I can feel your stress..
Keep monitoring that bias. Is the V+ now at a steady + and - 56V?
Hugo
Its always compromise. The bias generator transistor and R3306 (which is blurry but looks like 1 ohm) and which modifies the temp vs voltage characteristic of the generator is very difficult to get exactly right. If the output transistors pass a current peak (with the music) they heat up quickly (the die in the transistor) but the vbe multiplier just gets to see an average temperature of the heatsink. In other words it can not track exactly what is going on. If the bias goes down as it heats up then the vbe multiplier is over compensating but in practice it is not an issue at all.
The best you can do is set the bias when its all uniformly warm/hot and forget it.
The circuit shows 'something' in parallel with that 1 ohm. Perhaps a wire link in some versions.
Nope! All seems to be good. I tried with 50hz up to 10khz i did not find any oscillations. I also tried with a very small 0.1mv input signal to see if i could catch any crossover distortion, did not find it. The load was my sacrificial speaker which survived this test.
My ears though not so much from all the tone testing.
The only test that remains but i don't have th courage to do yet is hook it up properly to speakers and listen to my everyday music.
Last time it only lastd a few days but i had not set the bias properly. i've left it to 17.5mv on average for both channels after 90 minutes of it idling and being warm.
Time will tell i guess.. i'm still stressed for your information, and i have the new parts arriving from mouser on friday. I don't know if i will install them after all.. I have to think about it.. This amp has got me terrified! hahahaha
My ears though not so much from all the tone testing.
The only test that remains but i don't have th courage to do yet is hook it up properly to speakers and listen to my everyday music.
Last time it only lastd a few days but i had not set the bias properly. i've left it to 17.5mv on average for both channels after 90 minutes of it idling and being warm.
Time will tell i guess.. i'm still stressed for your information, and i have the new parts arriving from mouser on friday. I don't know if i will install them after all.. I have to think about it.. This amp has got me terrified! hahahaha
Thanks!
Quick question: I adjusted bias on both channels measuring only one of the two emitter resistors per channel.
This has 2 pairs of output transistors in parallel and 2 x 2x0.18ohm emitter resistors in series for each output pair if you look at the schematic. That's ok right? I did not measure the other emitter resistor on each channel.
Also since the bias tends to go down as this thing heats up wouldn't that mean that in a cold start condition the bias will be very high?
I had to slightly increase the bias trimmer every 10-15 minutes until it stabilized at around 17.5mv when it got warm enough. So the bias value will be higher at first
Quick question: I adjusted bias on both channels measuring only one of the two emitter resistors per channel.
This has 2 pairs of output transistors in parallel and 2 x 2x0.18ohm emitter resistors in series for each output pair if you look at the schematic. That's ok right? I did not measure the other emitter resistor on each channel.
Also since the bias tends to go down as this thing heats up wouldn't that mean that in a cold start condition the bias will be very high?
I had to slightly increase the bias trimmer every 10-15 minutes until it stabilized at around 17.5mv when it got warm enough. So the bias value will be higher at first
As I mentioned earlier
all of this is a compromise.It doesn't matter which pair you use to do the measurement but unless the transistors are identical the readings will differ. That's normal. You could perhaps set it on the pair with the highest reading if there was a large difference but there shouldn't be any need if the transistors are all reasonably close matched.
Yes, it could well be high when cold but that doesn't matter. It won't be excessively high and will fall quickly. Even if it is double the value it is still a low current in the scheme of things.
I've shelved the amp until the transistors from mouser arrived. They are here!!! I will replace the ones it currently has with the ones from mouser just to make sure.
I have a question / remark about the design on the amplifier. On Mooly's simulation on LTspice after running a few experiments i realized that if the bias transistor goes open either through a cold joint on the board or the transistor fails the bias skyrockets from a few milivolts to about 5 volts. here is an image showing that (let's assume the base goes open circuit):
I noticed when i first got the amp that it had a lot of cold solder joints. I'm thinking maybe i missed some and that caused the bias to skyrocket intemmintently and maybe that was the original cause that fried the outputs.
Of course in mooly's simulation there are some components missing compared to the actual design of the amplifier.
What do you think about this theory? Could that be a valid reason of the repeated failures? A cold solder joint on that transistor?
