AU-517 recap and clean-up (with nudies!)

[pete_mac]

Thanks to the skills of skippy124 (my father) I've now got a nice recapped/cleaned/deoxit'd AU-517 which is about to take over the main duties in my system. It will be replacing a AU-417 so whilst the general cosmetics are almost identical, the AU-517 is much beefier overall.

We used a mix of caps due to certain stuff not being readily available in Australia - a mix of Panasonc FC and FM, some Panasonic audio caps, and some Nichicon where I couldn't source Panasonics.

Prior to cleaning the amp (still dusty) with the driver boards removed:

The dreaded glue (before and after). Dad fashioned an EXCELLENT glue removal tool using a long strip of PCB material, with one end sharpened into a chisel-like implement. It made short work of removing the glue. As you can see, the glue had corroded some components:

Finito! Apart from some very minor scratches on the top edge of the front fascia, the amp is in WONDERFUL condition. DC offset is set to pretty much 0mV, bias is 20mV (was 20mV offset and about 60mV bias before adjustment).

Best of all, she sounds great! After a few hours of testing on small speakers to verify proper operation, I will open her up on some big'uns. Can't wait!

Thanks again Dad. :thmbsp:

 

[EPI-Center]

very pretty...great work...I have the 719 and I absolutely love the sound.

 

[lha1992]

Thats a very nice recap job ya'll did!

And one very nice AU-517!

Congrats

 

[RHX68]

Wow, what a nice job! Congratulate with you lovely 517.

Try set the idle bias current of o/p transistors to 80-90mA range (rather than 20mA as claimed in service manual). And you'll kiss this amp.

:yes:
Cheers,
Rudy

 

[pete_mac]

Wow, what a nice job! Congratulate with you lovely 517.

Try set the idle bias current of o/p transistors to 80-90mA range (rather than 20mA as claimed in service manual). And you'll kiss this amp.

:yes:
Cheers,
Rudy

Hi Rudy,

thanks for your comment.

Can you please describe the benefit of doing this? That is, what kind of audible improvement can I expect to hear?

Would I be correct in assuming that this would result in more heat being generated?

If any other Sansui gurus would like to chime in, that would be fantastic too

 

[scottrt]

Try to kiss it at that bias setting and you'll burn your lips. I think that is insane advice.

Scott T.

 

[EdibleHead]

great job Pete. Looks very similar inside to my AU-717 that I believe we bought from the same original owner?

Mine was unmarked and original and a little dusty. I have so far opened it up and gone over carefully with a vacuum then cleaned and lubed all pots and switches with Deoxit.
I was wondering about setting Bias and DC but I have no skills to do this. I might take it to a Tech one day but for now it sounds pretty good to me.

 

[LBPete]

It will certainly heat up at 80-90ma! I wouldn't recommend it.

It is interesting to see the back of an Australian Sansui. It is a single input voltage, 240v. The Japanese domestic models are all 100v. Most of the stuff sold here in the US is multi-voltage but some of the high production units like the 9090DB had some single input voltage models. I've never seen a 240v single voltage model.

- Pete

 

[pete_mac]

great job Pete. Looks very similar inside to my AU-717 that I believe we bought from the same original owner?

Mine was unmarked and original and a little dusty. I have so far opened it up and gone over carefully with a vacuum then cleaned and lubed all pots and switches with Deoxit.
I was wondering about setting Bias and DC but I have no skills to do this. I might take it to a Tech one day but for now it sounds pretty good to me.

Correcto! Same seller

If you have a digital multi meter you can quite easily check the offset and bias. The procedure and the measurements are outlined in a stickied thread in the Sansui section IIRC.

Depending upon your intentions for the 717 (ie. keeping it for a long time) it could be worth getting the capacitors replaced and the notorious glue removed. I was amazed at the extent of damage/corrosion caused by the glue. Based upon what we saw in the 517, Dad and I will be recapping the 417 and 317 in the near future.

 

[skippy124]

Nice Amp Pete!!! (but I am a bit biased).

It is interesting to see the back of an Australian Sansui. It is a single input voltage, 240v. The Japanese domestic models are all 100v. Most of the stuff sold here in the US is multi-voltage but some of the high production units like the 9090DB had some single input voltage models. I've never seen a 240v single voltage model.

