IME it is exceedingly rare. There is capacitance across the loom wires, but also an inductance. I also don't trust 'scope traces to give accurate voltage readings of switching transients. Even using a low-capacitance differential amplifier probe on a DSO, I have experienced such variable results after simply moving the wires around or unplugging the 'scope from the mains and running it on its internal battery that I have disregarded all the traces wrt their quantitative indications. And I have also experienced the effect where a solid fault goes away and the circuit works perfectly as soon as you clip on the 'scope probe.

Dave
(#106)

Hi all

Sorry for the slow reply, it's been a busy couple of days...

I've read the thread and PMs and I think we can sort out replacements. Please email me on support at reprappro dot com to make it official. I'll need your full name and address, RS order number (or proof of purchase), and let me know if your printed plastic parts are red or green, as we don't have access to the RS customer database. Please also send a summary of the problem, so I know what the problem is. Thanks!

Ian
RepRapPro tech support

Welcome back

I sendt you the info on email. have a good weekend and ty for response.

Regards
Christian