Sansui AU-717 TO3-P conversion

[Sui-Yogi]

Hello all ,
Lots of respected , educated techs in the forum has stated that there is no concern replacing the output transistors of an AU717/517 with other than Onsemis , the difference in the Mhz should be under consideration for “ faster “ amps . Others equal respected and educated are stating the opposite . I have NO clue and nothing to argue about !!!!
I’m just here cause I like the journey, feeding my enthusiasm for audio and particularly Sansui sound and last but not least enjoying the feeling of accomplishing a task.
For that and only ,I’m spending my weekend trying my best implementing work that Hyperion did in a BA-F1 in my AU717 hoping that I would not “verbalise Jack Rodinson “rather manage to go "(A)round Jack Robinson's barn" by the end of that particular “journey” .
The suggested TO3P complimentary pair are the :
2SC4468/2SA1695
I couldn’t find those from a trusted source so I’m going with the little more robust :
2SC3856/2SA1492 .
I have to thank @GrisPato for his generous offer to post me a couple of pairs for the C2344/A1011 for the drivers and a bunch of KSC1845(marched please) ….. Pato , I owe big !!!!!
I’m not gonna use this ones at the moment , I ll leave the MJE15032/33 and use the Sonys if all end up well .
Here is where I’m standing at the moment , paying attention to surgical details .

 

[GrisPato]

Nice work!
Watching with interest because I want to do this with one of my AU-D707 or may be even one of my AU-D907 Limited and then compare it with one with original outputs. So keep us updated

PS
The driver transistors are not Sony's but from Sanyo, considered the best audio driver transistors by many in the DIY community I was told!

 

[Sui-Yogi]

Nice work!
Watching with interest because I want to do this with one of my AU-D707 or may be even one of my AU-D907 Limited and then compare it with one with original outputs. So keep us updated

PS
The driver transistors are not Sony's but from Sanyo, considered the best audio driver transistors by many in the DIY community I was told!

Since I was a kid I was always confusing Sanyo with Sony ….. but yes you are right the transistors have great reviews !
Thank you again for getting your self into the trouble posting and for the offering !
I will keep you posted , for sure but first it needs to work and as I see now things are pretty tight around and it’s soooooo easy to crest a shortcut ….. let’s see ….
Honestly I don’t believe that it will be a deference in sound performance , very serious technicians that count their words carefully believe that the onsemis are doing the work perfect in this amp !

 

[Sui-Yogi]

Here is where am at , doesn’t look great But not bad either …. For sure nothing is touching nothing that is not supposed be teaching it …..

 

[Sui-Yogi]

Time to call it a day , finish with the right channel , DBT checked fine , bias and DC offset on point , I have sound and I did install Patos’s Sanyo drivers !
I have a strong believe that tomorrow I ll hear the real deal !!!
Picture below with the end result .

I have a question though .
Is there a way with a multimeter to check if the amp is not drifting to oscillation or even if it’s safe to tern up the volume with out fearing that something will “blow”?
I understand that bias, DC offset and voltages according to the service manual is the obvious to check but all this is at idle , what if you tern up the volume or even what if the amp is working continually for hours ?
Is there is something other than the above that something is not absolutely right ?
I know that the question is a bit “ stupid “ , that’s why the oscilloscope s are existing and the dummy loads ….. but I don’t know,might something more is there that someone can check so It can be at the safe side or at least safest possible .

 

[DougBrewster]

If you have a good quality multi meter you could monitor the output for ac volts/mV and frequency but it's a very crude method. It's all you've got I guess. My Fluke is good for up to 100KHz but considerably attenuates it's reading after that.

 

[GrisPato]

Looking good!

the amp is not drifting to oscillation

I don't think you have to worry to much.
The 2SC3856/2SA1492 haven't the really high fT as the later alpha models. They are just 20 fT, the same as the 2SA1105/2SC2580 used in the AU-417, which driverboards are the same as the AU-517/717 only a simplified circuit.
The Sanyo drivers have the same fT as the Toshiba 2SA968/2SC2238, 100 fT, these where used by Sansui starting with the AU-D907 Limited (1979), so no reference for that with 'slower' designs.

 

[Sui-Yogi]

If you have a good quality multi meter you could monitor the output for ac volts/mV and frequency but it's a very crude method. It's all you've got I guess. My Fluke is good for up to 100KHz but considerably attenuates it's reading after that.

Unfortunately I have the cheapest you can get !
Maybe I need to consider upgrading my game , imagine I made DBT just yesterday ….. enough is enough I need to do something about testing equipment !
There is a nice tread ( I can’t find it now ) from “ “ecowars” , mentioning start up equipment, the very basics ….

