Without seeing it all I would say (if you haven't already done so) that you should connect the speaker return to the main PSU star point and not to some point on the PCB.
Have you checked the output stage quiescent current ?
You could also try adding a small (less than 20pf) cap across the 20k feedback resistor and/or reduce the 1k feedback resistor to see if that has any effect on the poor sound quality.
Also try adding a 47uf (not critical) between the base's of the two output transistors.
Have you checked the output stage quiescent current ?
You could also try adding a small (less than 20pf) cap across the 20k feedback resistor and/or reduce the 1k feedback resistor to see if that has any effect on the poor sound quality.
Also try adding a 47uf (not critical) between the base's of the two output transistors.
That 47 if would probably make switch off worse with those darlingtons. The base emitter resistors on monolithic darlingtons are in the wrong place. You really want it to be inclusive of the emitter resistor. And you really don't want the 8k on the driver. The way they are arranged just begs to be run with the outputs in cutoff and just a trickle thru the driver.
If the TIP's are passing current under no signal conditions and the high current speaker return is connected so that it can not influence the amplifier by contaminating the feedback and signal grounds then you simply have to use a scope to see what the nature of the distortion is.
We are just guessing from here on in. This type of amplifier is very layout/wiring dependent. We need to see what the distortion looks like and how the amplifier behaves under load.
We are just guessing from here on in. This type of amplifier is very layout/wiring dependent. We need to see what the distortion looks like and how the amplifier behaves under load.
Are you absolutely sure? With those Rbe's connected improperly like that you measure both the driver and output stage currents together when you monitor across Re. And the first 10 to 15 mA of that is in the driver alone - you will have to well exceed that to get anything in the outputs. You need half a volt or so across those 40 ohms. It's why that connection sucks - it makes biasing a PITA. . You can get away with running the outputs class B if two things are true - if you have a butt load of global negative feedback and a butt load of current gain (i.e., triple). This has neither. So you gotta bias both the drivers and outputs.
If you bias with a vbe multiplier, you will see a sudden jump in how much current you get for a given rotation in the pot. When this happens with monolithic darlingtons you'll know the outputs finally "kicked in". Below that you will have noticeable crossover distortion.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.