What is the quickest way to determine what the issue with a class ab amp "not working"
Because its a feedback amp, you can't really check every small signal stage closed loop, so what I do is short both input terminals and check all biases. Is there a quicker way to determine any messed up solder join or bad transistor?
Because its a feedback amp, you can't really check every small signal stage closed loop, so what I do is short both input terminals and check all biases. Is there a quicker way to determine any messed up solder join or bad transistor?
The first thing I do is these tests without the power on.
1. If it's a bipolar transistor implementation, I attempt to measure the Vbe, ie, voltage from Base to Emitter on the output devices. Most meters have a "diode V forward measurement" mode whereby they source a fixed small amount of current from the positive lead to the negative lead then report the measured voltage on the display. This voltage should be around 0.7 volts, twice that if darlingtons. You may have to remove the transistors for an accurate test. The same goes for Mosfets only you are measuring the drain to source to ensure they are not shorted w/out power. Most solid state devices will fail short circuit across these terminals.
2. You could have bad power supply filter caps too. Power still off, put an ohm meter across them observing proper polarity and see if the resistance slowly rises as the caps charge. If not, the output devices could be shorting them out or the caps could be shorted themselves. They may have to be removed for a valid test. Note the meter will charge a capacitor and to repeat the measurement they will have to be discharged say with a low ohm resistor. An open cap will indicate open circuit. A leaky cap or partially shorted cap will not charge to infinity.
3. If this stuff is fine move onto the pre-driver small signal transistors. Once again with power off measure the Vbe of all transistors.
4. If this seems fine, with output devices removed or disconnected, and filter capacitors in the circuit, put some sort of dummy load say 100 ohms across these caps and see if you are getting the proper DC voltage under power. Use a scope to make sure you only see ripple voltage variation and not complete half wave rectified waveform which would be an indication of an open capacitor.
If this stuff checks out, you may have an open circuit or bad solder joint somewhere. To find this, assuming the thing doesn't blow fuses, and you've done the due diligence mentioned above, with no load, turn the power on and walk your meter around the circuit with the negative lead grounded to make sure the measured voltages make sense. It's all ohms law - never found an exception to it. You can measure Vbe on the active devices now using the regular voltage measure mode.
If you have replaced parts you may need to adjust the output offset - There's usually a pot for this. Also check for thermal drift. With no load, put the V meter on the output and GND. With the power on, after a minute or two, make sure the output voltage doesn't start to drift positive or negative. There's usually some thermal device mounted to the heat-sink. It compensates or negates thermal drift. It may have gone open.
Let's see is there more?
It can get tricky debugging a DC coupled design so you might consider breaking the feedback path. So long as you haven't got a speaker hooked up this is OK. You cold try steering the output by putting DC on the input. It should follow with an amplification factor of course so just use a little DC.
Hopefully, you have already found a bad device and you can just check for no offset, no drift and proper sound quality.
1. If it's a bipolar transistor implementation, I attempt to measure the Vbe, ie, voltage from Base to Emitter on the output devices. Most meters have a "diode V forward measurement" mode whereby they source a fixed small amount of current from the positive lead to the negative lead then report the measured voltage on the display. This voltage should be around 0.7 volts, twice that if darlingtons. You may have to remove the transistors for an accurate test. The same goes for Mosfets only you are measuring the drain to source to ensure they are not shorted w/out power. Most solid state devices will fail short circuit across these terminals.
2. You could have bad power supply filter caps too. Power still off, put an ohm meter across them observing proper polarity and see if the resistance slowly rises as the caps charge. If not, the output devices could be shorting them out or the caps could be shorted themselves. They may have to be removed for a valid test. Note the meter will charge a capacitor and to repeat the measurement they will have to be discharged say with a low ohm resistor. An open cap will indicate open circuit. A leaky cap or partially shorted cap will not charge to infinity.
3. If this stuff is fine move onto the pre-driver small signal transistors. Once again with power off measure the Vbe of all transistors.
4. If this seems fine, with output devices removed or disconnected, and filter capacitors in the circuit, put some sort of dummy load say 100 ohms across these caps and see if you are getting the proper DC voltage under power. Use a scope to make sure you only see ripple voltage variation and not complete half wave rectified waveform which would be an indication of an open capacitor.
If this stuff checks out, you may have an open circuit or bad solder joint somewhere. To find this, assuming the thing doesn't blow fuses, and you've done the due diligence mentioned above, with no load, turn the power on and walk your meter around the circuit with the negative lead grounded to make sure the measured voltages make sense. It's all ohms law - never found an exception to it. You can measure Vbe on the active devices now using the regular voltage measure mode.
If you have replaced parts you may need to adjust the output offset - There's usually a pot for this. Also check for thermal drift. With no load, put the V meter on the output and GND. With the power on, after a minute or two, make sure the output voltage doesn't start to drift positive or negative. There's usually some thermal device mounted to the heat-sink. It compensates or negates thermal drift. It may have gone open.
Let's see is there more?
It can get tricky debugging a DC coupled design so you might consider breaking the feedback path. So long as you haven't got a speaker hooked up this is OK. You cold try steering the output by putting DC on the input. It should follow with an amplification factor of course so just use a little DC.
Hopefully, you have already found a bad device and you can just check for no offset, no drift and proper sound quality.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.