• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Sansui AU-20000 Question

Thorbr

AK Subscriber
Subscriber
Looking for some advice/comments from anyone who owns/owned/serviced a Sansui AU-20000. Picked one up late last year and it is in my queue for detailed testing and restoration in the next few weeks. Unit appears to be original and was well taken care of by the previous owner. So, I must admit that prior to the purchase, I fell a bit for the hype on these, but I also had an AU-9900 that I restored with assistance from several AK members and am very happy with in terms of design and performance. After more considered research, I am reaching the conclusion that the AU-20000 is a much different beast and may be beyond my technical skill set. What I have is part of the first release, 1976 mfg date and does not include the factory mods in the newer revisions. For those that own/owned one, is this truly the "holy grail" worth the time and effort to restore and possibly pass on to those younger in the family, or am I looking at a bottomless money pit? Thanks in advance for any response.
 
Register to hide this ad
Looking for some advice/comments from anyone who owns/owned/serviced a Sansui AU-20000. Picked one up late last year and it is in my queue for detailed testing and restoration in the next few weeks. Unit appears to be original and was well taken care of by the previous owner. So, I must admit that prior to the purchase, I fell a bit for the hype on these, but I also had an AU-9900 that I restored with assistance from several AK members and am very happy with in terms of design and performance. After more considered research, I am reaching the conclusion that the AU-20000 is a much different beast and may be beyond my technical skill set. What I have is part of the first release, 1976 mfg date and does not include the factory mods in the newer revisions. For those that own/owned one, is this truly the "holy grail" worth the time and effort to restore and possibly pass on to those younger in the family, or am I looking at a bottomless money pit? Thanks in advance for any response.
I have a been kicking the tires on one as well over the last 60 days or so. It is local to me, and not likely to go anywhere unless I buy it. The one I am looking at is also an early serial number. I have done my share of rehabs/repairs of Sansui's (and a few Pioneers and Kenwoods, Luxmans as well), and I feel that I am equipped to manage this 20000 no matter the vintage. My feeling is that if you were able to work through a 9900, you can tackle a 20000. I suspect the only "new" challenge for you will be the pass-throughs (I cannot remember if the 9900 has passthroughs).
 
I have a been kicking the tires on one as well over the last 60 days or so. It is local to me, and not likely to go anywhere unless I buy it. The one I am looking at is also an early serial number. I have done my share of rehabs/repairs of Sansui's (and a few Pioneers and Kenwoods, Luxmans as well), and I feel that I am equipped to manage this 20000 no matter the vintage. My feeling is that if you were able to work through a 9900, you can tackle a 20000. I suspect the only "new" challenge for you will be the pass-throughs (I cannot remember if the 9900 has passthroughs).
Thank you for the positive response. Yes the "passthroughs" caught my eye and not something I have run across on other Sansuis.
 
Thank you for the positive response. Yes the "passthroughs" caught my eye and not something I have run across on other Sansuis.
Yes, I have not done passthroughs either. I have a derelict CA-3000 that has something like 40-50 passthroughs in it, which I have not yet started to tackle. But I think just steady as she goes - there's no reason why they should be a special challenge.
 
I own a AU20k that Ive been plucking away on the last year or so, got some great help from an AK member on testing some basics, theres a few other members here that are very knowledgeable on these, plus quite a few past threads on restoring them here & other audio forums out there.

I'm an amateur but have above average soldering skills, have worked on a few vintage amps & other gear, just recently did a full recap of a AU719 and replaced some other parts like bad output drivers & black flag caps, that was my first Sansui and was happy to find how easy the boards are to remove & work on... the AU20k is also a modular design and most the boards come out pretty easy so its not very difficult to work on, just be patient and realize how everything comes apart.

Get the service manual & study it to get familiar with the boards, then do some research on how others work on each board and what areas need attention like some vd1212 diodes, or the boards with the bad glue that needs to be removed, there are 3 of them. I havent got into the pass throughs yet but after looking at a few other resto threads it doesnt look too difficult, removing the heat sinks & separating them looks more involved than repairing or reflowing the pass throughs.

The Au20k has a lot of boards & parts to replace so how much you do depends on its current condition, some say every single part needs replaced, but as with many things, I follow the "if it aint broke dont fix it" approach for the most part... i left all original parts in the 719 besides the lytic caps and other items known to have problems or failures like black flag caps, but the rest of the transistors, resistors, etc tested good &/or work & sound good so I left them alone, nothing wrong with a complete replacement of all parts though if thats what you want to do.

