Sansui AU-X11 repair and restoration

Well my travels are over, but I've not yet had time to site down and sort out the FET checks and characterization, but I did want to share one little bit of info. Tonight there was a big thud outside... the type of thud that makes you wince a bit when you are expecting an amplifier delivery. The courier had scurried away before I could say hello. What is it that turned up on my doorstep? Could it be? Another AU-X11. And what do ya know? It was actually well packaged and didn't incur any new damage in transit!!!

Check out this pic if you have any doubts as to the size of these beasts.... I'm not a body builder but really, this amp makes my arms look tiny. Although, I've got to say, that pic really hides the bulk of the machine as you can't see the depth of it which is significant.

It is a dead machine also, but I was hoping it also 'just' had driver board issues and I could do the pair together. The seller reported smoke coming out of the unit during testing.... which other potential buyers attributed to him frying something, possibly the transformer, by not using a step-down transformer, but I had serious doubts about that theory. Anyway, I haven't done anything except cracked it open and checked the power supply and transformer. Aside from physically leaking main caps, the power supply and transformer are doing there job! (PHEW). I was partly expecting to find the remnants of some obvious catastrophe but all I find is a lot of dust, and a very suspicious smell from the rear driver board. I will check further when I have time and decide whether to start a separate thread to avoid confusion.

Again, will try to get onto the original AU-X11 soon... have put in a mouser order for full rebuild of the driver boards, however, Burning Amp 2016 is this weekend in San Francisco so I'm not sure much will get done shortly, but it will happen.
 
Last edited:
I got a chance this morning to check out AU-X11 #2. On the channel that has a smelly driver board all power output transistors are shorted, as is the bias transistor on the driver board and many of the resistors are open in similar positions as in my other AU-X11. One of the driver boards and outputs looks fine. I will wait till I have some more MJs to put in the outputs and then maybe hook up the 2 good driver boards to one machine for a sneak peak of the sound quality. But long story short is that I am now rebuilding 4 drivers boards instead of 2, which is pretty much what I was hoping for when I bought this 2nd unit... do both sets of boards at once.

Oh, one small difference I can see on the driver boards; the SV03 on the 2nd AU-X11 have a different package to the 'brown boats' which was seen on my other unit. I will measure them to double check they are electronically the same as the brown boats.
 
Still haven't gotten around to finish of characterizing all of the JFETs; I'm about half way through. But, I did go to Burning Amp 2016 on the weekend. For those who don't know, it's arguably the biggest DIY audio meeting. The guys who make the new FETs are the main sponsor and they are based in the SF Bay area. Very nice chaps. They are really committed to continuing to make matching P-channel and N-channel devices in single an dual-monolithic packages, not only for the stanford radiotelescope guys (apparently they ordered 2000 devices) and big firms, but for us. I would like to give special mention to Kevin, who handles a lot of the sales and who is the guy you will probably deal with if you email or call them. He is really making himself accessible by phone, and is quick on emails if its just for an order of a couple of devices. In my testing, I did have one or two dud units and chatted to them about that (they seemed very surprised indeed...), possibly I got very unlucky, or fried them myself (I tried to be careful those as I know those gates have fragile insulation). Will prob post a Burning Amp pictorial walkthrough. Oh, Kevin (and Paul) is also the guy talking videos at Burning Amp and making them available on Youtube... again, really enthusiastic guy! They have some more devices planned too. Of particular interest to us, possibly, is that Paul from the company said they are planning to release a device that sits in between the line-up of LSK389 and LSK489. Those are both units I'm testing for this thread. He said the 489 has lower capacitance than the 389, but some people need something in them middle.
 
A couple of shots of the driver boards from the 2nd AU-X11. The power supply caps are doing their job but there is some leakage and will eventually have to go:
rMkzAFj.jpg

2UbOm8A.jpg


Some cob webs:
2hQU5vf.jpg


And can you believe this pic is after I vacuumed and cleaned a bit?! I've seen more dust in a unit before, but this was some kind of industrial soot!
D0A47DS.jpg



A bit of heat related discolouration on those boards!
1wl76H1.jpg


One board looked good - didn't find any bad parts, but can't be sure as the amp is in protection mode. The 'bad' board has a similar level of damage as in my other AU-X11, with all outputs shorted, and many of the resistors and a few transistors fried. Check the colour bands on the blue resistors for example, some look really suspicious, and they measured very closer to open than to there desired value.
j8ktfSU.jpg


