Oznog
Active Member
I have a 386 audio circuit that will be driven by a power supply from 9v to 14.6v. It's using a NJM386 which can tolerate up to 18v and the pkg has an 800mW dissipation. 2V, maybe 3V peak is an expected audio output level for the headphones I have (not sure of the impedance, I will need to check), there are also some headphones up to 300ohms I know may be used. The project has a dsPIC on it.
I have a particular type of mono jack that someone might plug into it indefinitely. That jack will short one of the amp outputs to ground.
This is a trick. Now at full voltage the 386 will overheat at only 55mA if the output is shorted. I heard that a 10 ohm resistor in series with the output is common to provide short-circuit protection. I think that anything larger may fold back the bass response since the speaker's impedance is much lower there. But this won't protect it at this supply voltage! 2v pk/10ohm=141.4 mA RMS which is almost 3x the current needed to cook the amp. It looks likely to cook the amp even though average power is not this high.
This controller has unused ADC pins and can cut off the input to the 386 completely, since the output goes through a series cap to the headphones this will result in no current in a shorted output.
I'm trying to take in all my options here:
1. Put a thermistor on the amp and have the dsPIC shut it down when overheating. I can't shut it down forever though since somebody may swap in a stereo headset later. Really this isn't a proper mounting job either.
2. Sense the voltage at the headphone use that to determine if it's shorted. Well, dsPIC can't read negative voltages. At best I could put in a Schottky and then be able to read if it drove a signal that should have made a +0.3v signal but didn't. It might be hard to catch this too since the hardware and software weren't designed to look for when the signal would produce over 0.3v on the amp's output.
3. I could put a self-resetting Polyfuse in series with the power going to that amp, but it would be tripping and resetting constantly.
4. It'd be nice to read the current going to the amp, but that would need to be a high side shunt in a 12v system which would require an instrumentation amp which I'm not really ready to add to the cost and size for.
5. Lower amp voltage, or series limiting resistor with the amp's power line. I'd need a pretty big cap for this sort of thing to keep the power rail from bobbling around on heavy notes. There also may not be any voltage margin for it, the higher impedance headphones need as much voltage as it can get.
Am I missing anything here?
Any other ideas?
I have a particular type of mono jack that someone might plug into it indefinitely. That jack will short one of the amp outputs to ground.
This is a trick. Now at full voltage the 386 will overheat at only 55mA if the output is shorted. I heard that a 10 ohm resistor in series with the output is common to provide short-circuit protection. I think that anything larger may fold back the bass response since the speaker's impedance is much lower there. But this won't protect it at this supply voltage! 2v pk/10ohm=141.4 mA RMS which is almost 3x the current needed to cook the amp. It looks likely to cook the amp even though average power is not this high.
This controller has unused ADC pins and can cut off the input to the 386 completely, since the output goes through a series cap to the headphones this will result in no current in a shorted output.
I'm trying to take in all my options here:
1. Put a thermistor on the amp and have the dsPIC shut it down when overheating. I can't shut it down forever though since somebody may swap in a stereo headset later. Really this isn't a proper mounting job either.
2. Sense the voltage at the headphone use that to determine if it's shorted. Well, dsPIC can't read negative voltages. At best I could put in a Schottky and then be able to read if it drove a signal that should have made a +0.3v signal but didn't. It might be hard to catch this too since the hardware and software weren't designed to look for when the signal would produce over 0.3v on the amp's output.
3. I could put a self-resetting Polyfuse in series with the power going to that amp, but it would be tripping and resetting constantly.
4. It'd be nice to read the current going to the amp, but that would need to be a high side shunt in a 12v system which would require an instrumentation amp which I'm not really ready to add to the cost and size for.
5. Lower amp voltage, or series limiting resistor with the amp's power line. I'd need a pretty big cap for this sort of thing to keep the power rail from bobbling around on heavy notes. There also may not be any voltage margin for it, the higher impedance headphones need as much voltage as it can get.
Am I missing anything here?
Any other ideas?
