Sony-STR3800L. Output transistors.

[samjmann]

The output transistors on this amp have failed on one channel. They are:

2SD531. NPN. 50-60V C-E. 5A Max collector current. (TO-220C)

2SB521. PNP. Similar values to the above.

What would make a good substitute? I can find datasheets on these, but no actual parts.

The drivers have also failed, these are available.

Some have suggested a TIP41/42. The only thing that bothers me is these have a Transistion Frequency of 35Mhz. The 2SD531 is 8Mhz...

I don't want a short lived high power oscillator!

Thanks for any help. SJM.

[orbanp1]

Hi SJM,

Chances are the TIP transistors would be OK.
Doing a quick simulation the phase margin with the TIP transistors would be about 30 degrees.
(I tried to use transistors with similar parameters in the other positions.)
30 degrees phase margin is still stable, but it would not hurt to have a higher margin.
Adding a 10pF high frequency compensating capacitor parallel with the R708 47k feedback resistor the phase margin is increased to about 70 degrees.
I would try the TIP transistors with the added compensating cap.

Regards, Peter

[Diabolical Artificer]

This - https://alltransistors.com/crsearch....=45&caps=TO220 gives the MJE15032/3 as replacements and which seem to be available - https://cpc.farnell.com/on-semicondu...220/dp/SC06827

Andy.

[orbanp1]

Double checked the spice model of the TIP41, it has an Ft of 4MHz to 6MHz, depending on the Vce voltage.
Checked the Ft of the MJE15032 model too, it has Ft around 30MHz.
Running the simulation with the MJE transistors the phase margin of the amp is only 16 degrees, which is clearly not enough!
Applying the 10pF compensating cap the phase margin increases to 73 degrees with the MJE transistors too.
So do use the compensating cap if you use better transistors!
Also, I would make both channels identical.

Regards, Peter

[samjmann]

Thanks so much for such informative answers. I will indeed apply the 10pf cap on the feedback line to both channels. The other channel is running but the stability may well be compromised
as it also had had substitute transistors fitted!

Thanks again Peter and Andy.

SJM.

[Radio Wrangler]

When you compare Ft and Hfe values of different devices, be suspicious, and look carefully at the conditions the specs apply to. Sometimes you're lucky and see that devices are more similar than first meets the eye, sometimes you lose.

It's easiest to compare if you can find plots of Hfe and Ft versus collector current.

Spot values depend somewhat on which spot someone picks. They usually go for the best looking number, but best for whatever target market they have in mind.

David

[samjmann]

Finally got round to fitting the new MJE ouput transistors. The defective channel ran up first go no problem. I'd set the quiescient current pot to minimum current before switching on. I've also added the 10pf cap across the feedback resistor too.

The service manual quotes a set up current of 50mv across 0.94ohms. This equates to 53mA, this seems very high to me. I would have thought that 10-15mA would have been Ok for a small amp like this.

Thanks again to all who have replied. Christmas, long hours at work and and the flu got in the way of a quick job!!

SJM

[vidjoman]

Originally Posted by samjmann

The service manual quotes a set up current of 50mv across 0.94ohms

Are you sure it's not 5.0 mv. Many Japanese manuals write it that way and often the dot is only just visible.

[samjmann]

Thanks vidjoman. Around 5mA seems a reasonable current setting.
I've got a manual for a Sansui AU-919 which is a huge 100w twin transformer amp, this only sits at 30mA quiescent, so this 25W Sony must be way less than this.

Whilst I was working on it, a small trimmer cap which is on top of the tuning gang just fell apart. The FM oscillator is now way off but the calibration was nearly right before. It has '20' stamped on it so I assuming that's 20pF. I've ordered that and some near values and try to set it. I'm not touching anything else in the RF stages, it's easy to make things worse!

SJM.

[samjmann]

Update after fitting! After replacement of both output transistors all was well. Well at least for a while...

The set up went ok, both channels had much the same quiescent conditions. However a week later after switching on, the same two transistors went short again. Clearly something was wrong. I'd checked the bias diode D701/D751 and it read ok. I was having to now put the bias pot at near short circuit to stop the 2nd replacement set of output transistors failing. The SV04 bias diode now reads O/C but only when the leads are moved. Its mounted on a tag that's attached to the top output transistor.

The SV04 is a 4 diode package, the good one reads 2.3v on my meter
Does anyone know of a suitable replacement?

I could make up substitute with a 4 diode string. I'm not to sure what the current would be in the diode string, hence what would make a good substitute.

Regards SJM

[orbanp1]

Hi SJM,

The current through those diodes is close to the current of the VAS stage, in your case it is about 3.7mA according to simulations.
Small signal Si diodes, like 1N914 diodes, are fine there.

Regards, Peter

[G6ONEDave]

According to my Jaeger book the SV04 is a described as 'Si-Varistor' 'VDR/Varistor'. My ECA book shows it as being a 2.7v Zener, no current rating given. Might be worth checking carefully before replacement.

Dave