Repair of Audio Amp using Sanken SAP15N, SAP15P output transistors - FIXED

[EHT]

All,

i'm just adding this post in case it helps anyone with the same repair. I had an audio amp which blew the power transistors on one channel, taking with it a large number of other components. This was my fault (see my other thread) but replacing these transistors was not quite as straightforward as I hoped hence letting you know. I could identify the other damaged parts using DMM diode-test and resistance measurements. These were mostly SMD transistors & diodes close the output stage.

The Sanken SAP15N/P transistors are Darlingtons which include an emitter resistor (See attached).  Apparently failure of these embedded resistors was found to be common, and these transistors have now been discontinued due to this. Mine actually caught fire! (see pic)

Sanken make updated Darlingtons with no emitter resistors - STD03N, STD03P. It is possible to find people selling the obsolete SAP15s parts on fleabay but using the new parts is basically an upgrade. Still, i didn't replace those in the working channel as the transistors are about £5 each.

The SAP15s have 5 pins, one labelled "S" (sense) is actually the emitter and "E" includes the resistor. The STD03s have a blank stub where "S" was and "E" is the emitter with no resistor. Hence to fit the replacement you need to bend the E pin to where S pin was on the PCB, then attach an external resistor in place of the original "S" and "E" pins. See drawing. Pic shows the STD03s in place and the resistors panel mounted. The are 0.22 Ohm. 3W would probably have been sufficient in this amp but i decided to make it highly rated and used 5W bolted to the chassis, which is quite sturdy steel in this amp.

http://www.semicon.sanken-ele.co.jp/sk_content/std03p_ds_en.pdf
http://www.semicon.sanken-ele.co.jp/sk_content/std03n_ds_en.pdf

[Dazzamatazz]

EHT,

Many thanks for posting your solution! I have exactly the same problem with two Arcam P90 amps. One went up in smoke, so I bought another second-hand to replace it while I worked out the best way to get it repaired. Then the second amp went up in smoke with exactly the same fault. Neither amp had a short, as per your situation, and they were operating under normal minimal load conditions.

I spoke to Arcam who were no help at all, they just told me to speak to a dealer. I pointed out that this must be a known issue as it was an identical fault, but they just kept silent. The dealer couldn't help either and said I should buy a new unit.

I looked into third-party repairers, but the cost of shipping and repairs is very expensive.

A bit more research led me to this discussion, regarding a fella with an Arcam A500 with a charred Sanken SAP15:
http://www.diyaudio.com/forums/solid-state/2658-sanken-sap15-transistors-5.html

From that article, I found out that the transistor has an inherent reliability issue with its internal resistor. I'm very disappointed in Arcam. They must know about the issue. This is the age of the internet, so issues are going to become well known. Why not just admit there is a problem and suggest a fix? After all, it's not their fault, it's a third-party component failing.

Luckily, I was in the room when both transistors went up so I could switch them off quickly. Hopefully, the damage is limited to the single resistor, but I'll need to do a bit more work to check the other components. I'm no electronics genius by any means, but I'm fairly handy with electrical items and have a DMM.

My biggest quandary is whether to replace all the SAP15s with the new STD03P 4-pin replacements as per your fix, or just replace the single transistor: maybe the others will eventually fail. It's interesting that the RHS transistor has blown in your example and also in both of mine. I found the replacement transistor at Nikko Electronics for £6.05:
http://nikkoe.com/std03p-sanken-part-replacing-sap15p.html

I recently invested in an electric screwdriver (Panasonic EY7410LA2S31) which is fantastic and makes light work of getting all the screws out. The kids were fascinated by it, so I've enlisted them for disassembly work .  I should've bought one years ago! But even so, I don't want to have to fix the same amp multiple times if the other transistors fail if I can avoid it.

One thing that you didn't mention in your post is whether you had to re-calibrate the bias current for the new transistors? I can see the variable pot labelled as such. Also, would be very useful to know which other components failed so that I have a good idea where to focus if you have a minute...

I've attached pictures of both amps with the burnt-out resistors, plus a pic of the good side with the 2.5 Ohm resistor intact. I'm trying to find out what wattage the replacement should be: any suggestions gratefully received.

Thanks again,
Darren

[EHT]

Hi Darren,

hold your horses a minute. From your description, it is NOT the same as the damage I had to repair, so please don't rush to make the same changes without diagnosing it. Mine was a P85. I can't find the Service Manual for the P90, but from your pics it looks like it shares the same component numbers and I have the PDF for the P85.

So the resistor which has burnt in yours is R350 in the Zobel output filter (see pic). Probably that is C319 which is also damaged. The resistor which I was referring to is actually inside the Sanken transistors.  Yours don't show any physical damage, whereas mine had flames which reached a few inches out of the vent!

So there is good news and bad. Good news is it looks like you have a lot less damage to repair. Bad news is it is not the same as mine so you have to diagnose it!

Did you do anything "odd" to the speaker output? Like tried any kind of "bi-amping" or connected the ground to earth, or connected it to a subwoofer input etc etc. Alternatively maybe there is a component damaged in the speaker's crossover, or maybe the amp's output relay?

Here's what i recommend:

• check the above
• get the service manual, use the P85 one if it can't be found
• test the semiconductors in the area using diode test function. Include testing the output transistors but note the complex internals.
   You can compare the working channel to the damaged one for all components - even if you don't really know what reading you should get (the 2 channels are conveniently labelled such that the numbers on one channel are 100 +the other)
• test R351 and any other resistors in the vicinity
• look for other burn damage & test any components damaged
• order yourself replacements for any other damaged components plus R350 (5R6 2W) and C319 (100n, i think polycarbonate film - get one that looks the same)
• you may need to replace the silicone insulator on the heatsink. Mine ripped apart when i disassembled. This needs to be an insulator not just grease. You also need the thermistor refixed against the heatsink. This wasn't as easy to find as i expected. Can lookup if you need. Got it from Digikey as they had the STD03s in stock as well.
• after assembly test one channel against the other. Do with very low input level and/or with the speaker relay off. Do not attempt this if you are not experienced. You have live mains in there and 100V in the amp.

If you do have to replace the output transistors then do all of them in one channel. Looks like there are 2 like the P85, so you get a pair of new ones. I can lookup which resistor i used if you end up doing it. I used a panel mounted one. Yes, it does need the bias recalibrating after replacing the transistors (you need an accurate DMM for that as it is only a few mv).

BTW, I don't blame Arcam for this, although not great to just tell you buy a new one! It was totally my fault that mine got destroyed (i shorted it whilst the output protection was known to be inoperative  )

All the above it not necessarily easy.. depends how experienced you are. Also you may need to remove and replace some SMDs.

Best of luck.

[Jim Chinnery]

Thank you, a Laney amp blew for no reason, now I know why. I was about to order replacement transistors now I'll follow your advice.

[jimmyarcam]

Thanks for this, I have opened my P85 to find this issue.
The steps you mention are simple to follow.

Three questions:
Do you remember where you got the chassis mounted resistors? I cannot find a source.
Do you remember which silicone backing you got?
You mention resetting the bias. Please can you explain your procedure for this? I can see the measuring and adjustment points on the circuit diagram, and the stated mV figures for warm/cold).

Thanks

[mikerj]

The burnt out resistor in the Zobel network tends to imply the amplifier was oscillating pretty strongly (or less likely, the cap is shorted).  Hopefully not caused by special "audiophile" high capacitance speaker cables...