Mosfet questions

Discussion in 'Technical Repair' started by RAWtexas, Jun 12, 2018.

  1. RAWtexas

    Thread Starter New Member

    Jun 12, 2018
    5
    0
    Hi. I recently blew out my AIMS 3000 inverter. Upon opening it up I noticed that a couple of IRFB3206's had been smoked. I ordered and have received a dozen but I'm not confidant on a couple of things.

    1) One of the blown mosfet's, and a few others, has "2LIQ" written on it, while others don't. Is there a difference or should I assume all IRFB3206's set in row are the same?
    2) Is there any way to test those that don't have visible damage, short of unsoldering them?

    Thanks for any help
     
  2. N11778

    Member

    Dec 4, 2015
    163
    38
    Hook it up backwards did you.
    If so change them all.
    when you go to test it then, put a car headlight is series with the 12 v input if its 12 v.
    if something is still wrong it wont blow up and if its ok it should work with no load on it or a small 10w bulb
    good luck they are very self destructive if everything is not right. again good luck.
     
    olphart and RAWtexas like this.
  3. danadak

    Well-Known Member

    Mar 10, 2018
    1,916
    398
    From dasheet -

    upload_2018-6-13_5-27-41.png

    Regards, Dana.
     
    RAWtexas likes this.
  4. RAWtexas

    Thread Starter New Member

    Jun 12, 2018
    5
    0
    So that's just an assembly lot code. Thanks, Dana. I haven't found a reference anywhere to testing them while still on board so I guess that can't be done.
     
  5. ebp

    Well-Known Member

    Feb 8, 2018
    2,245
    787
    FETs will generally fail short-circuit between the drain and source. If there is sufficient power available they may then go open circuit - often blowing the package apart in the process.

    The problem with trying to to test in-circuit in something like a high power inverter is that there may be several FETs in parallel so they may all appear to be shorted. The gate often gets shorted to the rest of the slag and even with paralleled FETs each gate will typically have a series resistor to the driver. The resistor may be low in value (probably in the range of about 3 to 10 ohms) but it may still be possible to distinguish a FET with a gate short form others in a paralleled group.
     
    RAWtexas likes this.
  6. RAWtexas

    Thread Starter New Member

    Jun 12, 2018
    5
    0
    Thanks. If there are 24 mosfets and I don't have the time, inclination, and stock to replace all of them, what would be your best guess. 2 are visibly smoked, the other 23 look good. Those 2 are in a what looks like a group of 6. You think I may be ok just replacing those 6, or are mosfets so delicate that I'll likely be wasting my time?
     
  7. danadak

    Well-Known Member

    Mar 10, 2018
    1,916
    398
    If MOSFETs each have individual Gate R then start by looking for gate shorts.

    Then tack to each gate a ground, and then measure buss, should show some
    significant resistance. If true then one by one open ground, pull gate up to Vth or abouts,
    and see if buss low Z, then return gate ground and go on to the next. Tedious needless
    to say. This finds a MOSFET whose gate cannot turn it on for whatever reason.

    If buss, with all gates grounded, low Z then you will have to open drains, one by one.

    I am sure folks will look sideways at this but I would use a PSOC, wire up all the gates
    to GPIO (looks like Vgs of 5V enough to make this work), and write a little code, and use
    the Idac in PSOC to drive the buss and measure its R with ADC. That way once only wiring, and
    total test would be mS to complete.

    Regards, Dana.
     
    Last edited: Jun 13, 2018
    RAWtexas likes this.
  8. RAWtexas

    Thread Starter New Member

    Jun 12, 2018
    5
    0
    Well, that didn't work.

    POOF

    Apparently, the blown mosfets were a result but not the cause of issue. Actually, issue was probably just AIMS. Never did like how this 24/3000 got hot enough to fry a 24oz sirloin, draw or not. Its guts showed me why. One 3" fan and poorly designed heatsinking.

    Thanks for your help, people.
     
  9. danadak

    Well-Known Member

    Mar 10, 2018
    1,916
    398
    What went poof ? You got pic of area/parts on board that look heat damaged ?

    Regards, Dana.
     
    RAWtexas likes this.
  10. RAWtexas

    Thread Starter New Member

    Jun 12, 2018
    5
    0
    I didn't observe thoroughly but it looked like the same set of mosfets. It's ok. The inverter was costly but I ultimately recognized it as a garbage. I ordered a Samlex same day I ordered the mosfets.
     
  11. ebp

    Well-Known Member

    Feb 8, 2018
    2,245
    787
    I have no experience with inverters like that, so I can't offer any reasonable advice. If it appears that the six are in parallel, it may be worth replacing that whole group. If all the FETs are in groups, I would suggest checking each group for shorts before doing any replacement. Generally, FETs are pretty robust, though obviously there are things that will kill them - gross overcurrent of sufficient duration and overvoltage are killers.

    If you have appropriate cutters and can get access, cutting the leads near the body or just where they widen can make removal a lot easier.

    When you install new FETs, I strongly recommend installing the mounting hardware before soldering so you can be sure everything is lined up.
     
    Kjeldgaard likes this.
  12. N11778

    Member

    Dec 4, 2015
    163
    38
    My Aims is junk too. The bearings in the fan went out in 5 months. replaced it with a ball bearing one.
    Its been working for 5 years now, with the new fan. Let us know what the new one looks like inside.
     
Loading...