Yet another AU-999 restoration - with pics (now showing ;-)

Picked this 1972 AU-999 up a few month ago. Had some issues - popping in the left channel and very hot outputs on the left as well - The minute I touched the bias for a left turn, I got magic smoke. Outputs had already been changed in the left side, and the work done looked suspicious at best. Here is a first look:



Time to build her back up nice and clean. The less than humble filter caps are some brawny 12k/100v. Only option for a 40 mm fit to bracket in CDE 38x series.



First up for redo is the bottom - I remove the subsonics, crowbar and install a new 400v/25A rectifier - all else down here is recapped with Nichicon PW series. New transistors on the minus and plus ripple filter blocks and some 2.2 / 250 Polyprop bypass for good measure.



I pay extra attention to matching transistors on this one. The two differential pairs are matched within 1 % and set aside.



Then new outputs, mica and grease. Only two original transistors were left, so I replace the entire set with some rugged onsemi's MJ21194's



Next up are the driver boards. All new trimmers, diodes, caps, transistors and a Mundorf Polyprop for the input coupling. The bipolar feedback cap gets extra attention with a nice Nichicon Muse at highest voltage possible. (More on that here) Pre-drivers are robust 1220's/2690's.



Some detail pics....







Finally done - but not without a battle. One driver board keeps on blowing fuses. After having checked transistor pin-outs, measured all resistors, I turn to the traces where I continuity test and find two separate culprits. Both hidden from view under a seemingly sound looking blob of solder. In both cases the foil base have ripped from the copper trace right before the blob, like a hairline crack.



The remaining boards are recapped with Silmic II's and repopulated with ztx795a signal trannies, except for the head amp where two 458's are swapped with ksc1845's.
In the end, Dc and bias smoothly locks in. I got 1-turn trimmers, thinking the problem with old trimmers weren't the one turn, but rather the 40 year old pots acting up from age. I was wrong. Next time I'll go for 5 turns.

Close to 90 parts swapped.....



Cosmetically she is a real beauty. Not much to do here really.





LISTENING:

I first give it a run without the pre-amp / bass EQ mod. I am a tad disappointed. It is a good sounding amp, no doubt, just not quite what I thought. On a positive note, the amp is very quiet.

But the battle is not yet over. We still have the mods, so off to the bench.

The Pre-amp mod was rather straight forward. Here is the EQ bass mod...Oh boy oh boy - this is not hangover day work. Steady now...steady.



A good hour later and WOW - a completely different animal - Kevin - Kudos to you. That pre amp and bass eq mods are brilliant little engineering hacks that makes her sing to the high standard she was build.

This is a truly great sounding amp. No lack of bass output to my ears, super sweet mid and a great separation well into the higher notes.

Great work, looks awesome. I see you ripped out what appears to be the protection board? I would too, worthless piece of engineering.

Really admire the beautiful work you've done here. I happen to love the AU-999, as when I was a kid, I could only afford the AU-777. You are lucky to have such a beautiful piece of gear, and one that sounds great as well.

Thanks for the great pictures.

I love the look of flying saucer transistors. I'm thinking of a way to re-stuff them with modern replacements, like electrolytics

stereofun said:

The bipolar feedback cap gets extra attention with a nice Nichicon Muse at highest voltage possible. (More on that here).

Click to expand...

Very nice work and pictures; thank you for posting.

For the low-end cut-off capacitor, C807, did you replace with a larger capacitance value (e.g., 100µF) to obtain a F3 <2Hz (Doug Self recommendation)?

Regarding the negative feedback loop proper, I have found that replacing the feedback resistor (R819) to a 0.5W metal film type improves the sound noticeably; Sansui often used a resistor rated at 0.5W for its better temperature coefficient in later designs.

Also, in the schematic, there is a small 56pF ceramic capacitor (C831) in parallel with R819 for phase compensation, but it doesn't appear in your pictures of the driver board, is it on the back of the board? If present, this capacitor should be replaced with a better quality type (e.g., polyproplyene or C0G). Similarly, replacing the low pass ceramic capacitor (C803) with a better type would noticeably reduce distortion; Sansui often spec'd polystyrene in this position in the later designs.

Thanks for the kind words Leestereo, I had assumed the C807 is the feedback cap - at least it appears to be pr. block-diagram, where the loop goes back to the base of TR803, via R819 - however, I am no shark at reading schematics !. At any rate I kept the value at 47uf, but bumped the voltage to 50V. I do not have a C831 56pf. The only 56pf pr. service manual is C803, which I do have on the front of the board. Nothing on the back.

Thanks for the tip on the R819, next time in there i'll bump to .5 watt.

I'm sorry if I confused the issue, C807 capacitor is in the feedback and it determines its low end cut-off of the loop; with 47µF, the F3 is ~2.8Hz. And you are also correct that in using a higher voltage rated capacitor, you are lowering the distortion level; the lower F3 (<2Hz) of a larger capacitor will also lower the distortion a bit more.

Very nice.

quality work Tom. The big filter caps look great! And that is possibly the best pic of outgoing parts I've seen - nice arrangement. Do you domino them a couple of times before taking the pic?

