Hi Everyone and Merry Christmas,
I have a vintage Sony STR-6045 amplifier that I managed to restore with the help of the wonderful users of this forum. There was one last thing to do on it: to reduce the dc offset on the left channel that was 95 mV. The task was fairly simple as it was involving replacing Q701 and Q702 with a matched pair of transistors. In the process, I managed to break the legs of other transistors (Q756 and Q703). Since they were old and noisy, I decided to change all those transistors type (4 2SA706) and (3 2SC1124) with Toshiba TTA004B and TTC004B.
Stupid me, I misread the pinout and created a good light show with my bulb tester.
Since then, I have re-soldered correctly the seven transistors and get no more light but I have lost the left channel. So far:
I am at lost as to what I could have missed. Would someone with experience understanding circuits could suggest what else to verify? Voltages in BLUE are actual and in RED are expected.
As always, thanks a lot for your help guys.
I have a vintage Sony STR-6045 amplifier that I managed to restore with the help of the wonderful users of this forum. There was one last thing to do on it: to reduce the dc offset on the left channel that was 95 mV. The task was fairly simple as it was involving replacing Q701 and Q702 with a matched pair of transistors. In the process, I managed to break the legs of other transistors (Q756 and Q703). Since they were old and noisy, I decided to change all those transistors type (4 2SA706) and (3 2SC1124) with Toshiba TTA004B and TTC004B.
Stupid me, I misread the pinout and created a good light show with my bulb tester.
Since then, I have re-soldered correctly the seven transistors and get no more light but I have lost the left channel. So far:
- Q704 was damaged and I replaced it.
- Tested D702 (a varistor?), it has the same resistance and voltage drop as the right channel.
- Tested on board and out of the board a few transistors with diode tester function.
- Tested RT701 (the variable resistor that is used to set the 25mV bias). Resistance does change when turning it but no more voltage on R721-722.
- I have highlighted in yellow the voltages that are far from expectation. 30.7V at the emitter of Q757 being the one the most suspect, but I think the schematic is wrong since the right channel is working.
I am at lost as to what I could have missed. Would someone with experience understanding circuits could suggest what else to verify? Voltages in BLUE are actual and in RED are expected.
As always, thanks a lot for your help guys.
What did you replace Q704 with?
D702 is a diode.
If the VBE of Q704 (and others you have replaced), is higher than the original, you will get no quiescent current flow without adjustment of RT701 and it might not adjust far enough.
If D702 is damaged it will drop the wrong voltage again skewing your readings.
Across C706 you are looking for 1.8volts. If you have that, there is an issue with the rest that you have changed.
I would measure and compare left channel with right channel.
It's only an amplifier, not complicated stuff.
D702 is a diode.
If the VBE of Q704 (and others you have replaced), is higher than the original, you will get no quiescent current flow without adjustment of RT701 and it might not adjust far enough.
If D702 is damaged it will drop the wrong voltage again skewing your readings.
Across C706 you are looking for 1.8volts. If you have that, there is an issue with the rest that you have changed.
I would measure and compare left channel with right channel.
It's only an amplifier, not complicated stuff.
Thank you for your response.
Here is how it looks like:
All the values that I have showed in yellow look good on the right channel, so I'm not sure what is wrong with the left.
I did use a KSC1815Y which has a Vbe of 5 V. It replaces 2SC633A which has a max Vbe of 6 V.
The service manual calls this a varistor SV-31 and explains the following "The negative temperature coefficient of D702 provides thermal compensation for the complementary power and transistor circuits. D702 is attached to the power transistor’s heat sink to detect temperature increases in the power transistor."
I measured it like a diode. It drops about 0.159 Volt like the right channel and reads 1.78 in the other direction just like the one in the right channel.
Here is how it looks like:
You are right, I have 1.8 volt on the right channel and almost nothing on the left (0.07 volt).
All the values that I have showed in yellow look good on the right channel, so I'm not sure what is wrong with the left.
