Sansui JPD adventure

[Hyperion]

Just will have to get an understanding of the functioning of my scope

Set the 'coupling' to AC, and start off with say 10V/cm - then increase the sensitivity (to say 50mV/cm) while keeping the trace on the screen. If your scope doesn't automatically detect the probe setting (x1 or x10) you will need to account for that.

 

[sansui au719]

An interesting read so far. When you complete your restoration it might be interesting to some to hear more about your experience of buying & importing directly from Japan.

 

[GrisPato]

Some more pictures. Some glue, some weird resistors and filmcaps reaching to the sky. . .
And did some probing, right channel all 4 outputs give short, left normal. So I am 50% down on the nma/nmc 1012

 

[JoseHH]

Thai is a very nice amplifier. Too bad half of the outputs are shorted. Check the drivers on that channel, the emitter resisors on the outputs and also the base resistors on the drivers. In the long term, may be best to use TO-3p as output transistors before they discontinue them. The 2SA2223A is now in "End Of Life" status, so get a set for your amplifier. There are posts by member Ilimzn in this forum recommending the 2SA2223A/2SC6145A for this amplifier.

 

[GrisPato]

@JoseHH
Thanks for the advice. As mentioned before this one is for selling. So I am going to keep it as original as possible because of the preference her in EU.
That's the reason I bought an au-d907 limited and two au-d707's that will arrive in 4-6 weeks. The 907 limited and one of the d707 are going to stay and will get upgraded to modern components. This one and the best looking d707 I will sell in the most original state. I am just hoping I will end having some nma/nmc 1012 left.
I had some Sankens in my basket already, but the others :sc3263.
Would value your opinion why they are less good. Biggest difference I can find is basecurrent at saturation.

sc3263 : 200v 130w Ib at saturation 500mA COB 250 pF
sc6145 : 230v 150w In at saturation 50 mA COB 500 pF

 

[JoseHH]

I am sorry, I did not see that you want to re-sell the amplifier in a close to original form. I understand now. As I said, the recommendation was from a more senior member in this forum so I have no direct experience. Please look at post #76 in this thread to see where I got this information. Some of the earliest restorations work done on the AU-919 you may find in a Croatian audio forum, including the re-stuffing of the oval capacitors is due to Ilimzn. I believe that He and Echowars (a highly regarded contributor to this forum) were the first to suggest the use of TO-3P alternatives to the NMA/NMC transistors.
I purchased a set of 2SA2223A/2SC6145A but I still have not had any chance of installing them in my AU-D907, which is still functioning with a set of MJ21195/6. Hopefully I will change them to the TO-3P parts at some point, since it is generally agreed that this amplifier was designed to work optimally with output transistors with ft in the 30MHz zone. I hardly think I can distinguish the difference, so, given the complexity of this amplifier I have been postponing this upgrade.

 

[GrisPato]

which is still functioning with a set of MJ21195/6.

I hardly think I can distinguish the difference,

Good to hear this. I remember reading this already once or twice. Will give me an option to use the amplifier ( probably the d907 limited ) without having to rush a ,as you say and I agree the daunting task of fitting TO3P in there. I already went over Hyperion's thread where he fits them in the BA-F1. Without having to cut of one of the corners of the tab. This I tried already with some shorted Sankens I have laying around. The tab is copper or copper compound, really soft, went trough it like butter with a hand steelsaw.
The thread you rever to also mentions the sc3263 but recommends the sc6145 as even better. Will see, the sc3263 is at least not on its way out so I probably stick with that one.
What surprised me is his conclusion that the filter capacitors are often bad because of the high voltage from the rectifiers althoug the service manual states 55 Volt, so 63 Volt rating is a 15% margin!
I am going to check that when my step down transformer arrives.
I was more inclined to think that this could be caused by the packaging in the oval containers where they would have difficulty getting rid of the heat and we know heat is the biggest enemy, especially for lifetime.

 

[JoseHH]

Yes, I also cut some corners of the TO-3P transistors, and even drilled holes on them to pass a gold plated pin that i plan to use, but, i stopped there. My amplifier works well after I cleaned it and renovated it but my renovation is not as exhaustive as others have done. I intend to return to this at some point and tackle the output transistors.
With regards to the main power supply capacitors, I did not renovate mine. I did the same ripple test that John (Hyperion) suggested you on mine and the ripple was small. But i did the tests with the power supply unloaded, and I suspect that ripple will be larger when loaded. Old school advice for using electrolytic capacitors recommended a 50% voltage margin, so I can understand that in the AU-919 this criterion is not met. I do not know the reason for this but I speculate that electrolytic capacitors can exlplode if their voltage rating is exceeded, so it may have been a safety margin. Enjoy your adventure, and please let us know your opinion on these units and how they perform.
Cheers!!

 

[GrisPato]

Old school advice for using electrolytic capacitors recommended a 50% voltage margin

Ok, that's a lot I think. I love studying schematics ( like I loved to study maps when I was a kid ) and see mostly around 15% used, almost never more.
The ripple test I am going to do to. How would you do this under load. With dummy's, or with speakers? Mhh, a lot of fun ahead

 

[JoseHH]

Ok, that's a lot I think. I love studying schematics ( like I loved to study maps when I was a kid ) and see mostly around 15% used, almost never more.
The ripple test I am going to do to. How would you do this under load. With dummy's, or with speakers? Mhh, a lot of fun ahead

About the ripple I would do the test that John suggested you. I just did that test on my amplifier. I would think that up to 1% or so of the voltage may be acceptable for the unloaded unregulated power supply. I experimented with other power supplies at work (not with the one in the amplifier) and for a 45V power supply I put a 500 ohms 10W resistor right at the output to load it with 100mA . Nothing else connected to the power supply. The ripple clearly increased in that power supply. Now, I do not know what would be a good current load to test the capacitors in the power supply in the AU-d907. I checked out these videos to learn how to test the ripple; xratonyb and EEVblog but you may be able to find other sources.