Oh, I mentioned some difference packaged SV03 compared to the brown boats I found in the other AU-X11... here they are. 2x of these black diodes per board (other SV03 are the little orange spider eggs with white dot again).
tn06h69.jpg


If you look closely, these ones actually have "SV03" written on them, unlike the brown boats which just had "SV".
CqDMjLE.jpg


As you can see I also pulled the 2sk129A out. I want another sample for characterization.
Here is a pic of the board showing the pinout (which I believe for now to be correct). However, as GPS noted, this could be a bit weird if the package is formed by two single JFETS units stuffed into a single heat sink back to back (or flat to flat, if you like). So I will do some more measurements and make sure the 2sk129A is really symmetric (i.e. symmetric enough to flip around and interchange D and S) and if the pinout on the PCB is correct...
kM50dwH.jpg
 
Oh one important (?) point: I did measure those black SV03 packages that came out of the 2nd AU-X11 and the forward voltage is within 1% of the brownboat SV03s from 0.5mA to 20mA... so no need to adjust my previous charts and substitute suggestion.

And, HO HO HO, merry Mouser! the parts for one set of drivers boards has arrived....
 
Climbing Mount Everest here. What an arduous endeavor you've taken on.

Needless to say that second X11 seems to have had some work done in the past.
I wonder where they dropped the ball?

LOVN THIS THREAD!!

jcxd4ekcE.jpeg
 
I wonder where they dropped the ball?

I actually didn't find any tampering inside. The damage to the driver board is also eerily similar to the damage on my other bad driver board from the other unit. I guess it is a weakness of the AU-X11, rather than bad repair job.

I will check all of the black flags, but on my other AU-X11 which has very very similar damage on the driver board the black flags seemed okay. My guess is that a transistor let go, for any number of reasons, that caused the chain reaction. I am not sure if there was oscillation, it's certainly possible, but more and more I don't like the fragility of that bias transistor's thermal connection to the heat sink and that must be addressed (and GPS agrees emphatically on this point). Again, the Au-X1 has a bit of a different setup, and the thermal connection to bias transistor is slightly more sound IMO. Will post pics of my [hopefully] solution...

Yes, it's a big job, the two AU-X11! And not cheap. The two drivers boards cost about US$240 to restuff if using good parts and Mouser, not to mention the LSK FETS I ordered... they aint cheap either (they are reasonable at $3-10 a piece, but I ordered a large (~60) number of parts from them for testing). I suppose I could have replaced just the damaged parts, but there was so much damage it was going easier just to restuff the entire board. And I hopefully I considered parts values carefully and spent extra where I though it would help end up with either a better sounding, more stable or durable driver board. Fingers crossed. Anyway, it's going to be a one heck of a ride.
 
Last edited:
I've started a separate thread on the JFET replacements, as I thought it might be of interest to others.

See here

Once I finish the posting, the summary will be:

Dual N-JFET 2SK129A --> LSK389B
Single N-JFET 2SK163 --> LSK170A

which are both in current production.
 
I've started a separate thread on the JFET replacements, as I thought it might be of interest to others.

See here

Once I finish the posting, the summary will be:

Dual N-JFET 2SK129A --> LSK389B
Single N-JFET 2SK163 --> LSK170A

which are both in current production.

Yes it is ! - thank you !
 
Yes it is ! - thank you !

Thanks for your encouragement guys. it's appreciated especially since it took some time to do the measurements properly by hand without a curve tracer (I don't trust those). I've just finish off the data for the dual FET sub in that other thread linked a couple of posts above. But the summary is:

(old) 2SK129A: Idss ~7.5mA (Vds=10V,Vgs=0V), Gm ~ 18 (Vds=10V,Vgs=0V), Vp ~ 0.68 V (Vds=10v)

(new) LSK389B: Idss ~7.8mA (Vds=10V,Vgs=0V), Gm ~ 21 (Vds=10V,Vgs=0V), Vp ~ 0.62 V (Vds=10v)
 
Amazing work happening here.

No other restoration thread that I've looked at has covered the what and the why of a restoration as well as this thread has so far.

That thread on LSK389 is truly above and beyond...
And it's info I've been wanting for a long time.

Bravo X 1000!
 