To drop a voltage across C706 we must have current drawn through R710.
Look for the voltage across the preset.
The VBE is the voltage between the Base and Emitter, (V oltage B ase E mitter) not the reverse voltage, (Maximum breakdown reverse voltage) the base - emitter voltage of a silicon transistor cannot exceed 0.7volts before damage ... but I think you have a man made fault.
Check your work.
Look for the voltage across the preset.
The VBE is the voltage between the Base and Emitter, (V oltage B ase E mitter) not the reverse voltage, (Maximum breakdown reverse voltage) the base - emitter voltage of a silicon transistor cannot exceed 0.7volts before damage ... but I think you have a man made fault.
Check your work.
Last edited:
Well, I have been checking this circuit from quite some times. I still cannot get any voltage at the bias adjustment even if I turn the screw all the way counterclockwise:
If by VBE you mean the voltage drop you get when testing in diode mode with a Fluke, I get:
Left (non working) Right (working)
Q704 (KSC1845) VBE=0.645 Q754 (2SC633A) VBE=0.654
Q701 (KSC1845) VBE=0.655 Q751 (2SC632A) VBE=0.659
Q702 (KSC1845) VBE=0.655 Q752 (2SC632A) VBE=0.652
I also tried swapping the varistors D702 and D752. Same result.
I tried swapping Q704 and Q754, the left channel still dead and right channel adjustment had to be turned down a little to get to 25 mV.
Now, I forgot to mentioned that an eyelet is gone at the collector of Q701 and that the leg was soldered directly to the track. It was working like this before I start messing with it. I made a better jumper with a wire but same result. Does the wire gauge matters for such a short run?
To resolve this, I really need to understand what will make the current flow in this channel.
Thanks for all your suggestions.
If by VBE you mean the voltage drop you get when testing in diode mode with a Fluke, I get:
Left (non working) Right (working)
Q704 (KSC1845) VBE=0.645 Q754 (2SC633A) VBE=0.654
Q701 (KSC1845) VBE=0.655 Q751 (2SC632A) VBE=0.659
Q702 (KSC1845) VBE=0.655 Q752 (2SC632A) VBE=0.652
I also tried swapping the varistors D702 and D752. Same result.
I tried swapping Q704 and Q754, the left channel still dead and right channel adjustment had to be turned down a little to get to 25 mV.
Now, I forgot to mentioned that an eyelet is gone at the collector of Q701 and that the leg was soldered directly to the track. It was working like this before I start messing with it. I made a better jumper with a wire but same result. Does the wire gauge matters for such a short run?
To resolve this, I really need to understand what will make the current flow in this channel.
Thanks for all your suggestions.
Have you confirmed C706 is not shorted, it might have been collateral to your previous accident. The amplifier cannot be biased if this cap is shorted.
Thank you Kevin. I know how difficult it can be to help from afar. Much appreciated.
I tried to measure the voltages again and I managed to create a short. I took a few days off and realized that I blew one of the output transistors (Q901). I had a spared one and replaced it. It is now back to its state were I can't get any voltage in the bias.
I'm now a bit nervous now to put test leads on this board. 🤔
I just tested R713. It is 326 ohms and it has a voltage of 0.57 Volt across it. The good channel has 1.2 Volt.
In the process of finding this short, I tested so many components and found nothing. I also put back the original Sony Q701 and Q702 with same result.
This is really a mystery to me.
I tried to measure the voltages again and I managed to create a short. I took a few days off and realized that I blew one of the output transistors (Q901). I had a spared one and replaced it. It is now back to its state were I can't get any voltage in the bias.
I'm now a bit nervous now to put test leads on this board. 🤔
I just tested R713. It is 326 ohms and it has a voltage of 0.57 Volt across it. The good channel has 1.2 Volt.
In the process of finding this short, I tested so many components and found nothing. I also put back the original Sony Q701 and Q702 with same result.
This is really a mystery to me.