Best of luck!
 
Last edited:
The thing you need to pay attention to is unfused live power rails in the AU-20000. Before touching these, the power filters need to be discharged and disconnected. These are not an easy restoration if done properly.

However one really doesn’t have a Sansui collection without oneIMG_1880.jpeg
 

Attachments

  • IMG_6871.jpeg
    IMG_6871.jpeg
    262.7 KB · Views: 9
It has been a while, but I finally started on the restoration of this beast. First off, I've had it running for several weeks, with no issues to report, but at low volume, say 30% of volume full scale, the top of the case did get noticeably warm. Found that one of the two 220 ohm, 15 w, cement resistors, at the back of the power regulator and driver boards, plug in chassis board, is running hot, say 80 deg C compared to 45 deg C for the other resistor. Per the attached photo, the top one runs hot. Also the overall current draw, at idle state, is close to 1 A, which is much more than I have measured on other units. Bias adjustments were made with only minor changes. The other observation, is that some modifications were made to this unit. Even though it is an early build. June 1976, the driver boards temperature sense circuit is on the driver board, per the newer design revisions. The addition of the two 220 ohm resistors, is also part of the redesign. Plan to proceed with a refresh on the power regulator board followed by driver boards. Hopefully that will identify rogue component(s) that may be causing the temperature anomaly.
 

Attachments

  • 20260910_181602.jpg
    20260910_181602.jpg
    113.3 KB · Views: 11
It has been a while, but I finally started on the restoration of this beast. First off, I've had it running for several weeks, with no issues to report, but at low volume, say 30% of volume full scale, the top of the case did get noticeably warm. Found that one of the two 220 ohm, 15 w, cement resistors, at the back of the power regulator and driver boards, plug in chassis board, is running hot, say 80 deg C compared to 45 deg C for the other resistor. Per the attached photo, the top one runs hot. Also the overall current draw, at idle state, is close to 1 A, which is much more than I have measured on other units. Bias adjustments were made with only minor changes. The other observation, is that some modifications were made to this unit. Even though it is an early build. June 1976, the driver boards temperature sense circuit is on the driver board, per the newer design revisions. The addition of the two 220 ohm resistors, is also part of the redesign. Plan to proceed with a refresh on the power regulator board followed by driver boards. Hopefully that will identify rogue component(s) that may be causing the temperature anomaly.
The Sansui serial number database shows AU-20000 production from October 1975 to December 1977. 1975 is considered "early production" with "early design" features through at least November production. I would expect a June 1976 unit to have all the later revisions in place, but @richardpp has a lot of knowledge on the 20000, and he may weigh in with some insights.
 
It has been a while, but I finally started on the restoration of this beast. First off, I've had it running for several weeks, with no issues to report, but at low volume, say 30% of volume full scale, the top of the case did get noticeably warm. Found that one of the two 220 ohm, 15 w, cement resistors, at the back of the power regulator and driver boards, plug in chassis board, is running hot, say 80 deg C compared to 45 deg C for the other resistor. Per the attached photo, the top one runs hot. Also the overall current draw, at idle state, is close to 1 A, which is much more than I have measured on other units. Bias adjustments were made with only minor changes. The other observation, is that some modifications were made to this unit. Even though it is an early build. June 1976, the driver boards temperature sense circuit is on the driver board, per the newer design revisions. The addition of the two 220 ohm resistors, is also part of the redesign. Plan to proceed with a refresh on the power regulator board followed by driver boards. Hopefully that will identify rogue component(s) that may be causing the temperature anomaly.

You have really nice unit there but in need of attention.

The mains current draw is much higher than it should be. If you don't have a dim bulb tester in circuit I would use one before going any further. It will limit the amount of current the amp is drawing. And I would start by finding out why the 15W resistor is over-heating before doing any other work on the amp.

The 15W resistors are a part of the regulated power supply that feeds the control amp section of the amplifier so the problem is going to be either with the supply board itself or somewhere down stream in the control amp section. Meanwhile, it's important you keep that resistor clear of the adjacent wiring bundle.
 
You have really nice unit there but in need of attention.

The mains current draw is much higher than it should be. If you don't have a dim bulb tester in circuit I would use one before going any further. It will limit the amount of current the amp is drawing. And I would start by finding out why the 15W resistor is over-heating before doing any other work on the amp.

