STK-0050 replacement for SX-780 and others

[dr*audio]

He is asking if he can use the new discrete module in a Onkyo A-08 design. based on the STK-1040 design, I said I think so.
The discrete module bias generator is setup for a Re of 0.22 ohms, do not alter the design with a .47 Ohm 5W emitter resistor, it uses 200W bjts so it has plenty more power than the originals.

Ah, I thought he was asking about NOS Modules. Yeah, .22 Ohm is good for yours.

 

[zebulon1]

I have some misfit heatsinks, nine in all that have the 3mm threaded mounting holes enlarged.
I'll send them out for shipping only. Hopefully a guy could use them.
The next size up is a 4-40 standard but I think that might be smaller than the current holes.

 

[merlynski]

Usually the burrs are around the holes that were tapped. Inserting a screw can force burrs out of the threaded holes as you install. Before I use the plates I run screws through them to do just that and then do a final polish to the plate concentrating on honing around the screw threads to de-burr.
Just get new silpads and go at it again. One way or the other you need an insulator, you could use a MJF1503x which is an isolated tab device, two less places to have an issue with a short.

A 120 degree countersink bit will clean the burrs off and bevel the opening very nicely, then run the tap through again, then a screw, then a final polish. It is a bit more work but will eliminate the the burr, and greatly reduce the likely-hood of a some-what over-torqued screw raising a burr. If you countersink the backside of the metal transistor tab you get rid of any sharp burr that might be there as well. Yeah, I know, OCD perfectionism .

 

[Young46]

Hi all. I just finished a successful install of the STK0050 on my SX-780. I made a critical error on the first try and with the guidance from merlynski we fixed that.

I can tell you that from my perspective anyone who has the opportunity to do the upgrade should. It is a quality product all the way around. My 780 sounds great!

 

[BigWhale]

Greetings, is there a way to test this module without connecting it to the main unit? Here's my story:

I got a Sanyo DCA-411, tried it with some speakers and it was working for a while then the left channel died. I thought it was something with the cables and after a few more tries everything died and there was no output on the speakers. Long story short, 411 has two STK-0050 modules, the left one was blown and pins 2 and 3 were shorted and this caused both fuses on the transformer output to blow. Found this page and decided to try and replace the STK's. I had some problems with the gerber files and I couldn't import them in PCBWay or JLCPCB, so I just made a different board and I trimmed the board size to fit 411's heatsinks.

Here's the project on EasyEDA. I used the components that were listed on Mouser. I did my ordering at Farnell, because shipping is way cheaper. After I got PCBs from JLCPCB I build everything and checked for shorts on the collectors and did some sanity checking and everything seemed fine.

When I connected everything together and turned on the amp power meters went crazy. Jumping from 40% to 80%. If I connected 50W 8Ohm resistors to the speaker output power levels just spiked to 100% and then dropped and kept doing this indefinitely. After some debugging I figured out (smelled it actually) that R2 on the left channel is sizzling hot. Right one was ok. This was with volume turned way down and treble/bass set to zero.

My suspicion is that something else broke on the main amp board and that was the real cause for the original STK to fail. So, I wanted to check if the right channel was working as it should.

It didn't. I connected signal generator to the input of the amp. Just a modest 500Hz, 0.7V sine on the AUX input. Then I displayed input and the speaker output on the scope. Input signal was a clear sine and the output was a very noisy and slightly distorted sine. With no input signal, speaker output was also quite noisy with levels jumping up and down for a few milivolts.

I connected a small 8Ohm speaker to the output and the noise was very audible even at 0 volume. I was able to hear the test signal, but it was very noisy.

So, I need to figure out where's this noise coming from and I would like to rule out this STK replacement. If I connect pins 2 in 9 to the original power supply and pins 1 and 10 to the signal generator. Then pins 3 and 8 through a pair od 0.47Ohm 3W (this is what is used in 411) and connect this output to the scope (or a speaker) should I see and hear the amplified signal? And if everything checks out, I then can start digging around the amp board in the 411.