The service manual shows (supposedly) for the AU517/AU717 that the US got dedicated 120VAC, and Europe and Great Britian (and countries that had the same safety scheme as Great Britian at the time eg Australia) got a 220/240V switch selectable, and anywhere else got the 100/120 220/240 plug selectable version except Japan which got the 100V dedicated (as you have stated). IIRC the one in this thread (Australian delivered) has the 220/240 switch selectable input.

 

[RHX68]

>>>>>>>>Try to kiss it at that bias setting and you'll burn your lips. I think that is insane advice.

Scott T.

Scott, no offense...... Insane? Have you tried this setting ?

It's possible (and I think it's reasonable) to bias the o/p transistor in a linear region while the it still within SOA.

Sansui recommend 20mA bias current seems to be a trade-off of company sales/marketing policy rather than optimise the design for audiophile's favour.

I enjoyed this setting for over years and did no harm on my AU-517. (You should understad take this advise is actually for your own risk, where no one really want to bear responsibility).

Cheers anyway,
Rudy

 

[ferninando]

my 517 looks zakly like yours with no disassembly/recapping and sounds AWSOME. What else? It's a Sansui.

 

[OverLoad]

Wow, what a nice job! Congratulate with you lovely 517.

Try set the idle bias current of o/p transistors to 80-90mA range (rather than 20mA as claimed in service manual). And you'll kiss this amp.

:yes:
Cheers,
Rudy

........... It's possible (and I think it's reasonable) to bias the o/p transistor in a linear region while the it still within SOA.

Sansui recommend 20mA bias current seems to be a trade-off of company sales/marketing policy rather than optimise the design for audiophile's favour.

I enjoyed this setting for over years and did no harm on my AU-517. (You should understad take this advise is actually for your own risk, where no one really want to bear responsibility).

Cheers anyway,
Rudy

No offense either but I think you've been reading the TNT "put a tiger in your amp!" rubbish a little too much.

Just out of curiousity, where did you come up with the manual claiming 20mA of bias current? In actuality it claims 30mA :yes:

So Sansui seems to have recommended this mysterious 20mA (still like to know how you came up with that) figure to be a trade-off of company sales/marketing policy? There is a lot that goes into an amplifier design and I highly doubt a designer picked a bias current out of a hat or based on sales/marketing.

Lets assume the above was even true, then would it not be logical to also assume that there are other "trade-offs"? Lets take the transistors SOA you mentioned for example. Part of a transistor's SOA is much more than current limit. Another part of the SOA equation for example is power dissipation and costly heatsinking that goes along with that. Were the heatsink spec's picked out of a hat or a trade-off as well? I'd say they were carefully spec'd for the overall design criteria of the amp that involves many factors and without a great deal of extra margin.

It's great you may have had your bias set for 80-90mA for years without problem. I haven't crunched the numbers but I will go out on a limb and say that increasing the bias current threefold certainly wasn't part of the designer's equation and I certainly would not recommend it. I think it boarders on recklessness to simply suggest this to someone IMHO. More often than not, you'll be kissing your amp alright, kissing it good-bye.

 

[Hyperion]

No offense either but I think you've been reading the TNT "put a tiger in your amp!" rubbish a little too much.

Just out of curiousity, where did you come up with the manual claiming 20mA of bias current? In actuality it claims 30mA :yes:

So Sansui seems to have recommended this mysterious 20mA (still like to know how you came up with that) figure to be a trade-off of company sales/marketing policy? There is a lot that goes into an amplifier design and I highly doubt a designer picked a bias current out of a hat or based on sales/marketing.

Lets assume the above was even true, then would it not be logical to also assume that there are other "trade-offs"? Lets take the transistors SOA you mentioned for example. Part of a transistor's SOA is much more than current limit. Another part of the SOA equation for example is power dissipation and costly heatsinking that goes along with that. Were the heatsink spec's picked out of a hat or a trade-off as well? I'd say they were carefully spec'd for the overall design criteria of the amp that involves many factors and without a great deal of extra margin.

It's great you may have had your bias set for 80-90mA for years without problem. I haven't crunched the numbers but I will go out on a limb and say that increasing the bias current threefold certainly wasn't part of the designer's equation and I certainly would not recommend it. I think it boarders on recklessness to simply suggest this to someone IMHO. More often than not, you'll be kissing your amp alright, kissing it good-bye.

