Sansui AU-9900A Restoration Project

SanthoshA

Active Member
I myself was lucky enough to pick up the Sansui AU-9900A from a friend of mine.
This is a 220volt model/multivolt model, made for international markets i presume.

The amp itself runs OK, not in bad shape, there were also couple of issues which i helped my friend fix before i made the purchase.

The history of issues which were present in the 9900A are the following:
1. One non working phono channel ( this was a dry solder in the phono board for the 5 leg transistor).
I managed to get the output working, but then crept up a newer problem, the output on the fixed channel was noisy and grainy. Attributed to the same noisy transistor may be.
2. One could literally boil water on top of the amp.
The biasing was set very high and the amp sounded horrible producing enormous amount of heat.
The amplifier was then re-biased and it sounded so much better.
3. There is though no thermal grease on one of the TO-3 transistors, literally no grease present at all.
4. I do believe on of the TO-3 transistors installed in the amp is a replica and not authentic, just by looks on how it looks like this is what i could presume. I don't think i could even find the original output transistors, probably the On Semi MJ21193G and MJ21194G are the best replacements available.

I have also checked Hyperion(John's) and Dr.Audio's posts and do have some idea on what i am looking at.

The differences between the 9900 and the 9900A, atleast from driver board which i can make out is the missing offset adjustment. In paper the 9900A has a THD of 0.05% vs 0.08% of the 9900 and also a better frequency response compared to the 9900.

I plan to restore the amplifier to good working condition, better than it ever was, and i would like inputs from fellow senior members.

I have also decided to give a break for my AU-X11 restoration, which in fact is running in good health and shape.

And finally yes the serial numbers will be added to the AK database.

Pictures will be uploaded soon.
 
Good luck with this, I have never seen/worked on an AU-9900A, but have reconditioned an AU-9900, happy to help if I can, you've successfully done an AU-X11 so you probably won't have that many questions. :)
 
Last edited:
These are great sounding amps. If you get one channel dropping out intermittently, it could be that the pass throughs on many of the boards will need reflowing. These connections are the biggest potential headache with this unit IMO
 
Good luck with this, I have never seen/worked on an AU-9900A, happy to help if I can, you've successfully done an AU-X11 so you probably won't have many questions. :)

Thanks for your kind words John. I hope there will not be too many problems that i would face during the restore.

These are great sounding amps. If you get one channel dropping out intermittently, it could be that the pass throughs on many of the boards will need reflowing. These connections are the biggest potential headache with this unit IMO

I did check all channels in and out yesterday, i do not think this is could be due to a solder joint, but yes i have seen bad solder joints in the phono boards.

Here are some pictures which i manged to take of the amp , it is in good condition atleast by the external looks.
The needle on the main volume knob seems to be missing.

Some pics of the cometic condition:
Sansui AU-9900A.jpg

sansui au-9900a front closer.jpg

sansui au-9900a rear panel.jpg

au-9900a side1.jpg

au-9900a side2.jpg

au-9900a top.jpg

The tear down will happen mostly today or the day after.
 
The 2sa798's throughout the pre-amp sections need imho to be replaced (all of them) with matched pairs of ksa992's. Have been seeing more and more of them go noisy which in this amp really messes up the tone/filter section. Also when setting the bias allow at least twenty minutes with the covers on for things to warm up before setting to 50ma. There are alternate ways to set the bias using the emitter resistors if necessary. The tuner switch in the lower right corner if not thoroughly cleaned will also cause dropped channels. Will be watching and will add my .02 if required.
-Lee
 
That needle below the volume knob was a pain in the ass to find .... I had my au9900 for at least 5 years before I found one on the bay. EchoWars restored mine. It's a fine amp.
 
The 2sa798's throughout the pre-amp sections need imho to be replaced (all of them) with matched pairs of ksa992's. Have been seeing more and more of them go noisy which in this amp really messes up the tone/filter section. Also when setting the bias allow at least twenty minutes with the covers on for things to warm up before setting to 50ma. There are alternate ways to set the bias using the emitter resistors if necessary. The tuner switch in the lower right corner if not thoroughly cleaned will also cause dropped channels. Will be watching and will add my .02 if required.
-Lee

ZTX795A transistors can also be used in matched pairs in place of the 2sa798's. That's what I used on mine anyway. They have been working well. You have to do some weird bends to get all the legs to go where you want them. Was getting some popcorn-type noise with the old 2sa798's.
 

Attachments

  • F2570.JPG
    F2570.JPG
    116.1 KB · Views: 102
Yup, that serial number looks right. We only have 6 of these fine amps on file and two of them, like yours, were built on line 21 in April '77.

Thanks.

- Pete

That's a cool stat Pete! Thanks for your work re: the serial number database.

I really should update you with a bunch of serial numbers methinks...
 
Yup, that serial number looks right. We only have 6 of these fine amps on file and two of them, like yours, were built on line 21 in April '77.

Thanks.

- Pete

Thanks for the information and for the database updates.

Here are some updates, yes as my suspicion goes, we are looking at fake transistors in the left channel heat sink.
Point 4 in first post regarding the replica transistors.

