From voltage reported on Q601 base, transistor is almost certainly failed, exhibiting low Beta. Rough guesses suggest Beta is about 38 but data sheet claims at least 160. Get the pair replaced and things should improve. While transistor is out, confirm R605, R607 resistances.
Good luck.
Good luck.
Perhaps it's the 555 II that has the decimal-place error, and should be 4.7uF not 470uF. Anyways, it doesn't seem to cause 555 II's to have any weird DC offsets on startup, they settle out pretty quick.
I've taken the recommendation out of the ADCOM BOMs doc until I can investigate.
If the supply rise is delayed, your DC servo will not function for a bit. Meanwhile your DC offset would probably be whatever it is with the servo chip pulled if the op amp output is high impedance.
I hate this design with the diff pair unbalanced. Aside from DC offset that is entirely avoidable, an unbalanced pair doesn't subtract properly. That increases distortion. What you want more than anything is for the diff pair to be balanced and each have the same transconductance. Distortion will be minimum then. Personally, I'd rather a DC blocking cap in the ground leg of the feedback path and a balanced diff pair. DC offset should then sit within 50 mV max, and that is much lower than when amp starts up. We could have sub 5 mV offsets without servos in the 1970's and 1980s before servos.
In other words, there is no technical reason for a servo IC if the designer does their job and the parts are even semi-matched.
I hate this design with the diff pair unbalanced. Aside from DC offset that is entirely avoidable, an unbalanced pair doesn't subtract properly. That increases distortion. What you want more than anything is for the diff pair to be balanced and each have the same transconductance. Distortion will be minimum then. Personally, I'd rather a DC blocking cap in the ground leg of the feedback path and a balanced diff pair. DC offset should then sit within 50 mV max, and that is much lower than when amp starts up. We could have sub 5 mV offsets without servos in the 1970's and 1980s before servos.
In other words, there is no technical reason for a servo IC if the designer does their job and the parts are even semi-matched.
Hold on. That's engineering.
You've got to understand exactly how the entire circuit works before making changes. Often the designer/engineer is forced to do some things due to other circuit characteristics or even voltage levels. And the other massive thing to consider is that you do not have the instrumentation or experience. Just because you get something to work doesn't mean it works well, or properly.
Get it working as designed. Once you have experience and engineering knowledge you can examine the circuit and make some intelligent changes. Even high frequency stability may change with some simple changes.
You've got to understand exactly how the entire circuit works before making changes. Often the designer/engineer is forced to do some things due to other circuit characteristics or even voltage levels. And the other massive thing to consider is that you do not have the instrumentation or experience. Just because you get something to work doesn't mean it works well, or properly.
Get it working as designed. Once you have experience and engineering knowledge you can examine the circuit and make some intelligent changes. Even high frequency stability may change with some simple changes.
Agreed.
A few years ago I built a 25 WPC solid state amp kit which I wasn't too satisfied with as the heatsinks would get quite hot at anything close to its maximum power out and it had a current limiter as well which I didn't like. With the help of a switch mode power supply designer in a DIY solid state amp group on Facebook it was modified to make 93WPC into 4 ohms and would do so all day long no problem. I learned a lot about solid state amps, however that circuit is not the circuit Adcom used so it at best provided a general knowledge of how solid state amps with dual supplies work.
That said, had it not been for the servo, I might have had an easier time troubleshooting the Adcom amp.
Now if this was some lower end console stereo amp from the 70's, I'd have felt more like modding it as it would have been no great loss if it was screwed up.
A few years ago I built a 25 WPC solid state amp kit which I wasn't too satisfied with as the heatsinks would get quite hot at anything close to its maximum power out and it had a current limiter as well which I didn't like. With the help of a switch mode power supply designer in a DIY solid state amp group on Facebook it was modified to make 93WPC into 4 ohms and would do so all day long no problem. I learned a lot about solid state amps, however that circuit is not the circuit Adcom used so it at best provided a general knowledge of how solid state amps with dual supplies work.
That said, had it not been for the servo, I might have had an easier time troubleshooting the Adcom amp.
Now if this was some lower end console stereo amp from the 70's, I'd have felt more like modding it as it would have been no great loss if it was screwed up.
Yes, but honestly .... Consumers are completely unconscious to technical aspects beyond "runs hot", and my speakers go bang. I get enough concerned comments from customers about the woofers moving on Adcom product, they finally changed it. Beyond that, most consumers only notice smoke, odd noises, when their stuff is badly distorted or doesn't make sound at all. But heavens! If their stuff is new and has the smallest mark on it they want a replacement. Which is completely fair.
