DC output protection circuitry

[pinkmouse]

Originally posted by unclejed613
...even simpler is the one shown below...

What? Where?

[d3imlay]

I like traderbam's circuit. I've been working on a circuit and hope to publish soon. When I started I wanted to know what actually happened on the output terminals when an amp failed. I assumed that shorted outputs would be the failure mode, but was curious what happened a few milliseconds before the failure.

My test amp was a Phase Linear 400. Read on before you laugh. In order to expedite the process I removed 2 of the 3 outputs in the positive rail on one channel. I ran it a 1/3 power (about 70 watts) into 8 ohms based on the assumption that an amplifier runs hottest at 1/3 power. I used a sine wave at about 150Hz so I could capture the power supply fuse clearing which took a few milliseconds. I monitored the one xistor with a thermalcouple. The xistor temp ran up to 125C pretty quick and sat there. Ten minutes later it was still there and no failure. So I shut down the test and loosened the screws on the xistor. The temp went to 150C and after about a minute the amp failed. Unfortunately so did the trigger on the scope so I missed it. I triggered off of the emitter resistor and it took 4 failures to get it right. That was aggravating enough, but when you do get it right and forget to reset the scope trigger it's worse.

I did about 8 tests and noted that upon failure the output goes to the rail (big surprise) but immediately before the failure everything was normal. Curiously, there was never a failure in the negative rail. Most of the time when an amp fails it gets an output in each rail. I was amazed at how hard it was to get the Phase Linear to fail under controlled conditions. It makes me wonder what's different about living room conditions driving a speaker and the controlled bench testing into a resistive load.

The circuit I am developing will use an output relay and a crowbar triac, across the output, on the line side of the relay. I haven't confirmed it yet, but I'm convinced that standard relays can't clear 80 VDC without considerable arcing/welding. The crowbar will allow the relay to open up because it will fire before the contacts open due to mechanical issues with the relay.

[AndrewT]

Hi,
a resistive load on the amplifier is very gentle.
To force the amp into failing more quickly you would need to substitute a load of about half the specified minimum impedance or even less.

However, a reactive load does stress the output stage very severely.
Adding a capacitor in parallel to your specified resistive load to bring the load phase angle upto about 45degrees and this puts about double the current through the output transistor when it is passing significant voltage (Vce). Increase that capacitor (or increase the frequency) to give a 60degree phase angle and the current increases by about another 50%, to 300% of resistive. You will promote very quick failures this way. Almost as soon as the devices get hot (you won't need to slacken any screws).

This is the SOAR limit that gets talked about.