Questions about modifying a class AB amplifier

Thread Starter

GeorgeTR

Joined Feb 3, 2018
35
Previously, in this thread, I discussed my modifications to a class AB amplifier in a stereo console unit from the 1970s. Now I have another similar unit that I got at a rummage sale that I would like to do something very similar to, and now that I've already done it once, for the most part it's been easier. But, with the other stereo I had the original manual containing the schematic, and with this one I have no documentation. So, just by examining the PCB, I created this sketch of the heart of one channel of the amplifier. There are of course many more components in the circuit, these are just the ones that I particularly care about (and most of the other ones I wouldn't quite know what they're for).
IMG_20190807_195622634~2.jpg

My plan is to replace every component displayed here in order to allow the amplifier to sound better and be much more powerful. The DC blocking capacitor just before the connection to the speaker will be replaced by one that is 2200 μF instead of 470 μF to allow for much better bass, the 2.2 Ω resistors will be 5W instead of 1W, and the transistors will be TIP41AG for the NPN and TIP42AG for the PNP. Since the A738 and C1368 transistors were in a TO-126 package, and the new ones will be in a TO-220, the pin configuration will be backwards of what it was, but I have taken this into consideration.

My main questions, then, are these:

1. Does anyone know why a resistor and a diode in series may have been used instead of two diodes in order to bias the transistors? Would it be acceptable to replace the diode and resistor with two new diodes? There is a ~1.3V drop over the resistor and diode, so I would imagine two diodes would be pretty much the same, and since the goal is to match the characteristics of the diodes with the transistors, I would imagine it might be better to replace this configuration, especially since I am replacing the transistors. Last time, I used these same transistors and 1N4003 diodes, and it seemed to work great.

2. The use of 2.2Ω resistors seems to be quite inefficient. The previous stereo I modified used 0.5Ω resistors, which seemed much better to me. If I were to play a sine wave just at where it clipped (and I realize this isn't what music is like, but it makes the math easier), and I had a 4Ω load (which I plan to do, since I will have two sets of 8Ω speakers configured in parallel), if my math is correct, each resistor would dissipate 2.51W. And, even worse, the maximum power to the speakers would be at 21.9W with no resistors, and only 9.1W with these resistors. Would it be possible to reduce the value of these resistors, say to the 0.5Ω used in the other stereo (or to any other value)? Would this require additional modifications of the circuit?

If it would be helpful to know any additional details about the circuit, just let me know, and I will try to provide them.

Thanks in advance!

GeorgeTR
 

MrChips

Joined Oct 2, 2009
30,720
Hello George,

Everything you have suggested seems ok to me. What you are doing is not simply substituting individual components but replacing the entire output stage.

Emitter resistors serve a number of purpose with consequences.

1) Limits output current - provides some protection from short circuit and catastrophic failure.
2) Increases output impedance - reduces output power.
3) Provides better matching between two output transistors - less crossover distortion.
4) Increases negative feedback - less linearity distortion.

If your new design worked for you previously, go ahead and try it.
One difference you might encounter is that the current gain of the power transistors is lower. You might need more current drive from the preceding driver stage. Also the DC bias might need tweaking.
 

Thread Starter

GeorgeTR

Joined Feb 3, 2018
35
Thanks for the help! I guess I will try using new diodes and reducing the resistors to 0.5Ω, and if there's no thermal runaway, it will be good to go. If there is thermal runaway, then I'll have to worry about what to do about it. It makes sense why they used 2.2Ω resistors, since with 0.5Ω resistors, you could possibly burn out the transistors by turning the volume up to high.

I don't imagine the slightly lower current gain will be a problem, since the old transistors in the other stereo also had a slightly higher current gain. Also, since I used 2 1N4003 diodes with these transistors in the past, correct me if I'm wrong, but I think the DC bias will be good to go.
 

BobTPH

Joined Jun 5, 2013
8,814
To get significantly more power you would need to up the voltage and modify the stage before this since it would not have the needed voltage swing.

Bob
 

Thread Starter

GeorgeTR

Joined Feb 3, 2018
35
I do realize that this won't allow me to get that much more power, at best it would let me go from about 4W to about 20W, or about 7dB. However, when I did this to my other stereo, even before I replaced the dc blocking caps, to my ears it sounded better. Also, with this configuration I feel a lot more comfortable cranking it with a 4Ω load, something it probably was never really designed to do. If I were to increase the power significantly, I would probably just bypass the original amplifier and put in a Class D, but that's not something I intend on doing.
 

Audioguru

Joined Dec 20, 2007
11,248
Can the power supply in this low power amplifier cope with more than 5 times the original power (assuming 8 ohm speakers before)?
 

Thread Starter

GeorgeTR

Joined Feb 3, 2018
35
Thanks for considering that. I have thought of that, and so, similar to the last stereo I did, I will be replacing the bridge rectifier and main filtering capacitor to allow for the increased power draw.
 

Thread Starter

GeorgeTR

Joined Feb 3, 2018
35
After some time, and a couple frustrating setbacks (nothing to due with the design), I have gotten it working with the new setup. It seems that no matter whether you have the volume cranked until just before it clips, or you have it at a much lower, reasonable listening volume, the temperature of the transistors levels off around 75°C. I'm fine with this, even though I know it's not very efficient, and wouldn't be part of a marketed design. So I guess this is a success, as far as I am concerned.
 

Thread Starter

GeorgeTR

Joined Feb 3, 2018
35
It seems I spoke to soon. I thought that after running it for over an hour on full volume, its temperature would have leveled off. I also thought that because it quickly cooled off when I blew a fan on the heat sink, that this was just an issue of a large amount of heat that needed to be dissipated and not a huge heat sink, not an issue of thermal runaway. But, after continuing to run it, even though the volume is much lower, the temperature has very gradually continued to climb, so I now believe this must be an issue of thermal runaway. I'm pretty sure that replacing the emitter resistors with ones of the original 2.2Ω value would solve this, but I was wondering if anyone knew of a different way, since I would rather not have those resistors be 2.2Ω, for the reasons I mentioned above.
 

Thread Starter

GeorgeTR

Joined Feb 3, 2018
35
It seems I spoke to soon, again. I thought the temperature was continuing to climb, but it turns out it leveled off at 80°C. Counterintuitively, the temperature that the transistors stabilize at seems to be higher when the volume is lower. I would still like it if anyone new of a way to make this lower, 80°C is hotter than I would like for whenever the stereo is on. However, I've already gotten quite frustrated by this project a couple of times, so I don't know if I would want to spend time trying things out only to have them not work.
 

Thread Starter

GeorgeTR

Joined Feb 3, 2018
35
Seems my problem was I just needed a bigger heat sink. I slapped a bar of aluminum with some thermal paste on it onto the heat sink I already was using, and the temps plummeted down to 35°C and stayed there. The power consumption of the entire stereo also more than halved. Now I just have to fabricate a better heat sink out of all the scrap aluminum I have lying around.
 
Top