Definitely get some proper clip leads which should help to avoid additional shorts in the future.
Something like this will save you some grief going forward, and I often make the connections with the circuit unpowered for extra safety
https://www.amazon.com/Pomona-3782-36-02-MINIGRABBER-Banana-Plug/dp/B005T63B9C/
There are lots of other options, but you want something that is well insulated in use, crocodile clips are not your friend as you have found.
My next bit of advice is without power applied measure every single circuit node resistance and test each transistor with your multimeter to determine if it is still a transistor. (You do not need to remove to test.) 😀 Do the same for the good board and compare , there might be hint we can figure out. Record each detail no matter how seemingly trivial it might appear.
Something like this will save you some grief going forward, and I often make the connections with the circuit unpowered for extra safety
https://www.amazon.com/Pomona-3782-36-02-MINIGRABBER-Banana-Plug/dp/B005T63B9C/
There are lots of other options, but you want something that is well insulated in use, crocodile clips are not your friend as you have found.
My next bit of advice is without power applied measure every single circuit node resistance and test each transistor with your multimeter to determine if it is still a transistor. (You do not need to remove to test.) 😀 Do the same for the good board and compare , there might be hint we can figure out. Record each detail no matter how seemingly trivial it might appear.
I'm late to this thread, but you're seeing -3.3V at emitter of Q707 ???
It should be about -30V. The circuit is an emitter follower-filter on the negative supply. If the supply is at -3V, there's very little current available to drive the bias spreader. Have I missed something earlier?
It should be about -30V. The circuit is an emitter follower-filter on the negative supply. If the supply is at -3V, there's very little current available to drive the bias spreader. Have I missed something earlier?
No, expected voltages are in red and measured ones are in blue. I'm seeing 29.1 V at the emitter of Q707. Were you rather talking about Q757 where it shows -3.3 Volt? I was questioning this too originally:
Out of curiosity, what is an emitter follower filter?
Out of curiosity, what is an emitter follower filter?
Yes, typo error on my part. Expected -3.3V at Q757 emitter seems incorrect and -30.7V looks reasonable.
My description "emitter follower-filter" was clumsy. IIRC, they are often called "capacitance mutipliers" but implement a low-pass filter for supply hum/noise suppression.
My description "emitter follower-filter" was clumsy. IIRC, they are often called "capacitance mutipliers" but implement a low-pass filter for supply hum/noise suppression.
May I offer some observations and some constructive advice.
Several of the measured voltages noted the first schematic don't correlate in a rational way or are inconsistent with textual data. For example, the schematic indicates 2.6V at collector of Q703 and +0.7V at anode of D702. This suggests 1.8V across C706--- actually an encouraging voltage. But elsewhere near the end of post 3, you indicate only 0.07V across C706, so these data conflict.
You mention:
Tip: I often tack-solder lengths of hookup wire to points of interest so that I can bring them safely clear of the circuit and anchor them securely; then I can probe among a variety of points without accidents.
Several of the measured voltages noted the first schematic don't correlate in a rational way or are inconsistent with textual data. For example, the schematic indicates 2.6V at collector of Q703 and +0.7V at anode of D702. This suggests 1.8V across C706--- actually an encouraging voltage. But elsewhere near the end of post 3, you indicate only 0.07V across C706, so these data conflict.
I suspect these voltages may be with respect to ground rather than "across" the R713 leads? Any any case, I would expect only a few mV across R713 terminals. Try to be alert to description of readings and to data that doesn't make sense. I make mistakes too, eg my Q707 typo.
You mention:
I suspect a defect in the output stages, but does the voltage between collector of Q703 and anode of D702 (= voltage across C706) respond to the bias trimmer?
Tip: I often tack-solder lengths of hookup wire to points of interest so that I can bring them safely clear of the circuit and anchor them securely; then I can probe among a variety of points without accidents.
Last edited:
BSST, you just pointed out my stupidity. Yes, I measured these voltages with respect to the ground. Yes, I had roughly 0.57V on both sides of R713 so only a few mVs.
