Can an IGBT be intermittent?

Discussion in 'Automotive Electronics' started by NMNeil, Jun 18, 2019.

  1. NMNeil

    Thread Starter New Member

    Jun 18, 2019
    14
    2
    I have a problem with the ECM in my car. Whatever I do to try and fix it I still have an occasional misfire on cylinder #6 that will not go away. It takes days to get it to set a code and then the scan shows P0356, coil primary fault. Good power to the coil, 0.2 ohm resistance in the control wire to the ECM. Swapped coils, new plugs, PCV valve etc, no change. Reading the TSB's and recalls the IGBT coil drivers in certain Fords was prone to failure and Ford changed out not only the ECM/PCM, but all the coil packs (COP system) under warranty.
    My question is, can the coil IGBT, in this case the ON Semiconductor B8202N, have an internal fault, or are they either good or bad with nothing in between?
     
  2. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    9,936
    2,403
    I would expect almost any electronic component to be prone to an intermittent fault. The automotive environment is harsh. Vibration or temperature cycling could in theory affect, for example, an imperfect wire bond inside a semiconductor device such as an IGBT.
     
  3. Analog Ground

    Member

    Apr 24, 2019
    125
    28
    "In the old days", we suspected intermittent connectors with bad contacts or arcing. Your troubleshooting with disconnecting and connecting things may be temporarily fixing the problem.
     
    Last edited: Jun 19, 2019
  4. NMNeil

    Thread Starter New Member

    Jun 18, 2019
    14
    2
    I know just about enough about electronics to get myself into trouble, so I checked the IGBT for #6 cylinder in circuit. Problem is the readings for that IGBT and the other 5 was almost identical so no clues to if it's good or bad.
    As the B8202N is obsolete can it be replaced with any 20A 400V logic level IGBT designed for ignition circuits such as one from the EcoSPARK range?
     
  5. narkeleptk

    Member

    Mar 11, 2019
    255
    92
    Usually when its intermittent its a bad solder joint in the circuit for that coil.
     
    Last edited: Jun 19, 2019
  6. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    9,936
    2,403
    Proof that tracking down intermittent faults is a pain. Get your meter out and the device under test is immediately on its best behaviour :).
     
  7. NMNeil

    Thread Starter New Member

    Jun 18, 2019
    14
    2
    Unfortunately the PCB is glued to a heat sink, so I can't get to the underside. The IGBT is surface mount and visually the solder joints seem good. As the IGBT is logic level I'm assuming that there is no driver IC and the IGBT is controlled directly from the MCU. I checked the gate to emitter resistance and it shows about 16K, which matches the datasheet internal pull down resistor value. As the ignition IGBT only cost about $1.60 each I'm going to change them all out. It will also give me the opportunity to replace any lead free solder with some decent 60/40 solder.
     
  8. NMNeil

    Thread Starter New Member

    Jun 18, 2019
    14
    2
    Unfortunately the Ford 'engineers' made it impossible to remove the ECU without taking off the intake manifold, so in-situ testing is out of the question except with a scope.
     
  9. narkeleptk

    Member

    Mar 11, 2019
    255
    92
    I would recommend that you do not replace them all and only do the one you feel is in error. Without removing the board from the case/heat sink you will need a lot of heat to remove the driver and the pads will start coming up pretty easily. The risk of damage is much greater then the reward of the unneeded new drivers.

    What part number ecu are you working on? Maybe I can tell you whats on the other side.
     
    Last edited: Jun 19, 2019
  10. NMNeil

    Thread Starter New Member

    Jun 18, 2019
    14
    2
    I'll get the number later.
    I'm of the opinion that as the ECM is such a pig to remove and if one IGBT has failed there is the nagging concern that the others may not be far behind . ON makes excellent products but as this is a known problem the run of IGBT's from the foundry may all be bad. To me it makes sense to change them all out while I have it on the bench, and my Hakko has no problem putting out the heat.
     
  11. narkeleptk

    Member

    Mar 11, 2019
    255
    92
    The drivers don't typically fail. Mostly its the coils that blow them up. Fix them all if you want but its not needed and likely more harm then good. I always like to stick to the old phrase. "if it ain't broke, don't fix it" either way, good luck to you and hopefully it solves your problem..
     