 

[GrisPato]

There is a nice tread ( I can’t find it now ) from “ “ecowars” , mentioning start up equipment, the very basics ….

Is on top of the DIY section here on AK

 

[GrisPato]

But I have to agree, for the experimenting you are doing you will need a oscilloscope and some dummy loads to be able to verify what you are doing is working out.

 

[Sui-Yogi]

Looking good!



I don't think you have to worry to much.
The 2SC3856/2SA1492 haven't the really high fT as the later alpha models. They are just 20 fT, the same as the 2SA1105/2SC2580 used in the AU-417, which driverboards are the same as the AU-517/717 only a simplified circuit.
The Sanyo drivers have the same fT as the Toshiba 2SA968/2SC2238, 100 fT, these where used by Sansui starting with the AU-D907 Limited (1979), so no reference for that with 'slower' designs.

Make sense .
I was considering placing the KSC3503/KSA1381 for the TR5/7/8/11 now that I have the Sanyo installed , I have installed now the TTC04/TTA04. Maybe is a better mach in this combination;

 

[Sui-Yogi]

Left channel done too !
It took me the 1/3 of the time that I used for the right Channel , ones you find the way it become a standard procedure.

 

[Sui-Yogi]

I’m now 4 hours listening the playlist I’m listening everyday for the past 6 months through the AU717 with the TO3-Ps output transistors.
I have to remind to all, that they might read this, that my intentions posting my findings are to put something “on the table “ in “exclusively Sansui” forum rather than argue or doing the smart ass. As I got help ,I’m trying to give as much possible to the extend that I can. I have a big respect to the people in the forum, I’m not a tech, My native language is Greek.
I do not want to be misunderstood , I’m NOT here to guide or suggest ANYTHING.
I’m here to just tell you what I did and what was the results , I’m doing my best writing in English so if you don’t understand something use your imagination. It can help .
That been said , after the conversion to TO3-Ps , there is a deference in sound reproduction not quality . I do not have the vocabulary to give you a picture and I’m lacking knowledge too and this description would demand both so let me give an example:
If I had to try to explain to a 6 years old kid the transformation before/after Leestereo upgrade/update , I could easily do it in Greek and most likely in English .Same goes for cleaning switches/ potentiometers . I would have been able to find words to support my sayings/listening .
I this case I can’t , the transformation in sound reproduction is mild but present I can’t tell with certainty if it’s at the highs, mids or lows frequencies . In specific type of music is just more present than others .
If I had a gun in my head , someone threatening me to pull the trigger if I’m not gonna describe where I listen the deference, I would take a guess and say in the high frequencies and that shows my ignorance , I do not know how a faster transistor or amp should sound in comparison with a “slow” one .
I have to say with certainty that Kezvep is correct , there is auditable difference.
Echowars ,that suggest if you have the original transistors don’t replace them, if not ,put the Onsemis and forget about it . Correct !
I wouldn’t have miss the AU717 “magic” with the Onsemis.
Bottom line , if the amp is working with the originals don’t touch it, if not and you are comfortable with the work that needs to be done for the conversion then choose the TO3-Ps , if you are not feeling comfortable, then have the Onsemis in , there is no magic hiding there .
Other thing that I’m sure of is that this is not an upgrade, I can hear it , it’s just an action bringing the amp closer to the intensions of the Sansuis sound engineers .

Soon I will have my AU517 for a Leestereo treatment and it has the first generation Onsemis in it , after the restoration I will have the chance to do an A/B comparison and might there is something more to tell .

Any ways , sure thing , I enjoyed my time with the conversion, I’m happy and I’m listening my music through might ,one of the best; amplifiers ever realised in the market ……
Cheers
Kostis

 

[Leestereo]

@ARGOU: Kudos to you for your enthusiasm and willingness to experiment!

...to check if the amp is not drifting to oscillation or even if it’s safe to tern up the volume with out fearing that something will “blow”?...

When an amplifier is in oscillation, it is not uncommon to find that the output transistors/heatsinks heat up considerably .

 

[Sui-Yogi]

Thank you Mr. Ben.
It’s the second day now that I’m listening to the amp , it’s for sure not the same reproduction of sound compare to the Onsemis , Kezvep; is right and I feel someone should mention that, as he spend time busting his kidneys trying to prove he is not an elephant .
I have to mention that in the drivers I installed the Sanyos that GrisPato offered me , so might be a combination; this deference in sound .

