Heated Bed Fault ?

  • Deleted

  • administrators

    I will assume that your bed heater current is no more than the recommended maximum of 15A for the Duet WiFi. If it is higher then the PCB and/or the bed heater mosfet may be overheating.

    I suspect that either your power supply is overheating and its output voltage is dropping, or you have a bad connection in your power or heated bed wiring that is increasing in resistance as it heats up. You can monitor the voltage at the power supply output, across the VIN terminals, and at the bed heater to find out which. Check that the VIN and bed heater terminal block screws are tight.

  • administrators

    I'm sorry to hear of your misfortune. Which model of Duet is it? Is there any other sign of damage apart from to the CPU?

  • administrators

    Please check the SMD inductor labelled L7. It is next to one corner of the processor chip, the corner nearest the SD card socket.

    My guess is that you had a short between the bed thermistor and the bed heater - probably one that only occurs when the bed is at a high temperature - and that you will find that L7 has gone open circuit. Or it could have been a short between the extruder heater and extruder thermistor, but if it was only the bed temperature reading that was wrong then a fault in the bed circuit seems more likely.

    On the Duet WiFi we added a fuse to protect the processor against this kind of wiring fault. To my surprise, this fuse has already blown on about 5 boards, which indicates that this type of fault is quite common.

  • administrators

    Yes it could, although normally there are some charred remains there, or at least a burn mark.

  • Hi David,

    I have an Duet 0.8.5 laying around that I want to use for another printer. And after I blow the fuse on the DuetWifi I ask myself if there is an option to put in a fuse on the 0.8.5 to prevent to burn the 0.8.5.
    I dont have KiKad installed, so I cant look at the layout.
    Is it reasonable to desolder L7 and put it in series with a fuse to have the same protection as in the DeutWifi?


  • administrators


    Hi David,

    I have an Duet 0.8.5 laying around that I want to use for another printer. And after I blow the fuse on the DuetWifi I ask myself if there is an option to put in a fuse on the 0.8.5 to prevent to burn the 0.8.5.
    I dont have KiKad installed, so I cant look at the layout.
    Is it reasonable to desolder L7 and put it in series with a fuse to have the same protection as in the DeutWifi?


    L7 is not the right place to put he fuse. It would have to be connected between L7 and the VSSA pins, which are on the thermistor connectors and the expansion connector. This would require PCB traces to be cut.

    The Duet 0.8.5 has better circuit protection than most other controller boards, because the processor analog inputs are protected against shorts from VIN to the non-VSSA side of the thermistor inputs.

  • L7 is a simple ferrite choke that bonds the processor's ADC and DC grounds to help reduce noise generated by the rest of the board from coupling with the ADC.

    So if the ground lead from your thermistor were shorted to any form of voltage from a heater cartridge, fan, LED, etc with enough current to overload and destroy L7 (at least a few hundred mA), once that choke opens there is nothing left to protect the MCU which in turn is destroyed as well.

    Yes, perhaps some additional diodes could be added to help keep analog ground clamped to ground to help protect L7 (possibly inducing noise the choke is intended to prevent), but then instead of L7 being damaged you'd just burn a trace instead. Or maybe some kind of sensitive crowbar circuit could be implemented to help protect the board by immediately shorting the power rails as soon as it scenes positive voltage on the analog ground…but that risks damaging your power supply when the crowbar triggers. In my mind that's all over-engineering. It adds complexity, more points of failure, and expense. At the end of the day, the Duet is a 3D printer controller, not a missile defense system.

    And in the case of the new Duet WiFi, to protect the MCU the VASA fuse handles only fractions of a watt before it opens...this is deliberate. I don't know of any user re-settable fuses that are that sensitive enough to effectively do the job. Unfortunately, the fuse is also considerably more sensitive than L7 was on the "wired" Duets which is why there are a number of threads about them blowing...it's easier to pop the VASA fuse than L7. On the other-hand the fuse on the Duet WiFi is an easy fix, a new MCU and L7 on the "wired" Duets isn't.

    So at the risk of sounding a bit facetious, essentially what you are asking is the same as "if I pour water into the back of my television set, is there any way you can design it to not be damaged?"