 

[GrisPato]

Just didn't have the patience.


Doesn't look to bad . . .

 

[GrisPato]

And this:



Six "black flag" styrols on the backside of the flat amp board? Is this standard?

 

[GrisPato]

When you start it is difficult to stop. . .



Not the trend I was hoping for but ESR looks promising.



Cleaned them a little better



A little better view of the nma/nmc 1012's



I read something about using a wooden block. The power amplifier is out.

 

[GrisPato]

So now I could measure the driver/ pre-driver transistors. 2sb537 is shorted to.



Looks a little weird to.
The rest is ok for now.
Can I use these ones for replacement?

 

[JoseHH]

The capacitors appear to be ok.
The drivers are bad?. While this pair of transistors is suitable to be used as drivers in this amplifier, you need to make sure that they are not a counterfeit or copy with sub-standard quality.
The 2SC4793 seems suspicious to me, I think it does not look like an authentic Toshiba part. The 2SA1837 looks fine to me. Compare the legs and the number 3. The ones I have all look consistent with the 2SA1837 that you show. Please note that I am not saying that your 2SC4793 indeed is fake. All i am saying is that it looks odd.
Alternative drivers are:
2SC2344/2SA1011 (SANYO still available)
2SC4883A/2SA1859A (SANKEN, still available)
2SA1930/2SC5171 (Toshiba still available)
Old parts hard to find but will work
2SA968/2SC2238 (Toshiba)
2SB536/2SD282 (NEC)

 

[GrisPato]

The capacitors appear to be ok.
The drivers are bad?. While this pair of transistors is suitable to be used as drivers in this amplifier, you need to make sure that they are not a counterfeit or copy with sub-standard quality.
The 2SC4793 seems suspicious to me, I think it does not look like an authentic Toshiba part. The 2SA1837 looks fine to me. Compare the legs and the number 3. The ones I have all look consistent with the 2SA1837 that you show. Please note that I am not saying that your 2SC4793 indeed is fake. All i am saying is that it looks odd.
Alternative drivers are:
2SC2344/2SA1011 (SANYO still available)
2SC4883A/2SA1859A (SANKEN, still available)
2SA1930/2SC5171 (Toshiba still available)
Old parts hard to find but will work
2SA968/2SC2238 (Toshiba)
2SB536/2SD282 (NEC)

Yes, so true. I noticed the difference to and I am also not sure they are original ones. They measure good in the, between 140-160 but I realize that's no sure thing.
I like the Sanyo you're suggesting. Especially because it still has the metal tab. Considering, can never hurt to have some stock from high quality transistors.
Thanks for the information.

 

[GrisPato]

Tonight I wanted to get the nma / nmc 1012's out to prepare for putting the temperary replacements in.



Gave me the opportunity to measure them out of circuit and as I was silently hoping just one pair is shorted.
So that's a lot better percentage to work with.

 

[GrisPato]

To get to them I removed the finned part of the heatsink. Was surprised to find almost no thermal grease in between. What was there is transparent opposed to the white compound under the bjt's. I understand there is transparent silicon paste. Was there a special reason to use here a different one? And it was applied really scarcely.




Most of the mica's were burned in and would split removing them ( not breaking ). Well, some cleaning to do now.

 

[MIJ_Pioneer]

Awesome job so far!

I like how you checked the ESR on those filters, not an easy job to replace - I'm sure mine were doing ok as well. To be honest, I didn't really notice a difference with new replacements, certainly nothing night and day like I have experienced with other gear.

Do you plan on / have you looked into replacing the Thyristor for the protection board? I've read that this can be troublesome and often thought of replacing mine or at least ordering a spare. I am not sure what the modern replacement is.

The black flags on the back of the flat amp is documented, sometimes they used those other times they used the standard orange/tan ceramic caps as well.

I've ran into several (most?) Sansuis that use clear silicone (some type of grease) as a thermal coupler between heatsinks/micas. I'm not sure what the story is behind it but everytime I see it the grease still looks like it was doing the job - can't same the same for 90% of the white stuff. I don't think its worth reading into and they standard white compound we use is up for the job.

Dont forget to check those emitter resistors connected to the output terminals.

 

[GrisPato]

Do you plan on / have you looked into replacing the Thyristor for the protection board? I've read that this can be troublesome and often thought of replacing mine or at least ordering a spare. I am not sure what the modern replacement is.

Didn't consider that yet. Could be good to have some. I understand a possible replacement is mcr100. Found them in e few sec but they have mcr100-6 and -8. So have to find out which one I need.
About the thermal grease. I was more surprised that it didn't cover more 30 % of the contact surface. I didn't clean the finned part, didn't need to was practically no contact.
Emitter resistors all measured good. MV103's and SV03's also. So for now I am 1 pair of nma/nmc 1012 and a 2sb537 down. Could have been worse.