I have a question / remark about the design on the amplifier. On Mooly's simulation on LTspice after running a few experiments i realized that if the bias transistor goes open either through a cold joint on the board or the transistor fails the bias skyrockets from a few milivolts to about 5 volts. here is an image showing that (let's assume the base goes open circuit):
I noticed when i first got the amp that it had a lot of cold solder joints. I'm thinking maybe i missed some and that caused the bias to skyrocket intemmintently and maybe that was the original cause that fried the outputs.
Of course in mooly's simulation there are some components missing compared to the actual design of the amplifier.
What do you think about this theory? Could that be a valid reason of the repeated failures? A cold solder joint on that transistor?
I installed the new transistors from mouser 2sc5200 & 2sa1943. I was able ro get the bias stable at 20mv after letting it warm up for 45 minutes. Here are some pictures:
I have it playing for about 20 minutes and everything seems to be ok! I hope i'm done with this and the repair was successful.
Thank you all for taking the time and sharing your knowledge!!
I have it playing for about 20 minutes and everything seems to be ok! I hope i'm done with this and the repair was successful.
Thank you all for taking the time and sharing your knowledge!!
Ok, this is getting boring...
Now even though i have about 16-21.5mv bias on all 4 channels i notice that the right channel is more "quiet" than the left.
That occurs on both speaker a + b on the right channel.
This channel is also the one that was repaired multiple times.
What do i check? should i raise the bias more on that one? This channel has 16.5mv bias on speaker A and 19mv on speaker B, while on the left it's about 19.5-20mv on both A+ B. Both readings are taken after it warms up a bit. Could it be that on the "quiet" channel the bias transistor overcompensates and drops the bias really low as i noticed before replacing the new output transistors?
Both have ~2mv Dc Offset.
Now even though i have about 16-21.5mv bias on all 4 channels i notice that the right channel is more "quiet" than the left.
That occurs on both speaker a + b on the right channel.
This channel is also the one that was repaired multiple times.
What do i check? should i raise the bias more on that one? This channel has 16.5mv bias on speaker A and 19mv on speaker B, while on the left it's about 19.5-20mv on both A+ B. Both readings are taken after it warms up a bit. Could it be that on the "quiet" channel the bias transistor overcompensates and drops the bias really low as i noticed before replacing the new output transistors?
Both have ~2mv Dc Offset.
Definitely not
As a general statement bias does not affect the gain (we'll ignore what happens at the micro level).
Two things... easy one first... the volume control is mismatched between the two gangs which is common across all makes and models.
Secondly, the gain is set by the ratio of the two feedback determining resistors. Nothing else.
The first test would be to look at the input to each amp and see if they are the same or not.
1. Ok i can quickly check that by removing the rca bridge in the back. I'm lucky because this amp has the option to run it through an external preamp.
2. The feedback resistors have been replaced. they are both new 27k resistors and if i'm not mistaken there are two 1k resistors as well on each channel.
ohhh i haven't check that resistor over there the 1k5. will do! Thanks, i will return!
Unless a mistake has been made replacing those resistors its pretty much unheard of for them to be faulty.
Lets see if the theory holds...
Vin is 0.141 and Vout is 2.777. The gains is Vout/Vin which is 19.70
The feedback ratio is (27k + 1k)/1k5 + 1 which is 19.666
There are rounding errors in the calculation but you can see how the theory works. No changes to other parts including the transistors will affect the gain. Neither will the bias.
Lets see if the theory holds...
Vin is 0.141 and Vout is 2.777. The gains is Vout/Vin which is 19.70
The feedback ratio is (27k + 1k)/1k5 + 1 which is 19.666
There are rounding errors in the calculation but you can see how the theory works. No changes to other parts including the transistors will affect the gain. Neither will the bias.
Checked the feedback resistors and everything is the same on both channels. So i bypassed the preamp and fed signal straight to the amp. Same result.
But after feeding a 500hz signal and performing a signal trace on every stage with my oscillisope i found that the signal levels are almost the same up to one end of the relays. Meaning it reaches the one end of the relay perfectly. When the relay's are on though the output from them varies a lot. whic htells me that there has propably been sparks inside causing them to develop resistance (i guess).
The problem is on both sides. out of 4 channels:
But after feeding a 500hz signal and performing a signal trace on every stage with my oscillisope i found that the signal levels are almost the same up to one end of the relays. Meaning it reaches the one end of the relay perfectly. When the relay's are on though the output from them varies a lot. whic htells me that there has propably been sparks inside causing them to develop resistance (i guess).
The problem is on both sides. out of 4 channels:
- On speaker a : left has NO OUTPUT at all // right is low and sometimes cutting out
- On speaker b: both have output but with volume difference