ad: MyersEng-1

Homebrew SS amplifier zapping transistors

Discussion in 'Amateur Radio Amplifiers' started by KZ4AK, Apr 28, 2016.

ad: L-HROutlet
ad: l-rl
ad: L-Geochron
ad: abrind-2
ad: Left-3
ad: Left-2
ad: l-BCInc
ad: L-MFJ
  1. KZ4AK

    KZ4AK XML Subscriber QRZ Page

    I have a homebrew amp using the AN758 board from Communications Concepts with switched (multi-band) output filters. Works nicely, except... It does not have automatic protection or band switching. Some time back, I hit it with drive power with the amp set to the wrong band. Instantly zapped the outputs (MFR429). Replaced them with SD1726 devices which looked a bit more rugged. Well, 6 months later the idiot did it again! Gone in microseconds... Can't keep this up at $200 for pair of transistors. Supply voltage was 50 vdc.

    Would like opinions of what might happen to gain and power - and would it offer more resistance to high SWR (wrong filter) failure if I reduce the Vcc to 40 vdc.

    Thanks,
    Woody, KZ4AK
     
  2. WB2WIK

    WB2WIK Platinum Subscriber Platinum Subscriber QRZ Page

    $200 a pair is a lot. It's about double what CCI charges for a matched pair of MRF429s: http://www.communication-concepts.com/mrf429mp/

    I wouldn't reduce the Vcc, based on the spec sheets.

    MRF429 rated 150W output with 3W drive at 50V and still in linear part of the curve; same device at 150W out requires 5W drive at 40V, and is already non-linear.

    MRF429 @ 150W out -35dBc IMD (IM3) at 50V; it's only -30 dBc IM3 for same power at 40V.

    Seems it all starts to fall apart much below 50V.

    I'd think about using a fast acting circuit breaker in the DC bus that trips off quickly at 30A. Might help.

    On the early Ten Tec rigs using bipolar PAs, for many years the only PA protection was circuit breakers in the power supply and incredibly it seemed to work well.
     
  3. KZ4AK

    KZ4AK XML Subscriber QRZ Page

    Thanks for the reply! The 429 has a bit better IMD, but a bit lower voltage rating. At half the price (and for a matched pair), will probably go that way. I was wondering if over voltage or over current took them out. Happened very quickly - Key down was for only a second or two before I realized what was happening. FWIW, They failed open circuit, not shorted. One thing for sure, they don't like 20M drive with the 160M output filter selected!

    A fast CB might help (?) and I will probably build one. Have some high voltage/current FETs used for switching supplies (RDS = 0.014 ohms at 75 amps). Not if it is only fast enough!
     
  4. WW1WW

    WW1WW Ham Member QRZ Page

    I have been running home brew SS amps for a few years and experience has shown that running at 45 volts is less prone to failure than running at 50 volts. I would pop a PA module about once every 2 months when running at 50 to 52 volts, it has been over a year at 45 volts without problems.
     
  5. KZ4AK

    KZ4AK XML Subscriber QRZ Page

    That is encouraging to hear. I have reduced the Vcc from 50 to 45. There will be a little compromise in performance, but like your experience, hope it will add some margin of protection.
    _W_
     
  6. KF5FEI

    KF5FEI Ham Member QRZ Page

    Some radios supply a voltage on one of the rear connectors which varies according to the band selected. Maybe you need to set up an Arduino or similar to detect this voltage, compare it to the band switch on the amp, and disable transmit if they don't agree.
     
  7. KZ4AK

    KZ4AK XML Subscriber QRZ Page

    That is a good solution but I would need to get a KRC2 my older Elecraft K2. In the long run, might be cheaper than chronically replacing the PA finals!
     
  8. AH6LE

    AH6LE Ham Member QRZ Page

  9. W8JX

    W8JX Ham Member QRZ Page

    I would use a fast blow fuse sized for max normal current.
     
  10. KM3F

    KM3F Ham Member QRZ Page

    I see this situation a bit differently.
    For solid state devices, it usually is excess RF voltage that pops them on an instant basis, not current.
    RF voltage that rises above the solid state junction limit cause instant breakdown.
    RF voltage can exceed the DC operating level by huge margins to cause the issue.
    As opposed to current that will take a small amount of delay time for temperature rise and cause burn out from excess temperatures.
    I would bet you have a match issue where lowering voltage will lower the junction operating voltage conditions and permit buying some time, but does not address the route issue cause.
    Slow and smaller fuses etc are usually not helpful other than blowing until you hit the problem by luck. That costs a bunch of 'doe' in parts and frustration until you get lucky.
    Good luck.
     
    W8SPL likes this.

Share This Page

ad: HamInsurance-1