Just wanted to post that I've made a mcu solution that seems to be working. I've placed a mcu along with a digital potentiometer on the output of the op-amp, the output being ground reference for the ics. The emitters of each output transistor is also connected to the MCU so that measuring the difference in voltage will tell the bias current. The circuit will slam violently up and down when the amplifier is outputting though, but I don't think the mcu would mind as long as it's supply is properly regulated. I'm going to make it so that it can detect fast variations in bias current so that it would know if the amp is outputting or not so that it only does calibration when the output is zero. One issue is that the digital potentiometer and MCU will reset if the op-amp gets close to the top rail, but having either a MCU or a pot that has EEPROM will make it set the bias current correctly shortly after. I guess that is a worst case scenario though, having the op-amp constantly slamming into the positive rail. Here's a schematic of the solution:
One thing that might be an issue is to get enough resolution on the potentiometer. The one in the schematic is a 64 step which is not good enough, I hope that using a 257-step a long with resistors connected to the A and B terminals to limit it's operating area will give sufficient resolution. One can also reduce the value of R4 so that the difference between wanted bias current and maximum bias current is reduced. If a two-wire serial interface potentiometer like the one in the schematic can be used, a six-legged mcu will be able to do the job. Using the 30 cent attiny4 for this will make this a rather cheap solution, always wanted to use one of those for a project.
One thing that might be an issue is to get enough resolution on the potentiometer. The one in the schematic is a 64 step which is not good enough, I hope that using a 257-step a long with resistors connected to the A and B terminals to limit it's operating area will give sufficient resolution. One can also reduce the value of R4 so that the difference between wanted bias current and maximum bias current is reduced. If a two-wire serial interface potentiometer like the one in the schematic can be used, a six-legged mcu will be able to do the job. Using the 30 cent attiny4 for this will make this a rather cheap solution, always wanted to use one of those for a project.
Last edited:
More info
Please also check:
http://www.diyaudio.com/forums/soli...rating-class-ab-amplifier-22.html#post5104638
Please also check:
http://www.diyaudio.com/forums/soli...rating-class-ab-amplifier-22.html#post5104638
My way is to stay away from 3EF 2EF using CFP so I only have to deal with 2 Vbe from the driver BJTs instead of 6 or 4 Vbe with two in those at hight varying temperature output transistors junctions.
So with only two driver Vbe, it gets much easier right away. Then the key is thermal bounding of the driver transistors with the base spreader vbe multiplier or vbe diodes, which means tight thermal contact of cases and drivers with low Rth junction-case.
So with only two driver Vbe, it gets much easier right away. Then the key is thermal bounding of the driver transistors with the base spreader vbe multiplier or vbe diodes, which means tight thermal contact of cases and drivers with low Rth junction-case.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.