 

[rcs16]

Hello,

Welcome aboard.
I have a test jig to test out the ones I make, it is essentially a input stage, a voltage amplifier stage
There is not much to a STK-0050, all you really need is a DMM and measure ohms & in diode mode test the semiconductor junctions. Also test the resistors.
The only difference from pcbway and jlcpcb is the file naming. I use jlcpcb all the time now so the drill file needs to be re-named for a file extension of .xln.
I have had boards done at jlcpcb with liitle change to the original files. All I did was reverse the pot so CW increases bias.
I think what you need to do is what Merlin came up with, use the 1K resistors to replace the OPS and test your amp consisting of the input stage and voltage amplifier stage.
I'd have to take a look at the Sanyo DCA-411. Okay it is almost identical to the sx-780, except it has the over-current detectors, which short out the bias generator under a fault condition. It also has the tone control in the feedback loop which is a hassle when troubleshooting. Might want to isolate the tone ckt and wire up like the 780 for troubleshooting.
Take term 17 and wire to 7, lift R23, where it connects to Q15,17 and ground it to close the FB loop.
You can use Merlins' method and lift D13,15 as a start to isolate the over-current detection circuits as they might be causing your issue.

Good Luck
Rick

 

[merlynski]

Welcome to AK!!

So, I need to figure out where's this noise coming from and I would like to rule out this STK replacement. If I connect pins 2 in 9 to the original power supply and pins 1 and 10 to the signal generator. Then pins 3 and 8 through a pair od 0.47Ohm 3W (this is what is used in 411) and connect this output to the scope (or a speaker) should I see and hear the amplified signal? And if everything checks out, I then can start digging around the amp board in the 411.

You would have to build a test jig for the STK to test it, BUT you can troubleshoot your power amp section without the STK (or new module) installed.
Remove the STK module and connect 1kΩ resistors as shown in this modified schematic:

If there are failures in the amp they will show up as imbalance in the circuits. With the resistors in place the dc feedback path is closed and the amps dc balance will work. The 'STK' pin connections 0 and 1 will be about equal in voltage (4 to 10vdc, not critical), but opposite in polarity (0 will be positive and 1 will be negative). The dc offset at PCB pins 17 and 18 should be close to 0vdc. Of course since the STK is removed there is no current gain it will not play music. You can test the power supplies and other components under power for troubleshooting purposes. Once the amp balances properly put the STKs in.

EDIT: rcs16 beat me to it

 

[rcs16]

I did an edit, have to isolate the tone control in the FB loop too.

 

[merlynski]

I did an edit, have to isolate the tone control in the FB loop too.

@rcs16 How does the tone control affect the dc feedback loop? It is an AC connection through 10µf caps at the tone pcb input and output?
Isn't the dc feedback through R23 (56kΩ) and R13 (560Ω)? Won't connecting pin 17 to pin 7 create an additional dc feedback path? Lifting one end of R25 (5.6kΩ) would eliminate the tone control ac feedback without moving wires . . .?

EDIT: more

 

[BigWhale]

Thanks for the tips, I'll be trying these things tomorrow. Just a question about the 1k resistor, what power rating? I think I have only 1/4W on hand (that's what you end up with when working mainly with micro controllers ), so I'll have to get some with a higher power rating.

I'll measure the voltages on those pins on the left channel.

What can I do about the noise problem on the right? I was thinking about getting rid of the pre-amp and tone control but I wasn't sure how to go about it. Now I know what to do with the tone control, thanks. I'm still not sure if I can easily get the preamp out of the equation. I'll see if this helps. Thanks!

 

[merlynski]

Thanks for the tips, I'll be trying these things tomorrow. Just a question about the 1k resistor, what power rating? I think I have only 1/4W on hand (that's what you end up with when working mainly with micro controllers ), so I'll have to get some with a higher power rating.

I'll measure the voltages on those pins on the left channel.

What can I do about the noise problem on the right? I was thinking about getting rid of the pre-amp and tone control but I wasn't sure how to go about it. Now I know what to do with the tone control, thanks. I'm still not sure if I can easily get the preamp out of the equation. I'll see if this helps. Thanks!

1/4W should be fine, that is what others have used, it is not really a high current connection. If the dc voltages between pins 0 and 1 are too far separated you can go smaller, one person used 560Ω without a problem.
You can isolate the preamp from the power amp by lifting one end of R01 or R02, however, if you turn the volume to minimum it should (unless its broken) reduce the input to the power amp to nothing.