This is an interesting project, so I feel compelled to comment.
I agree that a DC servo tends to hide some of the sins of poor differential-pair matching and it inherently delivers low DC error at the speaker terminals. And offers a glitzy marketing claim.
One additional advantage of the servo is that it avoids the need for an electrolytic cap in the feedback divider which allows unity gain at DC. For example, a non-servo design might use a 33uF cap in series with R611 (1K), giving a high pass corner at 4.8Hz. But these caps are accused of being a distortion mechanism and thus an incentive for DC servo. (My suspicion is that these fears are overstated.). Given that the servo is already in place, I'd be inclined to retain it.
If the amp were mine, I would repair/upgrade as follows: I'd repair to get the amp back to as-designed state, and then try some improvements. I'd add a -12V supply to the servo opamp, incorporate a 100K pot from +12V to ground, and add a 2nd 1.5M from pot wiper to base of Q1. Use the pot to trim nominal servo voltage to 0V. You may able to make the servo a bit faster by adjusting R687 and R689. C623 should be a polypropylene film cap. The idea is to make turn-on transient small and tolerably fast. Maybe the turn-on thump can be made acceptable.
Another area I believe deserves attention is loading of Q603 so that there are approximately balanced currents in Q601 and Q603. If the values of R615 and R619 are accurate, their respective voltage drops can be used to infer the current in Q603. I'm also deeply distrustful of Q607 whose b-e junction would appear to be a non-linear load across R617, a possible distortion maker. My instinct is to remove Q607 until the the issues of Q601/603 matching are resolved. R617 value needs some analysis if Q603 current is changed--- there could be change in loop gain. Finally, I would consider connecting Q607 base to collector of Q601 so that any distortion won't be in the signal path. R615 value will likely need changing.
I agree that a DC servo tends to hide some of the sins of poor differential-pair matching and it inherently delivers low DC error at the speaker terminals. And offers a glitzy marketing claim.
One additional advantage of the servo is that it avoids the need for an electrolytic cap in the feedback divider which allows unity gain at DC. For example, a non-servo design might use a 33uF cap in series with R611 (1K), giving a high pass corner at 4.8Hz. But these caps are accused of being a distortion mechanism and thus an incentive for DC servo. (My suspicion is that these fears are overstated.). Given that the servo is already in place, I'd be inclined to retain it.
If the amp were mine, I would repair/upgrade as follows: I'd repair to get the amp back to as-designed state, and then try some improvements. I'd add a -12V supply to the servo opamp, incorporate a 100K pot from +12V to ground, and add a 2nd 1.5M from pot wiper to base of Q1. Use the pot to trim nominal servo voltage to 0V. You may able to make the servo a bit faster by adjusting R687 and R689. C623 should be a polypropylene film cap. The idea is to make turn-on transient small and tolerably fast. Maybe the turn-on thump can be made acceptable.
Another area I believe deserves attention is loading of Q603 so that there are approximately balanced currents in Q601 and Q603. If the values of R615 and R619 are accurate, their respective voltage drops can be used to infer the current in Q603. I'm also deeply distrustful of Q607 whose b-e junction would appear to be a non-linear load across R617, a possible distortion maker. My instinct is to remove Q607 until the the issues of Q601/603 matching are resolved. R617 value needs some analysis if Q603 current is changed--- there could be change in loop gain. Finally, I would consider connecting Q607 base to collector of Q601 so that any distortion won't be in the signal path. R615 value will likely need changing.
Well, if you retain the servo and provide a proper bipolar supply, then balance the diff pair as well. Otherwise there is zero benefit in making a negative supply.
The concerns over the large electrolytic are completely unjustified. There isn't any music down there to begin with for one. If there is, most speakers don't respond that low, if the do, their distortion swamps the circuit anyway. Plus, you wouldn't hear it.
The servo is very simply a way to save on matched pairs and/or manual adjustment. Period. The average consumer doesn't even understand what "DC offset" is, and what is acceptable. They don't even want to know, they just want to connect speakers without damage.
The concerns over the large electrolytic are completely unjustified. There isn't any music down there to begin with for one. If there is, most speakers don't respond that low, if the do, their distortion swamps the circuit anyway. Plus, you wouldn't hear it.
The servo is very simply a way to save on matched pairs and/or manual adjustment. Period. The average consumer doesn't even understand what "DC offset" is, and what is acceptable. They don't even want to know, they just want to connect speakers without damage.