The 15W resistors are a part of the regulated power supply that feeds the control amp section of the amplifier so the problem is going to be either with the supply board itself or somewhere down stream in the control amp section. Meanwhile, it's important you keep that resistor clear of the adjacent wiring bundle.
Thank you both for the quick responses and direction given. In reference to serial numbers, what I had in my notes, is that the older model numbers were from 1975 to 1977. Not sure where I pulled this, but probably a lazy google search. What I also noticed, is that related to the updates, they appear to be completed at a later date, based on the solder work. Very clean, but does not appear to be factory original. There were also several transistors replaced on all 3 boards. Caps and diodes appear to be original. Lot's more investigative work to complete on my end.
 
Just one more follow-up and question for today if I may. I checked the outputs more closely and they appear to be the original Sanken 2SC1585 & 2SA908 complimentary pairs. I plan on pulling them out and testing, but based on age, they will probably require a refresh. For pulling out the heat sinks, I see that the bottom connections are plug in type, but not the top connections, to the driver boards. Am I missing something? I've done this previously for an AU-9900 and it was all plug in connectors.
 
Thank you both for the quick responses and direction given. In reference to serial numbers, what I had in my notes, is that the older model numbers were from 1975 to 1977. Not sure where I pulled this, but probably a lazy google search. What I also noticed, is that related to the updates, they appear to be completed at a later date, based on the solder work. Very clean, but does not appear to be factory original. There were also several transistors replaced on all 3 boards. Caps and diodes appear to be original. Lot's more investigative work to complete on my end.
I can't add anything to the restoration questions, but I can confirm you should do it! :-).

I bought my AU-20000 (and a TU-717) directly from Sansui in Yokohama, Japan, on April 1, 1977. I paid the equivalent of US $453.00. I had met one of their Engineers while working for JVC. Interestingly, I received a "Factory Data" sheet from Sansui with the following measurements:

Total Harmonic Distortion
at 20hz - .05%. Measured: Left, .009%. Right, .012%
at 1,000hz - .05%. Measured: Left, .008%. Right, .010%
at 20,000hz - .05%. Measured: Left, .029%. Right, .034%

Signal To Noise Raio
from Phono-1 (3mv position): Rating: 70db, Measured: Left, 74.0db, Right, 74.0db
from AUX: Rating: 80db, Measured: Left, 94.5db, Right, 94.0db

Phono Equalization Accuracy
Deviation from RIAA Curve from 30 to 15,000hz (3mv position): Rating: +/-0.2db, Measured: Left, okay, Right, okay

Frequency Response
Overall (AUX to power output) from 10 to 50,000hz at 1 watt: +.00 -1db, Measured: Left, okay, Right, okay

Harmonic Distortion of Measuring Instruments:
0.0016% at 20hz, 0.0016% at 1,000hz, 0.0018% at 20,000hz

I could find no testing of total power output; however Many Moons tested it at 210WPC before clipping occurred (rating is 170wpc).

It's been decades since I went through my file. I have all the paperwork plus the all the brochures Sansui offered at the time for The Definition Series. The AU-20000 is not listed in the main brochure of The Definition Series, but the AU-20000 specific brochure lists it as belonging to The Definition Series. I attribute this to how quickly things were changing in those days. I remember being told that the AU-20000 was a combination of the BA-3000 and the CA-3000 but I have subsequently found out the pre-amplifier circuit is not the same as the CA-3000. I'm not sure how that affects the sound - all the features on the front panel are identical. Maybe Sansui figured out a simpler method to achieve the same results - I don't know.

In 2010 I had Many Moons Audio (in OK) completely refurbish it. Paul Hovenga owns MMA and has carved a niche in the Sansui world. Over the years, I have had many, many other amps, some of which came close to matching the AU-20000 in clarity, power and "sound".. But as you know, all these components are a subjective interpretation. But paired with my JBL's (4320, 4430 and C36-030), I've not found a better sound. I believe if you get yours working up to the factory standards, you will be very pleased at how it sounds.

Cheers,

Rich
central Texas
 
I can't add anything to the restoration questions, but I can confirm you should do it! :-).