Left Channel Heat Sink:

Left Channel (NEC 2SB541A)
Left Channel NEC B541A.jpg

Left Channel (NEC 2SD388A):
Left Channel D388A.jpg

Left Channel NEC (2SB541A) - Close Up Image
Left Channel NEC B541A close up.JPG

Right Channel NEC (2SB541A)(Original):
Right Channel NEC B541A.JPG

Just by looking at the transistors in the left channel board i can clearly make out the replicas.
Differences i could find :
1. The replicas have a flat top opposed to a slightly round top in the originals
2. The Red Font is some what different and more thin opposed to the thicker font used in the original transistors.
Please do pitch in if there are other differences between replicas and originals, as it would certainly be useful for the rest who attempt to restore and amplifier or source transistors their amplifiers.

I started with the left channel heat sink first, as i noticed the sansui white plastic tags were cut off to remove the driver board, the right channel sansui tags were intact and in order to remove the right channel heat sink i will have to cut off the sansui tags, these tags in sansui amplifiers are extremely helpful in quickly identifying which component was removed or board was removed for rework.

The amp was literally butchered to death by the previous technician who worked on the amplifier, the next image is painful sight to see.
Fuses shorted using wires .jpg

There are multiple strands of wires clubbed together and soldered onto the metal contacts in the fuse body top and bottom.
Both the fuses are blown.
A much closer image of the fuses.
close up fuse image.jpg
 
This will be quite the egg unscrambling.
200w.gif
 
This will be quite the egg unscrambling.
200w.gif

This is what i think would have happened.

The original output transistors would have been shorted, so the replicas were used in the amplifier.
But probably it kept blowing the fuses in the amplifier, and then they decided to short the fuse using the wire.
Another reason for the fuse to keep blowing would have been a higher bias current set, higher bias would have been set for the left channel with replicas, probably because it was noisy or there was too much cross over distortion. For some reason the driver board and the output transistors drew too much current that the 7 ampere fuse got blown.

I am still not very sure on areas touched on the amplifier when this repair work was done, probably it was restricted to only the driver board or on others areas too.
 
I think those transistors look OK - from a different manufacturing run, maybe separated in time by a year or two, but OK I think. :) 'Real' transistors can blow fuses just like fake ones can, ask me how I know. ;) Oh, and 'fake' repairs can blow fuses too!

Be careful of the (yellow) bead thermistors, they have very brittle leads and can easily be permanently damaged. ;)
 
I think those transistors look OK - from a different manufacturing run, maybe separated in time by a year or two, but OK I think. :) 'Real' transistors can blow fuses just like fake ones can, ask me how I know. ;) Oh, and 'fake' repairs can blow fuses too!

Be careful of the (yellow) bead thermistors, they have very brittle leads and can easily be permanently damaged. ;)

Thanks for your reply John, i also noticed the same batch numbers (I)R5Z floating in eBay, but either ways i will do a peak atlas curve trace and see between both channel transistors.
As i have limited equipment with me, and in some cases the rear of the transistors have an engraving underneath, and are not completely flat, i am not sure how it is in the other transistors in the right channel.
 
The erratic distortion on left channel has been verified is actually coming from the output transistors, i swapped them on semi MJ21193 and MJ21194, and it sounds far better.

I have sourced four output transistors from fellow member Kale, really thankful that he has spares for these vintage gems, and have served well for certain parts i needed for the restoration builds.

The transistors are shipped and are on the way to my place, i will then do a full comparison between the transistors doubted as fakes and update the thread accordingly.

Meanwhile capacitors for the driver board have been ordered, i have also removed the driver board for further inspection.
The driver looks in good shape, with no mod or repair work that has been done on it from what i can see .
Whereas the pre-drivers are all fully depleted of the thermal grease, and i would re-grease them.

Sansui AU-9900A driver board_1.jpg

All voltage measurements throughout the amplifier measured lesser than 7 percent, diagnosed the issue to the transformer being wired for 240 volts instead of 220 volts in my country.
The voltage swap was pretty much easy and straight forward to do so, i had to remove the white piece of plastic tab, and then put it in a direction where the arrow points to 220 votls instead of 240volts.

After the corrections were made the voltage measurements were falling in place.

There is though no DC Offset adjustment in the AU-9900A driver board compared to the AU-9900.
 

Attachments

  • Sansui AU-9900A driver board_1.jpg
    Sansui AU-9900A driver board_1.jpg
    123.7 KB · Views: 38
Based on suggestions i do believe we can use either ZTX795A or KSA992's to replace the 2SA798 transistors.
The KSA992 is available in two gains ,

KSA992FTA with a gain of 300 and
KSA992FBU with a gain of 430 ( probably risk's oscillation) ?

Please let me know your views on the same.

I also forgot to add that when i engage to tuner switch to "ON", no inputs connected(Tuner RCA's left un-connected) i get a feeble output in the left channel (AUX signal which is currently playing) whereas the Right Channel is silent.
 
Back
Top Bottom