Adjusting an old bias pot did some damage and I am unsure how to proceed

Administrator
Joined 2007
Paid Member
It would be nice of the power amp's left channel drove the left speaker and right channel drove the right speaker but I'm thinking this has no bearing on the issue(s) at hand, But I could be totally wrong about that.
The wires being connected to wrong channel on a working amp will only reverse what you hear as L and R but it also makes bias adjustment via the method in the manual meaningless because the method relies on using the speaker terminal as one of the measurement points.

There just seems to be so many man made errors with this :) that it makes it difficult to follow. Measuring voltage across any of the 0.47 ohms is the sure fire way to get the correct result.
 
Member
Joined 2021
Paid Member
if i am correct and although it isnt very clear, im sure this is that seciton of the print
so ignore this, what you have on those 2 i think are corrct.when princomes off it normaly shows a pattern of what was solder print, but i dont think it is in this instance as this is where those 2 transistors are located, and the print that has come off is both soler points and non solder areas.the diagrams are not particularly clear.
 
Member
Joined 2021
Paid Member
If you look at the attached, circled in red, to me looks like the solder pad is almost folded over and isnt connected to the component
The blue one could be connected to the other track, so i would take that off and re do the connection
 

Attachments

  • sony1.jpg
    sony1.jpg
    129 KB · Views: 14
If you look at the attached, circled in red, to me looks like the solder pad is almost folded over and isnt connected to the component
The blue one could be connected to the other track, so i would take that off and re do the connection

That circled in red solder pad had lifted but looked still intact and was soldered to the component. But to be certain I soldered in a bridge. I re-flowed the circled in blue connection and it is not in contact with that other track - no continuity.
 
The wires being connected to wrong channel on a working amp will only reverse what you hear as L and R but it also makes bias adjustment via the method in the manual meaningless because the method relies on using the speaker terminal as one of the measurement points.

There just seems to be so many man made errors with this :) that it makes it difficult to follow. Measuring voltage across any of the 0.47 ohms is the sure fire way to get the correct result.

Ain't that the truth.

To quote Monty Python...."My brain hurts."

Measuring again across all of the 0.47 ohm resistors, each on the left side group read a few millivolts. A little different than in post #196 but maybe that is because of varying lengths of time that the amp has been idling?

The right side group read 0.0mV across each one. And also as in post #196, bias is still 4.5 volts on both channels.

To avoid confusion, mostly mine, should I reverse that red and white wire on the speaker switch to what they were when I got the amp? The balance control will operate backwards but the power amp channels will be hooked to the correct speaker terminals.
 
Administrator
Joined 2007
Paid Member
Measuring again across all of the 0.47 ohm resistors, each on the left side group read a few millivolts. A little different than in post #196 but maybe that is because of varying lengths of time that the amp has been idling?
That sounds good. Yes, bias current does drift with temperature. All normal.

The right side group read 0.0mV across each one. And also as in post #196, bias is still 4.5 volts on both channels.
Bias is the voltage across the 0.47 ohm resistors :) Offset (which we are not measuring at this point) is the voltage at the speaker output as measured from ground. This should always be close to zero. If the voltage on the 0.47 ohm is 4.5 volts as measured from ground then you have DC offset problem which is unrelated to bias current.

To avoid confusion, mostly mine, should I reverse that red and white wire on the speaker switch to what they were when I got the amp? The balance control will operate backwards but the power amp channels will be hooked to the correct speaker terminals.
It sounds to me like you are measuring a DC offset on the faulty channel and that is what would happen if the speaker connections were reversed.

Given there is so much confusion over all this its best if you look at each speaker terminal feed and make sure it goes to the correct place. The output of each amplifier (the 0.47 ohm resistors) should ultimately go to the PLUS speaker terminal for that channel. The speaker negative terminal should go to ground.
 
Over 20 years ago (at the time when a friend became a member of eBay Germany, which had just been founded at the time), I received an integrated amplifier, and the owner told me that I too would not be able to get it back into working order (I think it was a NAD 3130).
In previous repair attempts, the output transistors were replaced 5 times without success (as were numerous other components on the main board around the power amplifier).
Since the service-friendly design of this model made it possible to quickly get an overview of the condition of the solder side of the main board, I decided not to start the repair until I had an other device of the same model with a power amp unit in working condition (solder site of main board around the power amp parts looks very similar to those from image in post #203.

I found a cheap model with various errors in the pre-amplifier unit (like crackled volume control) but with power amplifier in perfect working condition.

I then desoldered all components in areas where there was a soldering condition on the underside of the board as image of post #203 show.
Then I removed the solder and then the green solder mask (with very fine sandpaper). As next step I tinned all the conductor tracks and recreated the missing conductor tracks with pre-tinned copper wires. The second device with power amplifier in working condition served as a template for this.

After re-soldering the parts on the main PCB -,partly replaced with new ones - the procedure as described in post #113 (120) was carried out (additional introducing 0,47R emitter resistors on the power transistors, which were missing in the original state) and this integrated amplifier still works today without any problems.

Because a second device of this SONY-Model TA1150 isn't very hard to find, I think that the same approach leads to the goal most quickly here.
 
Last edited: