STK-0050 replacement for SX-780 and others

[Tala Mink]

I used a board based on this threads 0050 replacement in a Yamaha 1040 about a month ago.

I was a little concerned about stability because the Yamahas output chip is a triple and the 0050 replacement is a standard Darlington but the amp biased up fine and was stable to clipping.

The big issue is the mounting new output modules, most STK's use M3 screws whereas the Yamaha 1040 uses M4, Also I think that the mounting holes were further apart than the 52mm of the STK 0050.

 

[Mike Mag]

I am trying to order parts to build some STK-0050-A1 boards and Mouser is out of the 512-BD13910STU transistor. it seems like this is the recommended replacement, is that accurate ?

https://www.mouser.com/ProductDetail/512-BD13916STU

 

[saabracer23]

I am trying to order parts to build some STK-0050-A1 boards and Mouser is out of the 512-BD13910STU transistor. it seems like this is the recommended replacement, is that accurate ?

https://www.mouser.com/ProductDetail/512-BD13916STU

Yes, those would be perfectly fine to grab. If there’s anything else on the list, you can’t find. Let me know., I have everything needed for the boards

Dan

 

[saabracer23]

I used a board based on this threads 0050 replacement in a Yamaha 1040 about a month ago.

I was a little concerned about stability because the Yamahas output chip is a triple and the 0050 replacement is a standard Darlington but the amp biased up fine and was stable to clipping.

The big issue is the mounting new output modules, most STK's use M3 screws whereas the Yamaha 1040 uses M4, Also I think that the mounting holes were further apart than the 52mm of the STK 0050.

Appreciate that, yeah, I’ve put these on a few different pieces of equipment and as long as there’s physical room on the heat sink, shouldn’t be a problem, I have no problem, drilling and tapping for new mountings. It’s very good to know that these version one boards will work in this amp, did you have to do any modifications to make it work? I’m guessing it wasn’t the exact same circuit since the originals are going to be mk II boards.

Dan

 

[ConchBoy66]

Recently my sx-3800 that I rebuilt had an issue and I pulled it out of rotation for repair. I replaced it with my sx-780 that I had redone some time ago with these modules and I have to say it sounds marvelous. Almost a shame to pull it out and put the 3800 back in. Just throwing all this out there, these modules are great!

 

[Aldure]

0.47 or 0.5 gives better protection and helps with current sharing in the output devices if they aren't perfectly matched. The feedback reduces the output impedance so it doesn't hurt to use 0.5 Ohm resistors.

The forum format does not handle equations and structured calculations very well, so I provided my answer as a PDF.
This way the calculations for the specific 4 Ω / ±44 V case, as well as the real-world test conditions (100 W into 4 Ω with fan cooling), are shown clearly and unambiguously.

 

[chazix]

calculations

Very clear, indeed - nice presentation!

I think your calculations of the power dissipated in the emitter resistors might be incorrect, though. If I understood correctly, those calculations assumed that the full RMS current flowing in the output load is also flowing in each emitter resistor. But the current in a single emitter resistor in a class AB push-pull circuit is more like a half-wave-rectified sinusoid, no?

I'm not so sure about your calculation of damping factors, either. I don't think you can assume that the emitter resistance is simply added to what you designate as R(out,amp), when the emitter resistors are "inside" the global feedback loop.

No quibble with the finding that the higher-valued emitter resistors reduce maximum output power, but really, I don't think the reduction is enough to be perceptible to most human ears.

If I was choosing an emitter resistor value, after careful consideration of all the above, I would choose what I had on hand .

 

[Richard P]

Nice analysis, @Aldure . You mentioned use of the STD03N/STD03P in lieu of the complimentary pairs specified for these boards. Has this been done? I have quite a few STD03N/Ps with no current plans for use and would love to adapt them to this design but haven't seen it mentioned until your post.

 

[Aldure]

Nice analysis, @Aldure . You mentioned use of the STD03N/STD03P in lieu of the complimentary pairs specified for these boards. Has this been done? I have quite a few STD03N/Ps with no current plans for use and would love to adapt them to this design but haven't seen it mentioned until your post.

I don't know if you've seen this post of mine?

 

[Aldure]

Very clear, indeed - nice presentation!

I think your calculations of the power dissipated in the emitter resistors might be incorrect, though. If I understood correctly, those calculations assumed that the full RMS current flowing in the output load is also flowing in each emitter resistor. But the current in a single emitter resistor in a class AB push-pull circuit is more like a half-wave-rectified sinusoid, no?

I'm not so sure about your calculation of damping factors, either. I don't think you can assume that the emitter resistance is simply added to what you designate as R(out,amp), when the emitter resistors are "inside" the global feedback loop.

No quibble with the finding that the higher-valued emitter resistors reduce maximum output power, but really, I don't think the reduction is enough to be perceptible to most human ears.

If I was choosing an emitter resistor value, after careful consideration of all the above, I would choose what I had on hand .

Given the original Pioneer implementation and the SANYO datasheet data, I’m comfortable concluding that ~0.22 Ω is the intended design point for this class of single-pair STK0050-style amplifiers.

 

[Richard P]

I don't know if you've seen this post of mine?

Oof. Sorry. I'm not sure how I missed your earlier response. Especially since it was a reply to my own question. Maybe the STK-1050 reference threw me off. I'll take a closer look at that. Thanks!

 

[sregor]

The forum format does not handle equations and structured calculations very well, so I provided my answer as a PDF.
This way the calculations for the specific 4 Ω / ±44 V case, as well as the real-world test conditions (100 W into 4 Ω with fan cooling), are shown clearly and unambiguously.

Nice, except for damping factor analysis. The emitter resistors are inside the global feedback loop, so don't get added on to the output impedance.

 

[Aldure]

Oof. Sorry. I'm not sure how I missed your earlier response. Especially since it was a reply to my own question. Maybe the STK-1050 reference threw me off. I'll take a closer look at that. Thanks!

The STK1050 is not only compatible with STK0050-styleapplications, it was explicitly designed for backwardcompatibility.
While STK1050 includes integratedemitter resistors, these are not forced into theoutput path:
the module provides separate pinsfor emitter-resistor-free outputs, which are taken directlyfrom the emitters of the output transistors, bypassing theinternal resistors.


This makes STK1050 electrically and pinout-compatiblewith earlier STK0050-based designs where the emitter resistors wereimplemented externally (e.g. Pioneer SX-780 with0.22 Ω).


In other words, STK1050 does not change the output topology — itsimply adds the option of internal emitterresistors, while preserving full backward compatibility with earlierSTK modules.


STK1050 remains compatible with STK0050because it provides direct emitter output pins, allowing the internalemitter resistors to be bypassed and preserving backwardcompatibility.

 

[Aldure]

Nice, except for damping factor analysis. The emitter resistors are inside the global feedback loop, so don't get added on to the output impedance.

Looking at the Pioneer SX-780 schematic, the global feedback is indeed taken from the speaker output, i.e. after the external emitter resistors, so at DC and small signal their contribution is largely inside the loop.


However, the loop gain is finite and drops at high frequency, high current and near clipping. In those regimes the emitter resistors still cause real voltage drop, power dissipation and output swing reduction that feedback cannot cancel.


This is exactly why Pioneer chose 0.22 Ω external emitter resistors with STK-0050 in the SX-780, rather than 0.47–0.5 Ω, despite the feedback being taken after them.


So while a simple “add Re to Rout” model is not correct, it’s equally incorrect to say Re can be ignored in this specific STK-0050 / SX-780 application.

 

[Aldure]

I understand the general recommendation, but since thetopic is explicitly titled “STK-0050 replacement for SX-780and others”, I think it’s reasonable to expect commentsto stay within the context of that specific STK application.


We’re not discussing a generic BJT output stage here, but adirect replacement for STK-0050 in a specificamplifier, where the original design choices are welldocumented.


In this context, recommendations like 0.47–0.5 Ω— while valid as general rules of thumb — are not directlyapplicable, because both Pioneer and SANYO optimized thisapplication around ~0.22 Ω.


So this isn’t a purely theoretical discussion, but rather amatter of reproducing the original, proven topology.

What does the feedback “see”?​

✔ DC and small-signal AC error
✔ Static output deviation
✔At light load, most of the effect of ReR_eRe is largely “hidden”by feedback

What can it NOT eliminate?​

✘ The real voltage drop across ReR_eRe at high outputcurrent
✘ Reduction of available output swing, especially into 4Ω
✘ Power dissipation and loss in ReR_eRe
✘ Loop-gainreduction at large signal, high frequency, and near clipping


This is why ReR_eRe cannot simply be ignored,even when the global feedback is taken from the speaker output.

FAQ: Does the emitter resistor affect damping factor ifit’s inside the global feedback loop?​

Short answer:
At small signal and lowfrequency, much less. Under real 4 Ω / high-current conditions, yes.


Long answer:
In amplifiers like the PioneerSX-780 with STK-0050, the global feedback is indeed takenfrom the speaker output, i.e. after the emitter resistors.
Thismeans that at DC and small-signal low-frequency operation,the loop gain can largely suppress the apparent contribution of theemitter resistors to measured output impedance.


However:

• Loop gain is finite, not infinite
  
  
• Loop gain drops with frequency
  
  
• Loop gain drops under high output current and near clipping

In those real operating conditions (especially 4 Ωloads), the emitter resistors still:

• cause real voltage drop,
  
  
• reduce available output swing,
  
  
• dissipate significant power,
  
  
• and therefore affect practical damping and control.

This is exactly why Pioneer chose 0.22 Ω external emitterresistors with STK-0050 in the SX-780, rather than 0.47–0.5Ω — despite the feedback being taken after them.


Conclusion:
It is incorrect to simply addRe to Rout,
but it is equally incorrect to ignore Re inreal-world, high-current operation.

 

[Richard P]

The STK1050 is not only compatible with STK0050-styleapplications, it was explicitly designed for backwardcompatibility.

Good to know. There appears to be a seller on eBay (Semitronus) who has NOS STK1050s at a pretty reasonable price. I may pick up a few to try on a future SX-780 restoration. I'll still explore using the STD03N/P parts as well. Thanks for the great info.

 

[painttoad]

There appears to be a seller on eBay (Semitronus) who has NOS STK1050s at a pretty reasonable price.

i picked up a few legit stks from that seller. always the oddball devices. never the common ones

for giggles i bought some stk-0050 once and all i did was weight them and i sent them right back. they are, or at least WERE very good on accepting returns.

 

[dr*audio]

i picked up a few legit stks from that seller. always the oddball devices. never the common ones

for giggles i bought some stk-0050 once and all i did was weight them and i sent them right back. they are, or at least WERE very good on accepting returns.

Liberty Electronics on E-bay has good new STKs. Apparently they found some factory in China that makes quality replacements. I have tested them at full power and they are fine. I've been using them for a few years and had no failures. If you do receive a dud, they stand behind what they sell. They also have good optical pickups.

 

[sregor]

Looking at the Pioneer SX-780 schematic, the global feedback is indeed taken from the speaker output, i.e. after the external emitter resistors, so at DC and small signal their contribution is largely inside the loop.


However, the loop gain is finite and drops at high frequency, high current and near clipping. In those regimes the emitter resistors still cause real voltage drop, power dissipation and output swing reduction that feedback cannot cancel.


This is exactly why Pioneer chose 0.22 Ω external emitter resistors with STK-0050 in the SX-780, rather than 0.47–0.5 Ω, despite the feedback being taken after them.


So while a simple “add Re to Rout” model is not correct, it’s equally incorrect to say Re can be ignored in this specific STK-0050 / SX-780 application.

I don't have exact numbers but the Z out of the amp is considerably higher than .5 ohms, and goes up with frequency.

 

[Aldure]

I don't have exact numbers but the Z out of the amp is considerably higher than .5 ohms, and goes up with frequency.

I agree that in a voltage-feedback amplifier the emitter resistors are largely inside the global feedback loop, so it would be incorrect to simply add ReR_eRe 1:1 to the small-signal output impedance.


That said, it’s also not accurate to say they can be ignored altogether. Loop gain is finite, and it drops with frequency, output current, and near clipping. In those real operating conditions — especially with 4 Ω loads — the emitter resistors still cause real voltage drop, power dissipation, and output swing reduction that feedback cannot eliminate.


So a simple “add ReR_eRe” model is indeed wrong, but the practical impact of ReR_eRe on damping and control under high-current conditions is also very real, which is why designers like Pioneer still optimized around ~0.22 Ω rather than 0.47–0.5 Ω, even with feedback taken after the resistors.