Here is a pic of the parts for the 4 drivers boards. Just for the driver board! Actually, it's not really even for 4 boards... I ordered a full set of parts for one pair of boards and the other pair of boards is just getting the black flags replaced (so I can do a stock vs restored comparison as properly as possible). Two of the silver packets are actually spare On-Semi outputs too.

Just looking at the amps now and these drivers boards look like just 1/5 of the amp... so a lot of work to do here. But I guess the driver boards will be the key job to a happy ending...

fU10HIy.jpg
 
Here is a pic of the parts for the 4 drivers boards. Just for the driver board! Actually, it's not really even for 4 boards... I ordered a full set of parts for one pair of boards and the other pair of boards is just getting the black flags replaced (so I can do a stock vs restored comparison as properly as possible). Two of the silver packets are actually spare On-Semi outputs too.

Just looking at the amps now and these drivers boards look like just 1/5 of the amp... so a lot of work to do here. But I guess the driver boards will be the key job to a happy ending...

fU10HIy.jpg

"Boldy going where to man has gone before." Well...outside of Mr. Amp8 and the few other Japanese techs who have posted their X11 work online.

Since there is no service manual to refer to it's all the more fun. Nothing like a good mystery!
 
Thanks for your encouragement guys. it's appreciated especially since it took some time to do the measurements properly by hand without a curve tracer (I don't trust those). I've just finish off the data for the dual FET sub in that other thread linked a couple of posts above. But the summary is:

(old) 2SK129A: Idss ~7.5mA (Vds=10V,Vgs=0V), Gm ~ 18 (Vds=10V,Vgs=0V), Vp ~ 0.68 V (Vds=10v)

(new) LSK389B: Idss ~7.8mA (Vds=10V,Vgs=0V), Gm ~ 21 (Vds=10V,Vgs=0V), Vp ~ 0.62 V (Vds=10v)

Your comparison entitlted "
2SK129A dual N-JFET & 2SK163 single N-JFET specs and replacement" at
http://audiokarma.org/forums/index....3-single-n-jfet-specs-and-replacement.746462/
is gold!

Thank you Smurfer77!

giphy.gif
 
Your comparison entitlted "
2SK129A dual N-JFET & 2SK163 single N-JFET specs and replacement" at
http://audiokarma.org/forums/index....3-single-n-jfet-specs-and-replacement.746462/
is gold!

Thank you Smurfer77!

giphy.gif

Well as a fan of 16-bit entertainment, thank you! That reminds me, I should post the data on that single JFET sub for the AU-X11. My motivation to write it up was a bit low since that single JFET isn't used on the AU-X1 so I thought not many would be interesting, but probably the data is relevant to other amps (and I guess any data on single JFETs can be useful for making up subs for dual JFETs too). So I will get onto that...
 
Interesting find on the sv3 and 4h subs, will have to give the 1n4150 diodes a try next time I order parts refills.
Also, tried using an lsk389 sub for the original sony sk97fet and was not happy with the results, (as far as I could measure not enough gain to drive the rest of the amplifier driver chain in a AU-717) so the search will continue.
 
Interesting find on the sv3 and 4h subs, will have to give the 1n4150 diodes a try next time I order parts refills.
Also, tried using an lsk389 sub for the original sony sk97fet and was not happy with the results, (as far as I could measure not enough gain to drive the rest of the amplifier driver chain in a AU-717) so the search will continue.

Shame about LSK389 for the 2sk97.

Perhaps its ok for 2sk129 alone.
 
Interesting find on the sv3 and 4h subs, will have to give the 1n4150 diodes a try next time I order parts refills.
Also, tried using an lsk389 sub for the original sony sk97fet and was not happy with the results, (as far as I could measure not enough gain to drive the rest of the amplifier driver chain in a AU-717) so the search will continue.

It seems unlikely an LSK389 wouldn't have enough gain; they range from 8 up to 20mS transconductance which is pretty decent for a JFET. If the specs over here are correct then the 2sk97 transcodunctance is 6 to ? mS.

Which version of the LSK did you have? The LSK389 idss ranges from 2.6 (lower range of LSK389A) to 30mA (upper range of LSK389D) which I think is more likely to have been your problem. That said, I've not played with a 2sk97, but if someone wants to send me one I would be happy to fully characterize it as I did for the 2sk129A.

Cheers
 
Back
Top Bottom