Agree 100%

Reckless mumbo-jumbo.

Stick with manufacturers recommendations is my recommendation too.

Sorry RHX68 - I think you are wrong to suggest such a thing.

 

[pete_mac]

Cheers for the feedback gents.

Just out of curiousity, where did you come up with the manual claiming 20mA of bias current? In actuality it claims 30mA :yes:

Hmm... the copy of the combined AU517-717 service manual that I have lists 20mV, as does the link in the stickied bias current and DC offset thread here:

http://www.audiokarma.org/forums/attachment.php?attachmentid=82207&d=1205129320

http://www.audiokarma.org/forums/showthread.php?t=144580

If it is indeed 30mV, please let me know your source, as we've set the bias to 20mV based upon the above.

 

[OverLoad]

Cheers for the feedback gents.




Hmm... the copy of the combined AU517-717 service manual that I have lists 20mV, as does the link in the stickied bias current and DC offset thread here:

http://www.audiokarma.org/forums/attachment.php?attachmentid=82207&d=1205129320

http://www.audiokarma.org/forums/showthread.php?t=144580

If it is indeed 30mV, please let me know your source, as we've set the bias to 20mV based upon the above.

YES, that's correct it DOES says 20mV. However, lets be clear that your actually setting a "current".

In the SM procedure your measuring the bias "current" indirectly by way of measuring a "voltage" (mV's) between 2 test points. Those 2 points are across a pair of .33Ω emitter resistors connected from emitter to emitter of the output transistor pair. Since we know the voltage (20mV) and the resistance (.33+.33Ω) we can calculate the current at 30mA.

 

[pete_mac]

Ah, righto! I understand

 

[ferninando]

the svc mnl says 20mv across the terminals. NOT 20ma of current.

 

[RHX68]

>>>>still like to know how you came up with that?

Both the NPN and PNP beta's are not in good match during mass production. I doubt that Sansui will match it one by one. And after out of factory, the environmental factors which drift the parameters will also come up with trouble/complains that Sansui didn't want to deal with.

Provided you check that all the NPN & PNP are in good matching, or this bias increasing will also drift the final DC off-set (Another potential trouble to kick in the relay). Setting the final output transistor bias current to 80mA will also put the pre-stage transistor into a region with more linear hfe. Hence lower cross-over distortion.

And I definitely can hear the difference.

But I don't recommend a newbie to try this without good measurement and matching of their transistors.

Anyway, keep it at 20mA/30mA, it's factory recommendation and pretty good for you. Just enjoy the AU-517, it's a marvelous all round musical amplifier.

 

[pete_mac]

I undertook some further recapping of the AU-517 today, with Elna Silmic II caps throughout the few remaining boards that we didn't tackle first time around.

The dining table/work area:

Inside the beast with all internal trim panels removed:

Recapped boards:

Phase three will involve the replacement of the fuse resistors, the bias and offset trimpots, and maybe one or two caps on the driver boards with Elna Silmic II caps as well (I previously used the blue Panasonic ECA audio caps in one or two spots, so I might substitute those for consistency).

I've also got some Panasonic THA 15000uf 63v caps coming that I will probably fit too. It looks like it will be a bugger of a job to get these looking 'right', as I will need to solder some ring terminals onto the 'snap-in' pins on these caps in order to make everything marry up properly.

 

[pete_mac]

Did you or are you planning on replacing the 4 caps on F-2670 Pre-Main Switch Circuit Board? I missed those on my AU-517 until EW pointed it out. So I have an order for my L snap rivets (plastic) and capacitors that I'll be waiting on to finish my project. Looking good.

At this stage, I'm not going to bother.

I bought some replacement snap rivets and was all ready to do the job, but after reading EW's explanation of what these caps do, I don't think it is necessary for me to change these caps with any great urgency.

I only ever intend using this unit as a preamp for a larger power amp by selecting the 'separated' option via the switch, or as a integrated amp by leaving it as 'connected-direct coupled'. The 'separated-capacitor coupled' option is not relevant for my intended use of this amp, so replacing these caps is fairly low on my agenda.