After all those changes I made (soldering/desoldering of caps and transistors), I am due for a full review. I will take your advices and those of Kevin and do it safely. It may take a few days as I need to progress another project.
Stay tuned. I really appreciate your help guys!
After all those changes I made (soldering/desoldering of caps and transistors), I am due for a full review. I will take your advices and those of Kevin and do it safely. It may take a few days as I need to progress another project.
Stay tuned. I really appreciate your help guys!
Here is an update to this puzzle and a few questions:
I have done a complete measurements of the components (no power, in circuit only - I have measured the capacitors for shorts):
I then started to measure the voltages from ground with the amp powered on and I lost the right channel!
I was very careful to not short anything and I'm using a dim bulb tester so that's mysterious. Any way, I will find what I have damaged again.
My questions/observations:
I have done a complete measurements of the components (no power, in circuit only - I have measured the capacitors for shorts):
| | Left Channel (Bad) | | Right Channel (Good) | |||
Original Component | Nom. Value | Component | Measured | Comment | Component | Measured | Comment |
Carbon +-5% 1/4 Watts | 47K | R701 | 46 | | R751 | 46 | |
| 1,5K | R702 | 1,488 | | R752 | 1,489 | |
| 3,3K | R703 | 3,31 | | R753 | 3,32 | |
| 27K | R704 | 27,28 | | R754 | 27,15 | |
| 3,3K | R705 | 3,33 | | R755 | 3,37 | |
| 390 | R706 | 386 | | R756 | 383 | |
| 3,3K | R707 | 3,33 | | R757 | 3,34 | |
| 7,5K | R708 | 7,5 | | R758 | 7,43 | |
| 47K | R709 | 46,2 | So long to measure. Why? | R759 | 46,2 | So long to measure. Why? |
Composition +-10% 1/2W | 5,6K | R710 | 5,59 | | R760 | 5,79 | |
Composition +-10% 1/2W | 10 | R711 | 11,4 | | R761 | 11,1 | |
| 330 | R712 | 328 | | R762 | 323 | |
| 330 | R713 | 326 | | R763 | 323 | |
| 10 | R714 | 10,2 | | R764 | 9,9 | |
| 10 | R715 | 9,8 | | R765 | 10 | |
| 5,6 | R716 | 5,5 | | R766 | 5,7 | |
| 5,6 | R717 | 5,6 | | R767 | 5,6 | |
| 10 | R718 | 9,8 | | R768 | 10,3 | |
| 220 | R719 | 215,8 | | R769 | 217,9 | |
| 220 | R720 | 217 | | R770 | 217,5 | |
Wire-wound +-10% 2W | 0,47 | R721 | 0,5 | | R771 | 0,5 | |
Wire-wound +-10% 2W | 0,47 | R722 | 0,5 | | R772 | 0,5 | |
| 1,5K | R723 | 1,49 | | R773 | | Non existing? |
| 100 | R724 | 98,6 | | R774 | 98,3 | |
| 12K | R725 | 22,3 | | R775 | 21,6 | |
Composition +-10% 1/2W | 470 | R726 | 498 | | R776 | 524 | |
Electrolytic +100 -10% 50V | 1 | C701 | OL | | C751 | OL | |
Electrolytic +100 -10% 6,3V | 47 | C702 | OL | | C752 | OL | |
Electrolytic +150 -10% 10V | 220 | C703 | 0,157 | Makes a bip | C753 | 0,155 | Makes a bip |
Ceramic +-10% 50V | 47p | C704 | 1,5 | Measure 46,5p out of circuit | C754 | OL | |
Mylar +-10% 50V | 0,033 | C705 | 1 | Makes a bip | C755 | 1 | Makes a bip |
Electrolytic +100 -10% 16V | 47 | C706 | 1,323 | | C756 | 1,79 | |
