Restoring the AU-D907G Extra

[wmac]

The amplifier arrived operational but exhibited intermittent noise from oxidized controls. All switches and potentiometers were thoroughly cleaned and treated, after which the issue was completely resolved. The amplifier will then be fully restored as a preventive measure for long-term reliability.

Bellow the picture of it:

 

[frend2001]

It has a rare technical quirk or let's say trivia.

It mates the Super Feed Forward with floating ground design which later matured into x balanced amplifier line.

 

[wmac]

Yes, you are right just

It has a rare technical quirk or let's say trivia.

It mates the Super Feed Forward with floating ground design which later matured into x balanced amplifier line.

finished restoring an X1. The big issue with this one is accessing the boards; I will have to disassemble everything.

 

[wmac]

I started for easier part the F-3677 EQ that is in the top of the amp.

Picture before:



This is done already pic after




Capacitors used in this board: Rubicon and Vishay

 

[wmac]

To have access to the drivers and ouput transistors F-3679, I had to remove the heatsink:




Next step is to remove the output transistors to replace the mica

 

[wmac]

Finished working on the drivers and outputs for F-3679 as pic below:



Now start to work on the power supply F-3673; here is where the fun starts. I have to work for 2 days to remove all the glue and check all components that were affected by it

 

[wmac]

Work on the glue is boring so I decided to start and then I went to service the tone board and the speaker selectoras picutres shows.






This is where I am at the power supply board, half is done and other half I still have to clean

 

[wmac]

The power supply F-3673 is done, cleaned of that nasty glue that was very sticky, and recapped; transistors tested and cleaned.

 

[wmac]

Working on the Power amp board F-3678



The styro capacitors were removed, and all of them were tested open; they were replaced with PP Film caps 1000pF 160V



There was that nasty glue everywhere.




All Zener diodes were damaged by the glue corrosion.




Took me 2 days to clean this board; once I had to replace some components that were contaminated by the glue, remove most of the others, clean, test, and reinstall. The board is now cleaned:



F-3678 board is restored, new caps and zeners replaced, and now I am reflowing the solder and cleaning the back.





Finishing this, it is time to reinstall the board in place and start the calibration

 

[wmac]

Now boards are in place and output transistors installed with Berquist silpad






Now working on the board, F-3679 protector

 

[wmac]

Finishing with the F-3679, waiting for the delivery of the polish paste to clean the RCAs and mount everything back to start calibrations and tests.

 

[ToxicTele]

Nice work on the glue removal and restoration, kudos!!

 

[wmac]

Nice work on the glue removal and restoration, kudos!!

This glue was very difficult to remove. I used every tool that could do it; it was too sticky and took me about 4 days to remove and clean components and replace all that was contaminated

 

[wmac]

All put together, I will start to calibrate and test soon.

 

[wmac]

These are the boards and capacitor series and brands that were replaced:

Board / Position	Original	Replacement
F-3677 EQ	100 µF / 25 V Nichicon	Vishay MAL214651101E3, 100 µF / 50 V
F-3677 C25/C26	47 µF / 10 V Nichicon	Rubycon YXF, 47 µF / 50 V
F-3677 C21	10 µF / 50 V	Rubycon YXF, 10 µF / 100 V
F-3677 C5/C8 factory-mod positions	10 µF / 25 V Nichicon SL	Rubycon YXF, 10 µF / 100 V
F-3677	220 µF / 35 V Elna	Elna RJH, 220 µF / 35 V
F-3677	100 µF / 35 V NCC	Vishay MAL214651101E3, 100 µF / 50 V
F-3678 power amp	470 µF / 100 V Elna “For Audio”	Nichicon UFW, 470 µF / 100 V
F-3678 C23	470 µF / 50 V	Panasonic EEU-FM, 470 µF / 50 V
F-3678 parallel bypass	10 nF styrene	Vishay MKP1837, 10 nF polypropylene
F-3678 C3 / similar	1 nF styrene	WIMA FKS2, 1 nF / 100 V polypropylene/film replacement actually fitted*
Regulated PSU	1000 µF / 63 V Nichicon	Rubycon PX, 1000 µF / 100 V
Regulated PSU	470 µF / 50 V Elna AWD	Nichicon PW, 470 µF / 100 V
Regulated PSU	220 µF / 16 V	Elna RJH, 220 µF / 35 V
F-3679 protection	220 µF / 16 V Nichicon	Elna RJH, 220 µF / 35 V
F-3679	1 µF / 50 V polarized electrolytic	1 µF / 100 V WIMA polypropylene film
F-3679	100 µF / 6.3 V Nichicon NP	Elna BP / RSRBD, 100 µF / 50 V bipolar

 

[wmac]

First test was to power on under 200W DBT:




Calibration is done; BIAS and DC Offset:



Bias I had to do twice because the probes that I had did not fit together in the TP

Left and right side:






Tomorrow I will start the tests

 

[wmac]

Starting the tests:
First, I'll see if I have any phase issues:
Load is 8 ohms, 100 W



Change to a square wave, still with an 8-ohm load. Configuration: 1 kHz, 500mV VPP





Then 10 kHz at the same amplitude





At 20 kHz







For the stress and clipping test, I changed the load to 4 ohms, 200W. I could not clip because I had the input at 500mV VPP, but I was satisfied that at full volume I got 68.75 VPP, which is about 148 W. This amp was specified at 130W, so I did not change the input Voltage.




Then the THD for the same setup with a 4-ohm load:

Measured per channel, at full volume:



The other channel:



Now I am going to close the box and listen for a week, and measure the temperature.

 

[penmarker]

bravo!!

Why were the styro caps dead? Do we know the cause of damage? Glue?

 

[wmac]

bravo!!

Why were the styro caps dead? Do we know the cause of damage? Glue?

The big styros were open, and I had an accident with the small one, removing the glue, so I replaced both to match the channels with PP film caps.

 

[wmac]

Parts that were replaced:



75 Electrolytic caps
4 Styro caps
10 resistors
2 Diode Zener
2 Trimpots

93 Components


And spent two days carefully removing deteriorated factory glue