Do BJT finals go soft?

Discussion in 'General Electronics Chat' started by inwo, Jul 29, 2014.

  1. inwo

    Thread Starter Well-Known Member

    Nov 7, 2013
    MJ12005 was replaced with ecg238 in 2009 according to my notes.

    Unfortunately my notes don't show if machine was dead, or weak output.

    This is in a 20 watt 1mhz ultra sound. 120hz pulses ~4ms long.

    Looks good on scope.

    Checking into from 15 to 82 ohms.

    Much more or less power drops off drastically.

    If going soft is common, I'll replace it with proper one this time.

    As my methods of measurement are questionable, I might be better off leaving well enough alone.
    Last edited: Jul 29, 2014
  2. inwo

    Thread Starter Well-Known Member

    Nov 7, 2013
    Guess I'll find out. Ordered some new transistors.

    Best I can come up with is being about 1/2 power.

    My measurements could easily be off that far. :(
    Last edited: Jul 29, 2014
  3. inwo

    Thread Starter Well-Known Member

    Nov 7, 2013

    Thought some hams would know!
  4. #12


    Nov 30, 2010
    Theoretically, molecules move because of heat. More specifically, the diffused parts of a transistor can change over time. This can cause a low gain transistor, which is what I assume you mean by, "going soft". If you're good at this, you can calculate the gain of the transistor(s) even if you have to add an ohm in the drive circuit and measure across that.

    In practical terms, a transistor that lost its gain over time is a rare event. You almost always find them in the Q.C. department because they were out of spec the day they were delivered from the manufacturer. It is more likely that some other part has gone soft, like a overheated resistor, a capacitor, or a circuit board trace. Still, it's possible, especially when dealing with ECG or NTE relabeled parts.
    inwo likes this.
  5. vk6zgo

    Active Member

    Jul 21, 2012
    Some HF power transistors are really ICs---they have multiple Emitters.

    They tend to lose several Emitters when they fail.
    When you test them at low power,they show normal gain,but driven with full drive,they do not output full power
  6. inwo

    Thread Starter Well-Known Member

    Nov 7, 2013
    While waiting for parts, I'm trying to get a baseline power reading.
    1mhz @ ~ 1/3 duty cycle pulses.
    20watts average power showing on instrument meter. (whatever that means)

    Sorry to be long and boring, but this helps me get my thoughts in order.:D

    Tried my vhf and 27mhz watt meters.
    No reading on vhf.
    1.5W on 27 mhz meter. Not surprising that it's way off.:(

    Measure 130vpp into 82 ohms.23W @ 1/3duty cycle = ~8watts

    I don't trust any of these methods as duty cycle is a guess.

    At least I'll have something to compare!

    Last method:

    At full power. 20W on meter, 82ohm dummy load, heats from 25C to 75C in 58 seconds.

    Compare to 4W dc heats 25 to 75 in 70 seconds
    And 4.4 watts dc heats 25 to 75 in 51 seconds.

    I trust my heating method the most which tells me I'm between 4 and 4.4 watts average power.

    Most likely error between tests is my estimation of duty cycle.
    The pulse are tapered, slow to decay.

    By changing my math to meet results :eek:, a 20% duty cycle makes them agree.:)

    Still have no idea what the designer means by average power.
    Actual average power over time, or the average power in each pulse.

    If reference is average power in each pulse, I'm there, and no repairs are needed!:confused:
    Last edited: Jul 31, 2014