I’m fixing an older Heathkit AA-15 integrated amp for my father. It had many capacitors that were open, high esr, or highly reduced capacitance. Another major issue and the main reason I had it was that once channel had a pretty major failure. Nearly all of the transistors in the one channel were bad.
Original outputs were 40411 and are now MJ802G
Drivers were 40409 and 40410 and are now TTA004b and TTC004b
Predriver was 40408 and is now TTC004b
The two 2N3393 at the input of the amp are now 2N5551c
These were the closest replacements and were recommendations. With these changes the idle bias current is quite a bit higher than I’d like to see lol. Measuring across the 0.67 ohm emitter resistor I’m seeing close to 100 mV or just under 150 mA of current. Is it possible to put in a trimmers in order to bring that down a bit?
Thank you,
Dan
Original outputs were 40411 and are now MJ802G
Drivers were 40409 and 40410 and are now TTA004b and TTC004b
Predriver was 40408 and is now TTC004b
The two 2N3393 at the input of the amp are now 2N5551c
These were the closest replacements and were recommendations. With these changes the idle bias current is quite a bit higher than I’d like to see lol. Measuring across the 0.67 ohm emitter resistor I’m seeing close to 100 mV or just under 150 mA of current. Is it possible to put in a trimmers in order to bring that down a bit?
Thank you,
Dan
You'll probably have to redesign or replace the fixed bias circuit, which is D201, D202, D203.
Have you got crossover distortion? If you have add a 1n4001 with a 220R preset across it in series with the three bias diodes.
If you have too much quiescent current, place a 470R preset across one of the bias diodes.
If you have too much quiescent current, place a 470R preset across one of the bias diodes.
I would put a 1k potentiometer across one of the diodes d201 d202 d203. If that gets you down to 20 ma idle bias, clamp the pot with a schottky diode. Pot wipers are subject to the wiper losing contact due to oxidation. Or perhaps the schottky diode alone is close enough to give a fixed lower bias current.
MJ802? You like living dangerously? They don’t like 40V rails, despite the current and power ratings. MJ15003 is the “correct” part. It’s vbe is “too low” too, requiring the same bias adjustment.
That would be a bummer, at least more difficult than adding a variable resistor. They currently use three metal can diodes mounted to the heatsink, going through holes.
I don’t know, I haven’t measured the output quite yet as I don’t want to let it idle away at 100 mV bias voltage. Being that high I imaging it wouldn’t. I want to make sure I get this right.
So if there is crossover distortion I would want to do this? Does it matter which side of the three bias diodes it’s placed on?
If I have too much current then I’d want to do this? When you call the resistor a “preset” are you meaning a variable and setting it at that resistance to start off?
Beautiful!
Lol, this is why I love this place. I have plenty of MJ15003, I’ll place in a quad of them first and then tackle biasing.
Thank you,
Dan
Picture 1 is what jonsnell said, but If you have 150 ma of idle bias current there should be no crossover distortion.
Picture 2 is sort of what I said. I'd put a 1000 ohm pot at first, remote with a couple of wires, then set the voltage on the emitter resistors to 13 mv (20 ma), then depower and measure the pot value. Then put in a fixed resistor of that value. Or clamp the pot with a parallel schottky diode, which can be suitably tiny to fit right in between the two leads of one silicon diode. Say MBR1100 or BAT48 .
If you have a scope you can look at sine wave output for crossover distortion, but I do not expect any at 20 ma idle bias current.
Allen organ sold thousands of S100 amps with a pair of MJ802 as outputs, They were very reliable at +-35 v. This if I can read it, is +-46 v. A whole different ball game. MJ15003 or same price MJ21194. I would change the mica washers with new ones, or leaving out the heat sink grease, silicon pads. 50 year old mica washers are IMHO at expected life.
Picture 2 is sort of what I said. I'd put a 1000 ohm pot at first, remote with a couple of wires, then set the voltage on the emitter resistors to 13 mv (20 ma), then depower and measure the pot value. Then put in a fixed resistor of that value. Or clamp the pot with a parallel schottky diode, which can be suitably tiny to fit right in between the two leads of one silicon diode. Say MBR1100 or BAT48 .