There is no heat at the OP transistors / drivers the heat is at the same level as it was with Onsemis . The pre-drivers are getting hot (TR5/7/8/11) with the NPN’s getting pretty hot but it was like this before the conversion and even in stock condition. I assume that this is normal;
I even had the chance to crank it up to 12oclock today for an hour , when the neighbours left for work , everything looks like are well insulated with more voltage running through the OP transistors and they heated up accordingly; to the demands of the deafening level challenge .
So I don’t think I have a problem , let’s do not forget that in “Sanken discontinued list” is suggested the TO3-Ps (2SA1695/2SC4468) as a replacement of the TO3 (2SA747/2SC1116) I suppose they wouldn’t suggest something that even in one case , can cause oscillation…. Just me thinking now ….

Soon I will have my AU517 in my hands for a Leestereo update/upgrade and I will keep the Onsemis in for an A/B comparison with the AU717 , then I might find the English vocabulary to describe thoroughly what I hear deferent.

PS: For people like you , and a lot of others highly skilled in electronics this conversion is piece of cake , it will not take you more than 20min to replace all 4 OPs , it took me 20h .

 

[Leestereo]

...I have to mention that in the drivers I installed the Sanyos that GrisPato offered me , so might be a combination; this deference in sound ...

Have you tried the MJ21193/MJ21194 output transistors with the the Sanyo driver transistors? IMO, these ON Semi transistors are appropriate replacements for the original 2SA747/2SC1116 output transistors. The ON Semi transistors have the advantage of utilizing advanced perforated emitter technology (similar to multiple/ring emitter technology), which is considered to be SOTA for BJTs. And although the fT of the ON Semi transistors are lower than the original counterparts, this is fine since they are of the same order of magnitude, i.e., 6MHz vs. 15 MHz and 7MHz vs. 10MHz.

 

[Sui-Yogi]

Indeed I thought about it but I was to enthusiast and couldn’t help my self doing the experiment Sanyo/Onsemis combination .
I will perform the experiment in the AU517 !

I have to admit that this “transistors rabbit hole” I have found my self in is not very fun any more …it’s a bit mental , I’m just an enthusiast hobbyist so I’m kinda loosing track of my self here …
So last night I pull my self together and form a question to my self that I owe to answer in good time .
Who am I ?
If I’m a personality that , FOR EXAMPLE, concerns and loose sleep over the Cob of the pre-drivers transistor thinking that the originals have 6pf, the one I have instal are the TTC/TTA and have 17pf vs the KSC3503/KSA1381 with 3pf and I’m willing to spend money buying the last ones , pay shipping fees, wait , uninstall, reinstall etc etc then the only logical way for me is to install for the OPs what Sanken is suggesting .
If I’m so conserved about this Cob and finally I install Onsemis then I’m bipolar as much as the transistors I’m using!
Now if I’m perfectly fine with the Great Numbers that the TTC/TTA from Toshiba provides I can certainly live with Onsemis in the OP stage .
Simple as that ! Who am I?

And although the fT of the ON Semi transistors are lower than the original counterparts, this is fine since they are of the same order of magnitude, i.e., 6MHz vs. 15 MHz and 7MHz vs. 10MHz.

A simple thought about this ( this is how i torture my self ) :
If the ft of the Onsemis is min. At 4Mhz and arises up to 7Mhs for the needs of the signal reproduction that the topology is capable to provide , then a 16Mhz transistor that the Sansui order by Sanken to be made , where can be arise to? Is this something that the engineers have taken under consideration?
If yes ? Then a 20Mhz unit is closer to the intensions of the Sansui engineers….. that how it’s in my head .

Cheers
Kostis

 

[Sui-Yogi]

The ON Semi transistors have the advantage of utilizing advanced perforated emitter technology (similar to multiple/ring emitter technology), which is considered to be SOTA for BJTs

I’m not yet at that chapter In my “education”
Indeed there is some good writings ( very technical terms and difficult English ) about “horse shoe” vs “ring” emitter technology….. but I’m not yet there .

 

[Sui-Yogi]

A solution , how you fit a flat pin to a cycle socket . Indeed , is not an easy task to solder the golden pins to the transistor a lot of things can go wrong wile you are working with it and might in the long run if the soldered point is not solid .
So an easy , convenient and safe solution to this is to wrap around the pin ( not very tight ) a copper thread from a wire , let’s say speaker wire . I used a silver thread that they use for making jewellery. You need to wrap around the thread , not very tight , to the pin , solder it along the way of the pin and then file the corners with a 220 grid sandpaper….. and there you have TO3-P transistor fitting perfectly in a TO3 socket !