 

[rcs16]

Sure R23 is the DC FB R.
The more I look at the circuit, looks weird, something wrong, they show Q15,17 emitters at +43.5V which is connected to the o/p at term 17?
Lifting R25 opens the AC FB loop with the tone ckt.

 

[merlynski]

Sure R23 is the DC FB R.
The more I look at the circuit, looks weird, something wrong, they show Q15,17 emitters at +43.5V which is connected to the o/p at term 17?
Lifting R25 opens the AC FB loop with the tone ckt.

Good catch!
That has got to be mis-labeled, there is no way to get 43.7vdc across a forward biased B-E junction of Q15.

 

[Chainer211]

Well, that has to be cleared first. take the module out and figure out using the ohmmeter, what is shorted and replace/correct it.
Is it possible you have assembled them incorrectly, what exactly I do not know for sure.
here are a few possibilities.
Incorrect mounting of the TO-220 devices MJE1503x. shoulder washer correct orientation, incorrect use the thermal insulator, over-torqued the screws causing a short. Burrs on the heatsink protruding thru the thermal ins and shorting collector tab to heatsink.

Its been several months and I've had the time to make all the adjustments. Unit powers up and relay clicks when on the dbt. At full power relay clicks, but the right module gets very hot. Hot enough that I thought it might catch on fire. The left module stays cool throughout. What am I doing wrong here?

 

[merlynski]

Its been several months and I've had the time to make all the adjustments. Unit powers up and relay clicks when on the dbt. At full power relay clicks, but the right module gets very hot. Hot enough that I thought it might catch on fire. The left module stays cool throughout. What am I doing wrong here?

What are the dc voltages on pin 0 and 1 of the hot module? Go back on dbt.

 

[BigWhale]

Thanks for all the tips!

1/4W should be fine, that is what others have used, it is not really a high current connection. If the dc voltages between pins 0 and 1 are too far separated you can go smaller, one person used 560Ω without a problem.
You can isolate the preamp from the power amp by lifting one end of R01 or R02, however, if you turn the volume to minimum it should (unless its broken) reduce the input to the power amp to nothing.

I'll try with 1/4W resistor and see what happens. Thanks.

About R01 and R02, I could lift the resistor and connect the signal generator probe to it and ground from the probe to the terminals 2 or 5, right? This way it will completely bypass the preamp?

Sure R23 is the DC FB R.
The more I look at the circuit, looks weird, something wrong, they show Q15,17 emitters at +43.5V which is connected to the o/p at term 17?
Lifting R25 opens the AC FB loop with the tone ckt.

Yeah, I lifted R23, grounded it, and then connected the terminals 7 and 17 but it really didn't do anything. Well, it disabled the output completely. I'll try with lifting the R25 instead.

 

[BigWhale]

Another update, I tried and removed new modules and replaced one with the STK-0050 that is still working, the noise was still present, so it's not the modules, I'll have to look elsewhere. There are three things that I did after I opened the device:

- recapped all electrolytic capacitors (except one, because I ordered the wrong one), capacitors were all the same brand (Rubycon) and same values
- removed the wires from the main amp board and reconnected it
- cleaned all the pots with a Kontakt spray that is used to lube and remove corrosion from electrical contacts

So, now I'm stuck with diagnosing the left channel, which I'll do later today and the static/noise/cracking in the right channel.

 

[rcs16]

Its been several months and I've had the time to make all the adjustments. Unit powers up and relay clicks when on the dbt. At full power relay clicks, but the right module gets very hot. Hot enough that I thought it might catch on fire. The left module stays cool throughout. What am I doing wrong here?

Sounds like the bias is too high. Possible cause are the bias generator components incorrect, too much current through the bias generator will cause this too. Need to measure some voltages and check against the good working channel and the schematics that were posted. You can mark up a schematic with what you are measuring. You can measure the VAS current across one of the resistors in the path to calculate what you are getting.

 

[rcs16]

Hi BigWhale,

Good that it isn't the module, check one channel against the other, just using your ohmmeter with power off, it may uncover the fault location.

 

[Chainer211]

What are the dc voltages on pin 0 and 1 of the hot module? Go back on dbt.

Pin 1: -1.03
Pin 3: 0.08

This seems wrong to me as the opposite module reads 12 and 13v respectively.