If you have a scope you can look at sine wave output for crossover distortion, but I do not expect any at 20 ma idle bias current.
Allen organ sold thousands of S100 amps with a pair of MJ802 as outputs, They were very reliable at +-35 v. This if I can read it, is +-46 v. A whole different ball game. MJ15003 or same price MJ21194. I would change the mica washers with new ones, or leaving out the heat sink grease, silicon pads. 50 year old mica washers are IMHO at expected life.
15003 has higher gain at 15A, a better match for the original hometaxial types. That’s why someone suggested MJ802 (high gain at 20A), but it’s SOA as you know isn’t good enough for the voltage.
Well swapping over to the MJ15003G helped a bit all on its own. After swapping the bias voltage across one of the emitter resistors is sitting about 64 mV on one channel and about 70 mV on the other.
I will get the trimmers install installed and adjust for correct bias current. I will likely try to go for a fixed resistor once I measure the resistance. Appreciate everybody pointing out the MJ802 not being a good fit. And I did indeed mica insulator with a new one. I’ll report back.
Dan
There is no way a 40411 was going to stand 20 A, or 15 A unless as a switch at DC. Fully saturated. I suspect AA-1515 means 150 w or 75 w/ch. About state of the art for one pair in 1975. That is 3.1 A into 8 ohm. Gain at 15 or 20 A does not matter IMHO. No data on 40411 but 40636 had gain >20 at 4 A. Suspect the rails were +-46 because the transformer voltage sagged so badly at full power. Cheap thin windings. Heathkit was a price leader, after all.
Last edited:
Alrighty, so I put a 1k trimmer set for 470 ohm across D201 of each channel. It’s just a sloppy install as once I get the value I need I’ll take all of this out and put in a resistor and run the wiring appropriately.
Edit: I have found that the 1 kohm trimmer doesn’t really have much affect until you near the end, I meant about 850 ohms and I am close to the 13 mV target. Using these trimmers, I have in there it’s quite hard to get it dead on, so I’m going to be swapping in some 25 turn trimmers. Get a lot more accuracy.
Dan
Edit: I have found that the 1 kohm trimmer doesn’t really have much affect until you near the end, I meant about 850 ohms and I am close to the 13 mV target. Using these trimmers, I have in there it’s quite hard to get it dead on, so I’m going to be swapping in some 25 turn trimmers. Get a lot more accuracy.
Dan
Last edited:
Well measure the emiiter resistor idle voltage. 470 ohm is just a shot in the dark. 13-15 mv on R235 is 20-23 ma idle current.
Note R237 is in the collector of Q207, not the emitter. 40411 had some built in emitter resistance. Modern epitaxial transistors do not. I would put at least .51 ohms 5 W between Q207 emitter and negative rail. I did in my ST120 with NTE60s (green box MJ15003 I think), and got about 6 years out of it at up to 14 hours a day without a meltdown before a wire popped off and it started humming.
Note R237 is in the collector of Q207, not the emitter. 40411 had some built in emitter resistance. Modern epitaxial transistors do not. I would put at least .51 ohms 5 W between Q207 emitter and negative rail. I did in my ST120 with NTE60s (green box MJ15003 I think), and got about 6 years out of it at up to 14 hours a day without a meltdown before a wire popped off and it started humming.
Last edited:
I did see that they’re not truly both emitter resistors and actually I’ve been measuring across the resistor that’s on the collector of the output. Does that matter? I went ahead and swapped over to the .67 ohm resistor that was on the emitter. Now that I’ve swapped over to the new trimmers, I went ahead and set them for 820 ohms. That’s going to get me a lot closer to the 13 mV target. I will check for crossover distortion at 13, but maybe I should just set it for 15? Anything over the minimum for eliminating crossover distortion is just going to be wasted extra heat, correct? Once you eliminate that crossover distortion that is where the idle current should be set, correct?