  12. NMNeil

    Thread Starter New Member

    Jun 18, 2019
    14
    2
    I find the phrase "If it ain't broke, fix it till it is" is more appropriate to myself :cool:
     
  13. NMNeil

    Thread Starter New Member

    Jun 18, 2019
    14
    2
    OK, I got my new IGBT's today and went ahead with desoldering the old one.
    Even with the Hakko set at 850 the heatsink was pulling the heat out so fast that the solder under the D2PAK would not melt. Tried my 60 watt solder gun but that failed as well. Tried flooding the area with flux and lead solder, but still didn't work. If the IGBT wasn't bad at the start I figure it is now.
    Bit stuck as what to do now as the heatsink is glued to the PCB. I could hit the IGBT with the Dremel and get it down to the metal substrate, then I should be able to get it hot enough, but I would still have the problem of excess heat when I soldered in the new one.
    Any ideas anyone?
     
  14. bwilliams60

    Senior Member

    Nov 18, 2012
    1,126
    219
    I have done a few of those and used a dremel to get down to the base. You have to be very careful not to go through it as it is very thin and you need all the material you can get. I agree with @narkeleptk that you should only change the suspect IGBT and leave the rest alone. You may be biting off more than you can chew. As far as the drivers go, they do not fail very often and are best left alone where possible. The pads tend to come up easy on those. The IGBT failure is born from bad coils, not a bad batch of IGBTs so I would be comfortable just changing one and moving on.
     
    narkeleptk likes this.
  15. narkeleptk

    Member

    Mar 11, 2019
    255
    92
    If you get to the end up your rope, are in the US and need some help you are welcome to send it to me. Will not be free but I will give you a good price. Just PM me a I will send you my company info. I repair ford pcms every day. Can easily replace the driver, check the other side for bad joints, flash update it from ford servers and then test it out.

    You never mentioned the year and model for me to offer you any advice before messing with the driver but I guess that is of little concern now since you already got into it. Little more advice, be very careful if you attempt to remove it from the case. Most people who try damage the pcb beyond a reliable repair.
     
    Last edited: Jun 25, 2019
  16. NMNeil

    Thread Starter New Member

    Jun 18, 2019
    14
    2
    I agree about the biting off more than I can chew.
    I do question the assumption that the coils cause the PCM driver failure. If the coil primary shorts it will go from about 0.5 ohms to 0 and the current will go very high, but as the driver is rated for 20A continuous and 50A surge the coil power fuse should pop before it gets to that level. On the other hand if the coil fails so that it induces a high voltage back EMF the internal back to back zener diodes will clamp the voltage to about 450 volts, which is well within the margin of safety.
    But thinking about it the zeners are possibly not meant to handle repeated clamping and that's what fails.This could then lead to the gate pull down resistor failing so that the gate capacitance is never dissipated and the gate is floating, hence the erratic behavior.
     
  17. NMNeil

    Thread Starter New Member

    Jun 18, 2019
    14
    2
    I may be admitting defeat and taking you up on your offer.
    It's a 2005 with the 3.0 DOHC motor, PCM has many numbers, the most prominent is JJZ0
     
  18. narkeleptk

    Member

    Mar 11, 2019
    255
    92
    If that is the teartag # then I can say that I do not see to many of those come in with issue's but they are pretty easy to replace if it does get damaged. Usually that style case is not as hard to get into as some of the others so you might be ok.
     
    Last edited: Jun 25, 2019
  19. narkeleptk

    Member

    Mar 11, 2019
    255
    92
    From my experience @bwilliams60 is correct, most all driver issue come from the coil side of things. When they start to fail the high voltages from the secondary makes its way through the low voltage primary and wreaks havoc inside the pcm.

    there is one popular one that I work on from fords that I swear the main problem is the pcb design of pcm. I am no engineer but it seems to me the coils work the drivers so hard they overheat the area and since the controller for them is placed damn near directly under the drivers it fails from this heat, causing a big mess to multiple coils and the pcm. Still it was said to be faulty coils in the tsb put out on it and they designed new ones that helped suppress some of the feedback which seem to help a little.
     
    Last edited: Jun 25, 2019
  20. NMNeil

    Thread Starter New Member

    Jun 18, 2019
    14
    2
    Whatever the cause when I get the misfire sorted the car goes on Craigslist to be someone else's problem.
     
    narkeleptk likes this.
Loading...