Could transistor still be damaged if it passes diode test?

Discussion in 'General Electronics Chat' started by zero_coke, Mar 21, 2013.

  1. zero_coke

    Thread Starter Active Member

    Apr 22, 2009
    294
    1
    We have an amplifier here and we've had no luck in getting an amplified output.

    The manual says to place a dummy 100 ohm resistor in place of the transistors to adjust the bias voltage and we did that and we set it to 0.7V. We removed that dummy resistor and soldered the MRF454 NPN transistors in place. The biasing is correct but the transistors dont seem to come on because they're very cold and there are some parts of the circuit which are very hot like some of the resistors. We thought it may be the voltage regulator but it's providing the necessary 0.7V bias to the base when we tested it with the dummy resistor.

    We removed the transistors and did a diode test and it was showing 0.7V drop across BE and BC so it passed the diode tests. Could the transistors still be damaged even if they pass the diode test? We don't know what's wrong because the biasing is working correctly and we get no output. The amplifier is the AN762 by Motorola. The transistors are the MRF454 NPN.
     
    Last edited: Mar 21, 2013
  2. wmodavis

    Well-Known Member

    Oct 23, 2010
    737
    150
    Did you check the resistance between CE?
     
  3. crutschow

    Expert

    Mar 14, 2008
    12,974
    3,220
    Post your schematic.
     
  4. zero_coke

    Thread Starter Active Member

    Apr 22, 2009
    294
    1
    No I did not. But I did do a diode test and it showed nothing.

    As for schematic, please see the attached file:
     
  5. nigelwright7557

    Senior Member

    May 10, 2008
    487
    71
    A lot of meters have sockets for a Hfe check.
    I have found transistors with a Hfe of 1 which is clearly wrong.
     
  6. zero_coke

    Thread Starter Active Member

    Apr 22, 2009
    294
    1
    Would an hfe check be the most accurate check to see if a transistor is good or bad? What is a good range of hfe value for an RF transistor (MRF454) ?
     
  7. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    Generally small transistors will read a hfe of 100-300 or so on a multimeter hfe test.

    It's unlikely your transistors are dead and still measuring ok on a diode test B-E. That can happen in high voltage apps after enough time but is very unusual for short term faults in low voltage apps. Your transistors are probably ok.

    The most likely problem is you have miswired the circuit or have the transistors in the wrong orientation, or if this is an RF design it might be failing due to poor layout and not following good RF layout techniques. A PHOTO would be of great help!
     
    zero_coke likes this.
  8. w2aew

    Member

    Jan 3, 2012
    219
    64
    This application note describes how to setup the bias for the transistors. It will depend upon whether you want to run the amp in Class A, Class AB, etc. Because there is no emitter degeneration, the adjustment of the bias voltage is critical, so it must be done carefully.
     
  9. zero_coke

    Thread Starter Active Member

    Apr 22, 2009
    294
    1
    Hi guys:

    The photo of the amp we built is attached.

    @w2aew: We did follow the application note and there were two things we had to do. One was to set the bias voltage by placing a dummy resistor from base to collector and adjusting the trimpot till we get 0.7V and this passed successfully. Then we had to place the transistors on the board and connect the +Vcc to a multimeter as an ammeter and measure the bias current to be 100mA or more. We couldnt do this...the meter was showing 0mA. I don't know why it was doing that.

    I know my partner always plays with the trimpot when the amplifier is connected to the Vcc and has an RF input as well. You shouldn't be playing with the trimpot like that or it will change the bias right?
     
  10. w2aew

    Member

    Jan 3, 2012
    219
    64
    I don't see L3 and L4 populated on your board.
     
  11. zero_coke

    Thread Starter Active Member

    Apr 22, 2009
    294
    1
    w2aew: We didn't install those because we thought they were optional probes for current measurement...do we need them in there? Jeez if this is the problem then I'll shoot myself...not really but 1 week of troubleshooting would've been gone to waste!
     
  12. w2aew

    Member

    Jan 3, 2012
    219
    64
    Yes, they are absolutely needed. That's where the collectors get their bias. Be sure to use the correct beads as shown in the parts list. The only optional part is constructing them with a longer wire to be able to use a current probe.
     
    zero_coke likes this.
  13. zero_coke

    Thread Starter Active Member

    Apr 22, 2009
    294
    1
    Wow ok. Thanks w2aew! This is most probably why it hasn't been working! Jeez....thanks again!
     
  14. w2aew

    Member

    Jan 3, 2012
    219
    64
    It definitely won't work without them. Be sure to carefully set the bias after you install them. Setting the bias too high could damage the transistors.
     
    zero_coke likes this.
  15. zero_coke

    Thread Starter Active Member

    Apr 22, 2009
    294
    1
    The manual says to set the collector current bias to >=100mA or >=200mA total for both transistors. We will set to 150mA each I guess. Thanks once again w2aew! And also a warm thanks to The_RB for the photo request!
     
  16. zero_coke

    Thread Starter Active Member

    Apr 22, 2009
    294
    1
    @w2aew: Uh oh, we've run into a completely different problem now. And this problem, thanks to a mistake we made otherwise we wouldn't have ever figured it out, is with the transistors being mounted on the heatsink.

    According to the datasheet, the base metallic part of the transistor which is mounted onto the heatsink is not associated with any of the base/emitter/collector pins. However, when we screw it down, it shorts out the base pin to 0V. We put some isolation between the heatsink and the transistor by using 3M thermally conductive rubber pads instead of the heatsink paste and this prevents the base package from touching the heatsink but now the screws are the problems. The screws touch the heatsink and this is unavoidable unless we use nylon/teflon screws which we don't have.

    First of all, why in the world are we running into this problem? I've never heard of such thing...the metallic base of the transistor is not connected to any of the pins (we also tested using multimeter continuity check). I don't understand why it short circuits the base pin to 0V when it touches our heatsink (which is grounded, of course).

    Anyone have any suggestions? Could it be a manufacture defect of this transistor because seriously I have not found anything else wrong with the circuit. I replaced the old transistors with new ones ($40 a pop!) and still the same problem :(

    We wrapped the screws with electric tape and screwed them in and then we also placed that isolative thermal conductive rubber instead of the heatpaste but now our transistors starts smoking when we reach an output of 20V RMS into a 50 Ohm load (~8-10 Watts). I'm speculating that the rubber pads are not providing enough thermal conductivity which is why we see smoke at low input powers even.
     
    Last edited: Mar 24, 2013
  17. patricktoday

    Member

    Feb 12, 2013
    157
    42
    They make bushings to prevent a screw from touching the transistor or the inside of the ring it passes through. This is the wrong transistor type but just for example:
    http://www.ebay.com/itm/50-Units-He...hing-And-Film-TO-220-/150471475680#vi-content

    I'll be curious if someone has an answer on this because I've had similar issues in the past where I was not able to tighten the screws very much or the transistor would make contact with the heat sink.
     
    zero_coke likes this.
Loading...