    These controller boards are sensitive devices that require a certain amount of diligence on the user's part to ensure they can function reliably. Securely terminating connections, taking care voltage sources are sufficiently insulated, ESD precautions, heat management, double checking everything before powering on, etc are all examples of that. If an accident occurs that damages something, my advice is to swallow your pride and chalk it up as wisdom. It sucks but that's the way this any every other DIY hobby goes. I've been doing electronic design and repairs for years...I can't even begin to explain some of the stuff I've blown up before I learned to always look before I leap. And even then I still have the occasional mishap. ūüėČ

    Anyway, that's my $.02 on the subject.

  • I agree completely with the sentiment of W3DRK's post.

    I've been using a wired 0.8.5 for a year and a bit, have run thousands of hours of prints through it, have taken it to trade shows and other events, and the duet has been the most reliable and consistent aspect of the entire printer. It sounds like you either got very unlucky, and encountered a possible TBD manufacturing defect, or you made a mistake.

  • Chiming in, I've run around 1000 hours of print time through a Duet Wifi with no issues.

    One thing to consider with board protection - the more of it you add, the more the board will cost and the larger it will be physically. At some point there's diminishing returns as you protect against scenarios with very low likelihood.

  • I don't think anyone is blaming the OPs wiring. As an unqualified outside observer, it seems to me that there is a fault within the heated bed. Probably intermittent and which started to manifest itself at higher temperatures where something expanded a bit more than normal or some sort of bi-metal bending occurred. Is this a "normal" silicone heater, stuck to the bed with a thermistor somewhere in the middle?

  • If you're convinced your Duet simply self-destructed all by itself, than I'm sorry but that's just nonsense. Have a look at the schematic on Github and study the analog ground circuit to see for yourself. https://github.com/T3P3/Duet/blob/master/Duet0.8.5/Duet0.8.5_Schematic.pdf

    If L7 and the processor burned out, your printer somehow applied voltage to one of the thermistor ground leads. Period. And if your RAMPs setup worked than it must been an unfortunate coincidence. In fact, if you think about it by changing controllers you obviously needed to re-position at least some of the cabling to accommodate different socket positions, possibly tugging on a connection that was only marginally insulated from a voltage source…maybe that's a good place to start? And like the previous poster noted, take a good look at how the thermistor is mounted to the heat bed. There has to be an explanation.

  • administrators

    Have you checked the hot end too? As well as making sure that the thermistor wires can't contact the heater wires, make certain that they can't contact the metalwork of the hot end. Cartridge heaters sometimes develop a short between the element and the can, putting the whole hot end at +12V or +24V.

  • administrators

    According to the data sheet, the inputs of the chip on the PT100 daughter board can take +/- 45V without damage. So using a PT100 would offer additional protection. The chip on the thermocouple daughter board does not have this protection.

  • This all reminds me of a previous life when I was an automotive engineer and we used to get intermittent faults. You can have all the diagnostic kit in the world and run every check and test imaginable but if the fault isn't present at the time, you ain't never gonna find it. We had one customer who used his E type Jaguar every day and it never failed him but at weekends, he used to go away and after about 140 miles, would stop at a toll booth. The engine would cut and refuse to restart for about an hour, after which it would start and run fine until the next time he did the same trip. We could never find a fault - everything was perfect but clearly something was wrong because the customer wouldn't make that up. I could never prove it but just had a theory that it was some kind of heat soak problem and that it was most likely his ignition module which was mounted above the radiator. We fitted a new ignition module just for the hell of it and it cured his problem. We had contacts in Bosch and gave them the old one to check thoroughly but they couldn't find fault with it either. Yet there must have been some sort of heat soak related issue with the unit. I can't help feeling that the OP has something similar with his heated bed. Just the ramblings of an old man‚Ķ........

  • @CaLviNx:

    I like that, I too was once a Mechanic, sorry my bad Automotive Engineer.

    Not that it matters, but actually over about 35 years everything from apprentice, through project engineer to operations manager and for the final 7 years of that particular career I ran my own consultancy business. So I tend to lump the whole lot together as "Automotive Engineer".

Log in to reply