About the variable resistors, I really like the Bourns multiturn devices. My hand isn't steady enough for the single turn...

Nice job there Tom. I like my AU999, along with my Eight that I run mostly, I know, you used to have one too, :yes:.

I think you will find that the AU999 with the Kevzep mods will sound right in line like your Eight. At least by my ears it does.

Crank it up, let it play, have fun!

George.

smurfer77 said:

Do you domino them a couple of times before taking the pic?

About the variable resistors, I really like the Bourns multiturn devices. My hand isn't steady enough for the single turn...

Click to expand...

Yes David - indeed I did. I forgot the outputs, then had to rearrange and the avalanche began :sigh:

GeoKost said:

Nice job there Tom. I like my AU999, along with my Eight that I run mostly, I know, you used to have one too, :yes:.

I think you will find that the AU999 with the Kevzep mods will sound right in line like your Eight. At least by my ears it does.

Crank it up, let it play, have fun!

George.

Click to expand...

Right on George - The 999 reminds me more of my recently departed Eight, than it's 777a sibling, that is distinctly more tube sounding. Still haven't had too much listening time - but so far I really like how it sounds when cranked up a bit. Here it keeps steady, maintaining great separation and poise - even if it doesn't excel in one category. The 777a has more presence and depth, the Eight more top end detail and punch - but the 999 just rocks on in a very listenable way. It is a really nice amp.

Nice job Tom, well done......:thmbsp:

Great pics too....

Aren't they a joy to work on, nice and modular and logically laid out.....

Glad your enjoying the fruits of your efforts....

I just recently sold my AU999 to a good friend of mine, he loves it!!

stereofun said:

I pay extra attention to matching transistors on this one. The two differential pairs are matched within 1 % and set aside.

Click to expand...

Awesome job! :thmbsp:

Thanks for another great thread on the 999! Am just finishing off my restoration and recap.
In response to Leestereo
"Also, in the schematic, there is a small 56pF ceramic capacitor (C831) in parallel with R819 for phase compensation, but it doesn't appear in your pictures of the driver board, is it on the back of the board? If present, this capacitor should be replaced with a better quality type (e.g., polyproplyene or C0G). Similarly, replacing the low pass ceramic capacitor (C803) with a better type would noticeably reduce distortion; Sansui often spec'd polystyrene in this position in the later designs."

I have what looks to be 15pF small ceramic caps in parallel with R819 on the back of the driver boards on my amp. Wondering if I should stick to this value and just replace with polyprop or polystyrene? While I am at it will make the other tweeks that you suggested for the driver boards Leestereo.. I have done all the other mods and replaced my volume control and very happy so far

afroaudio said:

I have what looks to be 15pF small ceramic caps in parallel with R819 on the back of the driver boards on my amp. Wondering if I should stick to this value and just replace with polyprop or polystyrene? While I am at it will make the other tweeks that you suggested for the driver boards Leestereo.. I have done all the other mods and replaced my volume control and very happy so far

Click to expand...

I suggest that you replace with a 15pF capacitor, and since it is on the back of the board, it might be easier to use a C0G type. But, if you decide to use a polystyrene type, note that they are very sensitive to heat (use a heatsink when installing).

Very nice write up!

Nice work!

I might just add that 15pf on to my R819, since I dont have it. Afroaudio, if you dont mind, whats your serial number on the back, the first 5 digits - curious if yours are newer, and the 15 pf something they added to later designs.

Keep in mind you have no more than 5-8 mm on the back before the driver shield, wondering if front mounting would be easier depending on size ?

Another mod. I am contemplating is one from AK member trnsfrm. however, instead of adding electrolytics to the back of the mylars, I'll replace the mylars with polyprops and double in uf value.

I'll also replace the original drivers with onsemi mje15032/33. Since I already have mj21194's as outputs. In my Eight, those new transistors were an upgrade, better tweeter response, but not until I included the drivers.

The sound is excellent, albeit I 'll pursue a tad more resolution tweaking forward. Too me it is the kind of amp where the excellent tone controls are there to be used ad libitum - that is otherwise too flat for me

Hey Stereofun my serial no is 22 11 11 242. Have not looked into what date this means. Interested to know about the 15pF cap though! I see that cap is present on the schematic from Hifiengine and is named C819.

I have tracked down some NPO/COG replacements which may make a small difference I guess - at least peace of mind. Once someone tells me there may be even a tiny improvement by upgrading a component I am compelled to do it!

Not wanting to hijack further but am slightly confused about the procedure for DC offset and bias adjustment and what order to do these for the 999.. I need to visit my friend with a dummy load resistor on Friday and get this dialled in.

Hey - I am currently listening to my modded 999 - my favourite amp - I have done all these mods also (including the EQ and bass-boost resistor mods) - except for replacing the outputs - was there a substantial benefit in terms of sound improvement do you think in fitting the Onsemi outputs - I'm a bit loathe to replace my factory outputs as they were obviously selected by Sansui for a reason, and I would hate to lose the characteristic sound of this amplifier - however is this something that you would recommend?