Sansui au 11000a - smoking resistor - please help!

[jayjay99]

Hi everybody,

This is my first post hoping to get some assistance. I managed to got hold of a beautiful sansui au 11000a which worked ok but was a little bit noisy and perhaps not as crisp as I would have expected. Based on my readings on this forum, I thought I would attempt a basic restoration, and replaced all electrolytic capacitors (with same specs), the large power filter capacitors (with United Chemi-Con 10000 uF 100 VDC), the small transistors that have a metal tab with KSC3503, the VD1212 diodes replaced with two series 1n4148 and the fuses (with same specs).
When I plugged the amp for the first time after making these changes, after about 5 seconds, the F2661 board started smoking. From visual inspection, I could see that R60 was burned. I tested some of the other resistors and realized most of the 470 ohm, 68 ohm and 1k ohm were significantly off specs so I replaced all of these resistors. I also double checked the orientation of the replacement capacitors and transistors and throroughly cleaned the botton of the board (there were a number of of small solder projections although i am not sure enough to create to short circuit). Plugged it and again R60 started smoking. Reading the forum for similar issues, I realized I needed a dim bulb tester. I made one, plugged the amp. The bulb lit up a little but the resistor still started smoking.

So I know I am a rookie, I shouldn't have embarked in something I don't fuly understand etc. But I thought I could handle it having recapped another amp and speaker. Now I am stuck with a fantastic amp that is broken and I have no possibility of having it looked at by a decent professional (I live in Morocco and can't find anyone who has any expertise of hifi). So I want to try everything I can before giving up. But I need some help please...

1. why is the resistor still burning despite using a DBT?
2. what should I look at having already checked the orientation of the replacement parts? is the issue necessarily with the 2661 board or could it be elsewhere?

Please help me. I wish I could have it repaired by a competent technician but I just don't have that option unfortunately. I am attaching a picture of the 2661 board. The service manual has a schematic on p14 and can be found at http://www.amplimos.it/images/Sansui AU-9900A, AU-11000A Service.pdf

Thank you

 

[Hyperion]

and replaced the small transistors that have a metal tab with KSC3503,

You do know that there are (at least), two distinctly different types of 'transistor with a small metal tab' as used in an AU-11000A ? - and if you have replaced ALL of those with one type - i.e. KSC3503, I am not surprised something is smoking.

1. why is the resistor still burning despite using a DBT?

I am very glad you are using a DBT - otherwise much more would have been damaged. DBT's do not prevent all components from getting hot and others from being permanently damaged in the case of mistakes. They simply limit the damage so that it should be easier to recover after a problem has been identified.

It wasn't smoking before, so I suggest you check your work starting with those transistors - which incidentally have a different pinout as compared to the originals. 2SA818, and 2SC1628 (the ones I think you mean?), are EBC (legs down, transistor facing you), KSC3503 are ECB (legs down, transistor facing you).

The above pinout issue might explain everything - if you have replaced those transistors with the two types required (NPN & PNP)

 

[jayjay99]

You do know that there are (at least), two distinctly different types of 'transistor with a small metal tab' as used in an AU-11000A ? - and if you have replaced ALL of those with one type - i.e. KSC3503, I am not surprised something is smoking.

Thank you for answering. I did pay attention to the pinout issue. But I replaced 2SA818 with KSC3503 which is not right according to this post https://audiokarma.org/forums/index.php?threads/once-more-sansui-transistor-substitutions.583009/. I will put back the original 2SA818 and report back

 

[spark1]

2SA and 2SB = PNP transistors

2SC and 2SD = NPN transistors

NPN can never substitute for PNP, and PNP can never substitute for NPN.

 

[jayjay99]

I put the 2SA818 back (double checked orientation including comparing with the picture of an original board) and replaced R60 with a new resistor. When switching on, light bulb still lighting up a bit. This time, I thought the issue was solved: no smoke, protection turned off, but after about 30 seconds I started smelling funny smells and the amp went back into protection mode. R60 is burned again.

Any ideas where should I look next?

 

[Hyperion]

Any ideas where should I look next?

Check the output transistors, especially the one nearest to R60 on the schematic, you will need to remove them to test them properly. Followed by the driver transistors, same approach - remove them to test. Draw diagrams, take pictures, but record where everything goes and make sure everything goes back in the right places.

Anything the was installed incorrectly is highly suspect - I recommend you discard them and fit new ones.

Good luck.

 

[jayjay99]

Thank you for guiding me.

I checked the output transistors: both SC1403A (tr01 and tr02) are shorted on the channel with the burnt resistor (close to 0 voltage drop from collector to emitter; 0.4 to 0.8 voltage drop from emitter to base and from collector to base). On the other channel only one of the Sc1403a (tr01) is bad (close to 0 voltage drop from collector to emitter; 1.2 from emitter to base and from collector to base).

The 2SA745A output transistors and the driver transistors (tr23, 24, 25 and 26) on f-2661 are good.

I have a few questions:
- what are the recommended replacements for the output transistors? Some older posts mention MJ21194 but it’s no longer being sold. Is MJ21196G a good alternative (it’s the only one that is recommended on mouser as a replacement) or any other suggestion? I assume I can keep the 2SA745A which tested ok?
- Could the replacement of the 2sa818 with NPNs explain why the output transistors are shorted?
- what else should i check anything else at this stage? It’s going to be quite expensive to get the parts delivered to Morocco so I’d rather make a single order if other parts will need to be replaced

thank you

 

[jayjay99]

Having done more research I found a website that has a few MJ21193G and MJ21194G in stock.

Can someone please advise if i should replace both SC1403a and sc745a even if only the former are blown? Or would the amp work fine with a combination of MJ21194 and the old sc745a?

Any other suggestions as to what i should check to avoid blowing the new transistors (We are talking €100 to replace all 8 of them). I already tested the driver transistors (tr23 to 26) and they are fine.

Thanks!

 

[Hyperion]

You should never 'mix' old and new output transistors in a channel - so you should change all transistors in the faulty channel at a minimum. However it is generally regarded as best to change even the 'good' transistors (in the other channel), for a set of new transistors to keep the characteristics of both channels the same.

Be very wary that the shortage of MJ21193/94 transistors may have increased the possibility of fakes/counterfeit devices being offered for sale - 'caveat emptor' is the watch phrase. Note that MJ21195/96 are suitable alternatives, if you can find them from preferred suppliers like Mouser, Digikey, Newark, Farnell, RS Components, etc. Preferably not from auction site vendors located in the far east.

Make sure you test the driver transistors as well. There have been cases where someone has replaced output transistors after some failing and going short circuit, not realising that some/all of the driver transistors have failed as well. Changing only the output transistors in these circumstances can cause the new set to be immediately destroyed by the faulty driver transistors, - you have been warned.

 

[jayjay99]

Thank you Hyperion. I assume you mean to test all transistors on the driver board (not just the ones that are immediately before the output transistors - tr23 to tr26 - which i already did)?

 

[Hyperion]

Thank you Hyperion. I assume you mean to test all transistors on the driver board (not just the ones that are immediately before the output transistors - tr23 to tr26 - which i already did)?

I didn't realise you had tested the driver transistors, but testing others just to be sure, especially TR13/14 isn't a bad idea.

 

[jayjay99]

Will do. I will order the output transistors and report back when i will have installed them - and hopefully not fried them

 

[Hyperion]

and hopefully not fried them

Keep it on DBT until you are sure you have control of 'DC offset' and that bias can be adjusted, and you shouldn't have problems.

Note that you can probably sell the output transistors from the good channel, which will recoup some of your costs.