I am in the process of learning about different amp topologies and designs, and so far, the only thing I have built is a Pass A40. With class A, biasing seems simple in many ways, just give it a lot of bias and call it good. In trying to learn class B, the exact bias point matters, and how to get there is eluding me.
I understand the bias is supposed to be set high enough to eliminate crossover distortion, putting the transistors at the point of conduction, but without going over and drawing too much current. The only thing I came up with is looking at the emitter resistors, but I don't know what to look for exactly.
Without having any knowledge of specs (except supply voltages and general topology), is there an easy way to set the bias to the correct level with just a DMM and o-scope? Any help would be appreciated.
I understand the bias is supposed to be set high enough to eliminate crossover distortion, putting the transistors at the point of conduction, but without going over and drawing too much current. The only thing I came up with is looking at the emitter resistors, but I don't know what to look for exactly.
Without having any knowledge of specs (except supply voltages and general topology), is there an easy way to set the bias to the correct level with just a DMM and o-scope? Any help would be appreciated.
ClassAB EF output stage is optimally biased when Vre ~26mV.
The difficulty is that the re includes the internal transistor resistance as well as the external Re.
As the external Re is reduced, the proportion of Vre across the external resistor reduces to maintain that optimal bias condition. This is best done by monitoring the crossover distortion and setting the output bias to suit. That method is not available to most builders.
Settle for a Vre~18mV for Re=0r1 and rising to ~24mV for Re=0r47. Self listed the intermediate values of Re and the total Vre across the series pair. Self also defined his ClassB the same as what everyone else labels optimal ClassAB.
The difficulty is that the re includes the internal transistor resistance as well as the external Re.
As the external Re is reduced, the proportion of Vre across the external resistor reduces to maintain that optimal bias condition. This is best done by monitoring the crossover distortion and setting the output bias to suit. That method is not available to most builders.
Settle for a Vre~18mV for Re=0r1 and rising to ~24mV for Re=0r47. Self listed the intermediate values of Re and the total Vre across the series pair. Self also defined his ClassB the same as what everyone else labels optimal ClassAB.
Thanks for the help, everyone. I have an old Sony STR-V6 receiver that I was fixing, and when I first got it, the power supply thermal fuse kept blowing after a while, but when I replaced the fuse it sounded ok. I figured the caps were leaky/old and the bias would be off. I was only seeing ~10mv or so across each .47ohm Re, and it didn't seem right according to what I'd heard. There is indeed a set of 4 diodes in series, and an adjust pot per channel. I've replaced all the major electrolytics, and an open resistor, and now I need to adjust it. Is it possible to see the crossover notch on an o-scope, or will the resolution be too low on a digital scope such as a Rigol 1052, etc?
As soon as I get a feel for understanding how all this works, I'm going to try building a class B amp, but haven't decided what project would be a good one yet.
As soon as I get a feel for understanding how all this works, I'm going to try building a class B amp, but haven't decided what project would be a good one yet.
The crossover distortion is easiest seen with a low signal level like a volt or less on the output with a frequency of 10-20kHz. Sometimes its not a right off flat, but rather a little kink in the waveform around the zero crossings, so watch for these straightening out.
Adjust until the flat parts around the zero crossings just about disappear, check the voltage across the emitter resistors, then let the amp sit turned on for ~15 minutes to let things stabilize and settle down, then recheck the bias and readjust if needed, then let it sit for another ~15 minutes, recheck bias again and your done.
I'd make a tutorial on doing bias adjustment on an amplifier, but after all flack i've gotten from ppl that my circuits, measurements and other stuff are flawed, i'm not going to as i will likely be doing this wrong as well.
Adjust until the flat parts around the zero crossings just about disappear, check the voltage across the emitter resistors, then let the amp sit turned on for ~15 minutes to let things stabilize and settle down, then recheck the bias and readjust if needed, then let it sit for another ~15 minutes, recheck bias again and your done.
I'd make a tutorial on doing bias adjustment on an amplifier, but after all flack i've gotten from ppl that my circuits, measurements and other stuff are flawed, i'm not going to as i will likely be doing this wrong as well.
should follow datasheet of power trans. search detail of it and adj correctly value current cut off of power transistor.
Sorry to ask so many questions, but I think this will be the last one on this thread. Well, I put the amp on the scope, and fiddled with the bias pot till the crossover disappears and the waveform looked normal at 10khz, which was only around 10mV Re or so... Going beyond that to 26mv would more than double the dissipation, such is class AB right? As this amp was built to be class B, is there any reason (performance or otherwise) other than heat that I shouldn't put it at 26mV and call it good? The heatsinks just barely feel warm after 30 min at idle, and are of substantial size, being at the tail end of the 70's receiver power wars.
Sounds like you can definitely try giving a value up to 26 mV a shot. You're still in underbiased territory now, and if the amp can handle more without getting too warm, you can obtain even better distortion performance. But don't overdo it, since SOA on those old power transistors may not be that luxurious and they die more easily when hot.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.