I bought my AU-20000 (and a TU-717) directly from Sansui in Yokohama, Japan, on April 1, 1977. I paid the equivalent of US $453.00. I had met one of their Engineers while working for JVC. Interestingly, I received a "Factory Data" sheet from Sansui with the following measurements:

Total Harmonic Distortion
at 20hz - .05%. Measured: Left, .009%. Right, .012%
at 1,000hz - .05%. Measured: Left, .008%. Right, .010%
at 20,000hz - .05%. Measured: Left, .029%. Right, .034%

Signal To Noise Raio
from Phono-1 (3mv position): Rating: 70db, Measured: Left, 74.0db, Right, 74.0db
from AUX: Rating: 80db, Measured: Left, 94.5db, Right, 94.0db

Phono Equalization Accuracy
Deviation from RIAA Curve from 30 to 15,000hz (3mv position): Rating: +/-0.2db, Measured: Left, okay, Right, okay

Frequency Response
Overall (AUX to power output) from 10 to 50,000hz at 1 watt: +.00 -1db, Measured: Left, okay, Right, okay

Harmonic Distortion of Measuring Instruments:
0.0016% at 20hz, 0.0016% at 1,000hz, 0.0018% at 20,000hz

I could find no testing of total power output; however Many Moons tested it at 210WPC before clipping occurred (rating is 170wpc).

It's been decades since I went through my file. I have all the paperwork plus the all the brochures Sansui offered at the time for The Definition Series. The AU-20000 is not listed in the main brochure of The Definition Series, but the AU-20000 specific brochure lists it as belonging to The Definition Series. I attribute this to how quickly things were changing in those days. I remember being told that the AU-20000 was a combination of the BA-3000 and the CA-3000 but I have subsequently found out the pre-amplifier circuit is not the same as the CA-3000. I'm not sure how that affects the sound - all the features on the front panel are identical. Maybe Sansui figured out a simpler method to achieve the same results - I don't know.

In 2010 I had Many Moons Audio (in OK) completely refurbish it. Paul Hovenga owns MMA and has carved a niche in the Sansui world. Over the years, I have had many, many other amps, some of which came close to matching the AU-20000 in clarity, power and "sound".. But as you know, all these components are a subjective interpretation. But paired with my JBL's (4320, 4430 and C36-030), I've not found a better sound. I believe if you get yours working up to the factory standards, you will be very pleased at how it sounds.

Cheers,

Rich
central Texas
Many thanks for the additional information. Yes this definitely a keeper and well worth the effort to restore. In regards to my connector question, I was a bit gun shy to pull on the top connectors for the driver boards. All good now.
 
So I have tackled the outputs and heatsinks. Doing this a$$ backwards compared to my original plan, but it is probably the better approach. Pulling 3 of the 2SC1585 transistors I get get hfe readings of between 27 and 28. If I see the same on the others, then I am tempted to reuse these after cleaning and new mica/paste. Suspect these are original, but were worked by the previous owner, as the paste looks relatively fresh, a bit messy though - lol. Any thoughts and if replacing is the consensus, are there any recommendations as to part numbers? There is a very detailed BOM, by an AK member, from several years back, with part numbers MJ21194G MJ21193G, for the complimentary output pairs. Digikey has these in stock, but is this the best option? Cost is obviously always a concern, but a few dollars difference, is not a deal breaker. The other question is in regards to the pass throughs boards on the heat sinks. I haven't pulled these yet, as some desoldering is required for the output connector boards, before I can get to these. When I do, is there any advice as to what to look for and correct? Thanks in advance for any guidance offered.
 
The original output transistors are best. IF they are the same type numbers for both channels (i.e. not a mixture).

Sometimes it is wise to use a set of new output transistors (the MJ's) until the unit has been fully restored and is stable. Any mistake during restoration could damage the 'Family Silver' i.e. the precious original output transistors. ;)

Just something to consider...
 
All good advice, as always, but I plan on using my dim bulb tester and variable voltage transformer, when I reinstall the outputs/heatsinks. The PNPs on one channel measured at between 41 and 44 hfe, so unless the other channel is erratic I will stay with the original outputs. Attached is a photo of the Right channel F-2497 board. Appears that the pass throughs (rivets) were soldered and capped at some point. I see no reason to redo this. Would like to replace the emitter resistors however, before I button everything up.
 

Attachments

  • 20260915_132030.jpg
    20260915_132030.jpg
    126.5 KB · Views: 5
Don't worry about the hfe values of the original output transistors. They don't deteriorate with age, and are the best thing for your AU-20000. The hfe figures you are posting for them are quite normal.

Replacement emitter resistors are a good upgrade, but to improve on the originals you need to spend quite a lot.
 
Back
Top Bottom