Dan
No, but 10 amp peaks would be possible (40 volts of swing, until the transformer sags) with 4 ohm load. 15 amp hometaxial types had better beta hold up at 10 or 15A than 15 amp epitaxial types. Even MJ21194 is flat to 8 amps, then falls like a stone and is no better than 15 at 15A. A 20 or 30 amp epitaxial type will have better gain above 10A. Which is why higher current types were suggested. You might not need or want 15 or 20 amps, but relative linearity matters. And the original 40409 and 40410 don’t like being overloaded having to drive crazy high base currents. 2N3055’s will kill them because they’re so hungry.
13 mA, 15 mA, whats the difference?
Attachments
Well holy crap, dialing these in is definitely not linear. Even with these 25 turn pots 1/10 of a full turn will take it from 20 mV to 10 mV. It is very touchy. The smallest amount of resistance will change the bias fairly dramatically. I don’t think substituting in a resistor will work very well. Let’s say I get the trimmer to 940 ohms, if I were to put in a 910 or 1k that difference will make a large difference in bias current. Really, I am quite shocked by how touchy this is on a 25 turn trimmer.
Dan
Dan
Dynaco said aim for 20 ma idle current. ST120 had a problem not achieving that cold and low wattage, had a reputation as "world's worse amplifier". Djoffe designed a closed loop circuit add on for ST120 that ensured 20 ma all the time, which made my ST120 sound a whole lot better cold & low wattage. I don't have a scope, so I just go with the 20 ma standard.
13 mv or 15 mv, not ma. Divide by 67 ohm emitter resistor. A little high idle current is better than a little low, IMHO. I had fans on the heat sinks of ST120. when the sensor transistor on the djoffe circuit blew b-e junction, it ran 200 ma idle bias for a couple of years. 200 ma idle bias sounds fine but produced some excess heat. With fans, that's okay, without, not really. I abandoned the ST120 because a purchase of another organ meant I was going to need 18' long RCA cables to keep the amp behind the record rack to hide the fan noise. I was already having AM radio pickup problems with 12' long RCA cables, I had to increase the ST120 input filter capacitor just for those. I bought a Peavey M-2600 that has huge heatsinks and no fans, sits on top of the new organ and only thing running 12' to 18' are speaker cables. No more AM radio problem. No more cold & low wattage problem.
saabracer, you may find idle bias current changes as the amp heats up. Set pot cold, blast the amp with a hair dryer for a while, see what the emitter voltage reads now? Ensure not below 20 ma (13 mv) hot or cold.
Last edited:
So you’re at 20 some odd mA. And the debate continues as to whether 20, 26, or 30 mA is the optimum class B bias current - without resolution. Can’t tell without instrumentation which is more sensitive than human hearing.
200 mA is too high, even for mosftets. The ST120 was a POS, anyway.
200 mA is too high, even for mosftets. The ST120 was a POS, anyway.
I think I’m gonna dial it into 15 mV. There is zero possibility of getting it set there with a standard resistor so yes, 100% I am going to leave the trimmers. Unfortunately, I don’t have a whole lot of Schottky diodes on hand. I have a bunch of these which are of course intended for surface mount. Though I’m sure I could get them to work by either putting legs on them or I could just solder them between the two trimmer legs. They are DL5819, 40V 1A. This should work just fine, correct?
Out of curiosity, how would a germanium do here? I have quite a large collection of different germaniums. I’m just asking about them as they have a lower forward voltage like schottky vs the silicon. I have no idea if they would work, which is why I’m asking. I grabbed several of the Russian options from a seller in Ukraine. I also have some of the American options like 1N34A
Edit: I forgot to ask, after I set the schottky diode across the trimmer it would be wise to set bias again? How much affect will that have on the bias current?
Dan
Out of curiosity, how would a germanium do here? I have quite a large collection of different germaniums. I’m just asking about them as they have a lower forward voltage like schottky vs the silicon. I have no idea if they would work, which is why I’m asking. I grabbed several of the Russian options from a seller in Ukraine. I also have some of the American options like 1N34A
Edit: I forgot to ask, after I set the schottky diode across the trimmer it would be wise to set bias again? How much affect will that have on the bias current?
Dan
Last edited: