JoseHH

Well-Known Member
Hello everyone.

Last August I got this AU-D707X Decade amplifier here in Bangkok. The amplifier turned on, came out of protection, and it sounded good when we tested it at the shop. We used a pair of Yamahas NS1000 and a Marantz CD-63 CD player. I observed that all the functions work with the exception of the subsonic filter which did not engage when I pressed it. The amplifier has a little damage at the lower left corner of the particle board side panel. But the front and back panels are in good condition, and a quick look inside suggested that all the boards except for the regulated power supply (F-4292) did not get any noticeable work done on them. So I took it home and I had no time to do anything until nowFrontPanel_AK_Small.jpg.

There are no threads discussing renovation and adjustment of the AU-D707X amplifiers, which are the Japan market equivalent of the AU-D11-II. Thus, I studied every thread I found in this site about the most similar units (AU-D11-II and AU-G90X). Big thanks to @Hyperion, @phonomac, @LWB and @leesonic for all their extremely valuable threads. A very useful site was the amp8.com, a company that repairs amplifiers in Japan which shows pictures of the units they repair. I wish to renovate this amplifier, so I start this thread to get advice from users with experience in this kind of amplifiers.

The first thing I did was to carefully inspect the unit internal parts, and had a close look at the F-4292 power supply. The board had damaged electrolytic capacitors, and a broken tab on the upper left corner. There were bulging electrolytic capacitors, a ceramic resistor, five delaminated/damaged solder pads on the solder side. So I decided not to power on the amplifier until I fix this board. Clearly somebody worked on this amlifier before. Unfortunately that person was not carefull or did not care about the unit very much... Here is a picture of the initial state of the power supply board:

F4292_DiodesInitial_AK.jpg

For this board, the circuit diagram is the same as that fr the AU-G90X.
I removed all the electrolytic capacitors, and cleaned the glue.
I noted that on the -30V V rail, there was a 1N4007 diode installed in place of a constant current diode (10DY4.5). Thus, they found the need to install a 3W, 150Ohm resistor where there should have been a 33 ohm resistor rated at 1/2 watt. The 4.5 mA current regulating diode can be replaced by an E-472. I replaced both current regulating diodes. The installed zener diode in the -30V rail was a 1W version which has a Izener of 8.5mA. With the 4.5 mA current regulating diode in place, the zener diode is not working on the steep part of the I-V curve, so the negative rail would have less than -30 volts. So i replaced both zener diodes, using 0.5W diodes whose Izener is 4.2 mA so they would regulate the voltage properly. I installed non inflamable 33ohm and 10 ohm resistors where appropriate, and then put new electrolytic capacitors ased on what I had available. I used Panasonic FC 470 uF 63V for the unregulated power supply. For the regulated power supply, instead of the pair of 470uF 50V, I felt that there was no harm on using a pair of Nichicon UFG 1000uF 63 V which fit nicely in this location. The zeners were bypassed with Panasonic FC 100uF, 63V, and the output of the regulated power supply were upgraded from 100uF 35V to 220uF 63 V (Panasonic FC).
Then I fixed the delaminated pads, using 60minute epoxy glue. For the loose connector I installed some inserts using 60 minute epoxy to glue them in place.Once the epoxy cured, I filed the tubes flush with the top surface of the PCB and then soldered the cnnector back in place.PadsRepair_AK_small.jpg

I fixed the broken tab by gluing a piece of pertinax using epoxy. Work still in progress. I will post more details after test this board.

Jose
 
Everything worked as expected; the voltages out of the regulated power supply measured +29.9 and -29.8V. I used regular 5% zener diodes (Fairchild 1N5256B). Here is the final look:
Repaired_F4292.jpg

To test this board I unplugged the connector that brings the transformer wires to the main power supply (F4291) , them plugged the corresponding three connectors from the transformer to the F4292. To avoid damaging any of the circuits, at this point I did not connect any of the outputs of the F4292 board.

After this initial test, I reconnected both power supply boards, and I turned on the amplifier using a dim bulb tester: no urprises up to that point. Next I powered with full mains using a 220V to 100V transformer and the unit powers and comes out of protection in 7 to 8 seconds. Since the amplifier is partly dismantled, I went on to measure the DC offsets and bias. I did not change any My results were the following:
At the control board terminals 38, 39 -0.2mV
At the control board terminals 41, 42 +1.1mV (adjusted to 0.1mV)

Hot-to-Cold Right channel 66mV
Hot-to-Cold Left channel 60mV

Hot-to-GND Right channel 502mV
Hot-to-GND Right channel 455mV
These four were unstable, and fluctuated between zero and the values above. Here I discovered that the long metal strips that connect the speaker terminals have an intermitent connection. The pads at the board F4553, were damaged. There is a suspiciously bad looking solder blob and I can see that the track that goes to that solder point where the speaker binding post is soldered is loose. I think I finished damaging this when I was trying to remove this board to take note of the capacitors.:( I take the blame because I would have noticed an intermitent conection when I tested the amplifier with the seller. Anyhow, this will have to be the next board I repair before changing the trimpots and the electrolytic capacitors.
For the bias currents:
Bias currents measured at transistor emitters:
At Q62-Q64 V=27.5mV (Rigth channel)
At Q61-Q63 V=29mV

At Q61-Q63 V=28mV (Left channel)
At Q62-Q64 V=29mV

I think this amplifier is biased abit too high. I did not attempt to adjust the DC offsets and bias on the amplifier stage yet because this could lead to a kaos with the old trompots. Acording to the AU-G90X service manual the voltages across the emitter resostors (a total of 0.54 Ohms) should be 27 mV while for the AU-D11 II (uses the same emitter resistor values) is specified to 20mV. I do not know yet what would be the final setting on this unit. Eventually, with new muliturn trimpots I will try to see if the oscilloscope shows crossover distortion when the bias is adjusted to give 20 mV across the resistors.

What is the correct way to remove the speaker protection board?. I was thinkinking on removing the screws that hold the RCA connectors and then remove the back panel to access to it. Any recommendation is appreciated.

I will work some more and report how it goes.
 
What is the correct way to remove the speaker protection board?. I was thinkinking on removing the screws that hold the RCA connectors and then remove the back panel to access to it. Any recommendation is appreciated.
If I recall correctly this is what you have to for the AU-D11 II (i.e. remove the rear panel) - the speaker terminals are then left in 'mid air' supported by the Protection PCB only.
 
Here is the speaker protection board (F4553). Interestingly, it has only two relays for commutation of the speaker systems. The inrush relay is missing but the board appears to have the tracks and pads implemented but they have not been populated with the components. The schematic appears to follow that for the corresponding board in the AU-G90X except for this missing components. Notably, looking at the pictures posted in the amp8 website, it appears that none of the AU-D707X have this inrush current control installed but the boards in principle support this circuit. I wonder if this is a cost-saving decision or rather a regulation that impedes the use of this kind of protection for some safety reason. For the time being I will not attempt to populate this part of the circuit until I understand why it was not there from factory.
The main difficulty with this board is that the pads where the speaker binding posts are soldered are partially lost. Out the 8 pads where the metal strips are soldered, four are good, one lost only the portion not covered by the solder ask and the other 3 lost half of the perimeter part of this was caused by previous work, but I may have done part of the damage. I had to carry the amplifier barely packed from the shop where i got it to a place 250m away where I could get transportation. - and yes, my back still hurts from this exercise -
Also, I was trying to make my way to take pictures of the capacitors in this board and I may have moved the board. The second probems is the not so good speaker binding posts, and finally the electrolytic capacitors that I wanted to renovate. In any event, here are pictures of how the board looks after I removed it:
F4553_before_renovation.jpg
I removed the speaker binding posts, which were in good shape, despite of not being a very strong design. I was studying the excellent thread by @Hyperion who is one of my references, especially for this kind of amplifier. Unfortunately I do not have the means of doing careful work on acrylic plates here in bangkok nor I could get decent binding posts like the ones He used. But I plan to use the same modification posted by Hyperion in a future revision of this work after I get the amplifier working reliably. For the time being I opted to refurbish the existing binding posts to be able to use banana plugs or bare wires. I do not care about reversing this to the original configuration. I decided to keep the black plastic frame and the long metal ribs that are soldered into the board. Then I used the nuts and the shafts of readily available binding posts:
Refurbished_Speaker_binding_posts.jpg

I am not exactly proud of this refurbishment, but I believe it is decent and I am confident that it will work until I carry a better upgrade.

Although the pads in the printed circuit are partially damaged, I could solder this assembly onto it provided that I reinforce the connection using wires to a zone of the track where there is good integrity of the bonding of the copper foil to the phenolic board.
But I am still trying to decide what would be the best route considering a future replacement of the binding posts. Suggestions are welcome.
Thinking aloud, there is no room for implementing anything using screws and nuts. Thus the soldering is the mechanical support as well as the electrical connection. One possibility would be to drill the board and use tin plated copper inserts of the type I used on the power supply board shown above. This way the epoxy bonding of the copper foil does not take all the stress.

I will continue with replacing the electrolytic capacitors and will post that soon.
 
I repaired the pads around the rectangular holes on the PCB where the metal strips are soldered. I used some tin plated bushings that generally used to improvise a plated-through hole. The external diameter of the tube on the non flared end is 2.65mm and the internal is 2.15mm. The overall length is 3.0mm. They can be easily reshaped into rectangular bushings by pressing with pliers while using a rectangular template inside. The rectangular holes on the pertinax PCB need to be slightly enlarged, which i did just scraping the internal walls with an "exacto knife", perhaps about 0.2mm on each side. Ideally after applying epoxy glue one would use a drill press to flare the top end using a drill blank of about 5mm in diameter on the top and bottom sides. Then, the shaping tool should be used to make sure the rectangular clearance is sufficient for the metal contacts to pass freely. Then after the epoxy sets, soldering can be done. I believe that this repair will provide extra support for the printed circuit which is held by the metal trips that connect it to the binding posts.
Here is a picture of the results:
F4553_Pads_repaired.jpg

And here is a sequence of pictures describing what was done:

RectangularPadsRepair.jpg
 
I received the capacitors yesterday, so I could install the new components on the speaker protection board. I replaced the 4x 100uF 16V bipolar type with Nichicon ES-muse bipolars trated at 25 volts (locations C31-C32). The 2x 0.1uF 50V electrolytics were replaced with WIMA 0.1uf 63V (MKS2). The rest of the capacitors were standard ELNA and were replaced with Nihicon FW (1X 220uF 50V C1, and 2X 1uF 50V for C33, and C36), regular ELNA (1x 33uF 50V) Nichicon VR (2 x4.7 uF 50V in the flip flop) and Nichicon SH for the 0.47uF 50V that was in C34.
This was a good time to replace the relays. I used Omron G4W-2212P-US-TV5-HP-DC24 for both speaker commutation relays. Thanks to @LWB for this. There was no inrush relay in this board.
I do not know what kind of capacitors were used in KC68-69 for the Zobel network. But because they were bulky and difficulted the installation of the speaker terminals. For this reason I replaced capacitos with those recommended by @Hyperion in his thread that are polyester rated at 250V. (Thanks John:)). Here are all the new components installed:
F4553_renovated_top.jpg

After double checking that the capacitors were correctly oriented, it was time to solder the wires. (I think it is more convenient to install the speaker posts at the very end. So, I soldered the single wires first, leaving the more delicated ribbon cables for the end. With the back panel removed, I started by holding the board rougly in the final orientation and populated the wires from botom to top, beginning with wire No 45. Then I installed the ribbon cables. Then the speaker terminals. I think I will put the amplifier together, check the DC offsets and bias and if everything is ok I will listen to it before renovating the capacitors and trimmers.
 
Before renovating the capacitors in this amplifier, I wanted to see if it still works since I disasembled it. So I put it together. I checked the DC offsets after leaving the unit warming up for about 40 minutes. Hot-Cold voltages were 45mV(R and 30 mV and stable. The hot to GND values were near 1 volt.
I took the chance to adjust the amplifier using the original trimmers. I used the service manual instructions for the AU-G90X. First i adjusted the two trimmers on the control board behind the front panel, to 0.0mV. They were close but I tried to get them closer to 0.0.
Next I adjusted the Hot-to-Cold offsets (100 Ohm trimmer on the driver boards). I could bring both channels down to specification, but they fluctuated in the +/- 3mV range.
For the Hot-to-GND i lowered to 0 but the values fluctuated between -100mV and 100 mV approximately. I left it there.
The bias currents gave all near 28 mV voltage reading between the corresponding emitters. These seemed a bit too high and here I decided to lower them to 20mV. I quickly rechecked the DC offsets once more, and then put all the covers and I connected it to my speakers and CD player.


The first thing i noticed is that there is a "hum" on both channels that becomes evident when the volume knob is at -42db and it gets louder wit the volume. There is no change upon disabling the tone control. I do not recall hearing this at the shop but the area was quite noisy.
Other than that the amplifier in its current state sounds quite good but not as good as my renovated AU-D907. I have to listen to it a little more.
At this point the only issue that worries me is the hum. I am thinking that I should check the ground wires, and the rippe in the power supply. I have all the components to replace the electrolytic and the trimmers. I wonder if I should first replace them and then track down the cause of the hum?. I will appreciate any suggestions about the best way to proceed.
 
I am listening to this amplifier a little more, and i realized that the the bass is more pronounced than that for the AU-907D. Even without putting new capacitors this thing sound quite good. It does not get hot at all.
In addition to the hum, the volume and balance controls and the switches need cleaning. The subsonic filter dos not enagage (the pilot lamp lits but it does not stay on (the switch does not latch).

With regards to the hu noise, its level increases with the volume, so this suggests me that it originates in the circuits prior to the power amplifier stage.
 
It does not get hot at all.
These amplifiers run very cool - which is quite normal.
With regards to the hu noise, its level increases with the volume, so this suggests me that it originates in the circuits prior to the power amplifier stage.
I found the positioning of the cables near the volume control to be critical for minimisation of hum. Also have you seen my posts where I mention providing an extra ground wire from the rear panel ground point to the star ground on the underchassis near the transformer?
 
Hello John,
Thanks for your comment. I have read your post about the extra ground wire but I can not locate the star ground point on the underchasis near the transformer. May be this amplifier os different in that respect. I am attaching a picture of the area near the transformer from the bottom side:
Bottom_View_PowerSupply.jpg

There is a star ground point on the back panel and there are two wires that are soldered from the bottom of the main power supply, into the points where a pair of 6.8K resistors form a "center tap" coming from the positive and negative rails. Thease wires (one white and one black) that are wired to the star ground on the back panel.
I tried to take a picture of these:
Ground_Wires.jpg
 
With regards to the wires near the volume control, that is a very important clue. I will see if that is what causes this hum.
Thanks again!!
 
Ah yes, I see that the under-chassis design is a bit different to the AU-D11 II - in your picture the star ground would be to the right of, and slightly above the white push on connector - and located on the front to back chassis member. But no sign of it or any of the associated wiring shown in your picture as compared to the AU-D11 II.

Your description of the 6.8K resistors, and the way they are connected suggests that they dealt with the issue I found, and solved, - but in a different way. ;)

Again regarding hum - all bets are off if the internal shielding covers aren't fitted. ;) (this will very likely be model dependant - you may not have so many internal covers as on the D11 II)
 
Thanks John.
I double check and indeed there is no star ground near the transformer. I think you are right, Sansui made changes in the ground scheme. My power board is version 3. But I can not see a version number on the power supply board.
There are only two internal covers in this model. One of them runs parallel to the right side panel, shielding the equalizer board, and the selector controls. Then there is another internal cover behind the front panel shielding the back of the control board.
I will see if i can resolve the issue of the hum by altering slightly the wires near the volume control and reinstalling the covers.
 
Wow!. John you are the man!. Although not completely resolved, what you recommended made a huge difference. Thank you!. Now the hum appears when the volume knob reaches -11db. Before it i can hear it at -42 db. What I did is the following. I removed the top cover and the shieldings. Then i took the bundle of wires that pass above the volume potenciometer, and used two cable ties to hold them more tightly grouped. There was a twisted pair of wires that connect via a molex connector to the regulated power supply at points numbered as 82 and 84. They power the control board) which I had it going on the component side of the regulated power supply but these wires had a lot of slack, so should it be routed on the foil side of the regulated power supply? I experimented with this wire but tt made no perceptible change on the hum. My initial picture shows it going on the component side, but it was not neatly bound to the rest of the wires.

Now I have the amplifier on, and I can hear a "hiss" which is noticeable even with the volume control at its minimum. This hiss becomes a lot weaker when I engage the tone control but it may still be heard. Its presence is constant up to aproximately -9db, then increases with the volume control. When the tone control is engaged it is constant up to about -42db and then it increases with volume but at -10db the hum quickly becomes the dominant noise. Further fiddling with the wires does not alter this.

I was planning to replace all the elecrolytic capacitors. I am thinking if I should start with this at this time..
 
I was planning to replace all the elecrolytic capacitors. I am thinking if I should start with this at this time..
I reckon you are ready too. ;)

Glad the tip about the wiring helped with the hum - that was hard won knowledge, :) discovered while I was chasing a much lower level and persistent residual hum in one channel only.
 
Back
Top Bottom