Electrolytic +100 -10% 50V | 100 | C707 | 0,795 | | C757 | 1,2 | |
Electrolytic +100 -10% 50V | 100 | C708 | 1,15 | | C758 | 2 | Makes a bip |
Diode | | D701 | 0532-OL | | D751 | 0,541-OL | |
Varistor | | D702 | 1,7-0,155 | | D752 | 1,8-0,157 | |
2SC632A (NPN Ic=0,2A, Vcb=50V, Vce=50V, Veb=6V, hFE=120) | | Q701 | Vbe=0,658, Vbc=0,642, Vother=OL | | Q751 | Vbe=0,658, Vbc=0,642, Vother=OL | |
2SC632A (NPN Ic=0,2A, Vcb=50V, Vce=50V, Veb=6V, hFE=120) | | Q702 | Vbe=0,652, Vbc=0,642, Vother=OL | | Q752 | Vbe=0,652, Vbc=0,642, Vother=OL | |
2SA706 (PNP Ic=1A, Vcb=60V, Vce=60V, Veb=6V, hFE=70) | | Q703 | Veb=0,602, vcb=0,603, Vother=OL | | Q753 | Veb=0,605, vcb=0,603 | |
2SC633A (NPN Ic=0,2A, Vcb=25V, Vce=25V, Veb=6V, hFE=65) | | Q704 | Vbe=0,649, vbc=0,641 | | Q754 | Vbe=0,650, vbc=0,640 | |
2SC1124 (NPN Ic=1A, Vcb=160V, Vce=140V, Veb=5V, hFE=50) | | Q705 | Vbe=0,613, Vbc=0,611, Vother=OL | | Q755 | Vbe=0610, Vbc=0,607 | Vcb=0,640 and Veb=1,99 |
2SA706 (PNP Ic=1A, Vcb=60V, Vce=60V, Veb=6V, hFE=70) | | Q706 | Veb=0,608, Vcb=0,604 | | Q756 | | |
2SC1124 (NPN Ic=1A, Vcb=160V, Vce=140V, Veb=5V, hFE=50) | | Q707 | Vbe=0,611, Vbc=0,608 | | | | |
2SA706 (PNP Ic=1A, Vcb=60V, Vce=60V, Veb=6V, hFE=70) | | | | | Q757 | Veb=0,609, Vcb=0,605 | |
2SD290 (NPN, Ic=5A, Vcb=80V, Vce=60V, Veb=6V, hFE=60) | | Q901 | Vbe=0,105, Vbc=0,477 | All good out of circuit | Q951 | Vbe=0,105, Vcb=0,450 | All good out of circuit |
2SD290 (NPN, Ic=5A, Vcb=80V, Vce=60V, Veb=6V, hFE=60) | | Q902 | Vbe=0,105, Vbc=0,467 | All good out of circuit | Q952 | Vbe=0,105, Vcb=0,426 | All good out of circuit |
I then started to measure the voltages from ground with the amp powered on and I lost the right channel!
I was very careful to not short anything and I'm using a dim bulb tester so that's mysterious. Any way, I will find what I have damaged again.
My questions/observations:
- Some capacitors show low continuity. I assume because they are in circuit. However, they have different values between channels. Do you see anything significant there?
- I have found the supplement and the correction to the service manual. This supplement also indicates -3.3V at the emitter of Q757. I'm starting to think this is the critical information that we have omitted or should not have discarded.
What components would affect this value?
I'm quite certain that -3.3V is a drafting error (intent was -33V) that was not noticed even in supplement/correction. Follow that supply trace into the amp and you'll find it drives R704 and R710. R704 is approximating a constant current source and R710 needs large voltage in order to drive the output stage.
| 47K | R709 | 46,2 | So long to measure. Why? |
When you first apply your ohmmeter, I believe you're measuring R706 via C702; as C702 charges, you are able to observe R709.
| 1,5K | R723 | 1,49 | R773 | Non existing? |
I could't spot R723 in the schematic. Would you guide me, please?
Thanks.