2
\$\begingroup\$

I've made an amplifier from a circuit I found on the internet (below, link). I made it more or less exactly as described (same components, big heat sinks) and it seems to work great (can run at a loud volume for extended periods of time). Sometimes, however, for no apparent reason, the fuses blow and the MOSFETs have become completely conducting. I understand this can happen if a high (> 20V) voltage is applied to the MOSFET gate. No other component is damaged.

I've completely re-made the circuit about 6 times and it always happens on the same channel. The amplifier attached to the other channel might last for weeks, until I test it on the bad channel, then it breaks within a day or two. So I think what's happening is that one channel of the sound source (a cheap bluetooth receiver powered by a questionable 5V supply) occasionally has short voltage spikes, causing an over-voltage on the MOSFET gates and breaking them.

I know I could get a better source, but I want to make the amplifier better. So my questions are:

  • Is this a plausible diagnosis?
  • If yes, what can I do to make this circuit more robust against a "bad" input?

My first idea is to put Zener diodes between the gate and source of the MOSFETs (say 15V threshold diodes, the gate voltage should never get close to that during normal operation). Is this a bad idea? Or is there something else I could do?

enter image description here

Improve this question
\$\endgroup\$
  • \$\begingroup\$ Have you tried simulating this? The purpose would not be verifying if it works, but to give you the insight into how things work, since you can very easily probe for voltages or currents without the fear of perturbations, or that things might blow up. \$\endgroup\$ – a concerned citizen Jun 6 '20 at 6:35
  • \$\begingroup\$ I have not, and it's a good idea. Any recommendation for a simulation tool? I've played around a bit with LTspice but it's quite clunky and annoying to use. \$\endgroup\$ – Simon Coop Jun 6 '20 at 8:46
  • 1
    \$\begingroup\$ If there's anything I can safely say about LTspice is that, currently, there's no other SPICE software that outmatches it in terms of speed, both execution and productivity. Of course, I am not advocating for it, you're free to use whatever other simulator you fell comfortable with. \$\endgroup\$ – a concerned citizen Jun 6 '20 at 11:20
2
\$\begingroup\$

Place back to back Zeners of say 15V across each gate source .The mosfets are expensive and the zeners are cheap .I use gate zeners a lot and find that most other engineers accept this .Sure you get the odd one that wants to remove the gate zeners.For high reliability stuff it is just not worth it .

Improve this answer
\$\endgroup\$
  • \$\begingroup\$ Do you think this will affect the sound at all? \$\endgroup\$ – Simon Coop Jun 6 '20 at 8:45
  • 1
    \$\begingroup\$ @SimonCoop Well, there might be a slight case of nonlinearity involved, a bit more acute as you get closer to the limiting points, but I think it would far outmatch the risk of blowing those transistors up. After all, if you get that high with the gate voltage, chances are that your THD already increased considerably. \$\endgroup\$ – a concerned citizen Jun 6 '20 at 11:22
  • \$\begingroup\$ @ Simon Coop .Your sound will be fine .The Zener volts at 15V is above the normal voltage when the Amp is operating ,Most mosfets are said to be fully enhanced at 10 Volts . \$\endgroup\$ – Autistic Jun 7 '20 at 0:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.