STK404-140S Question

[Enzo]

Weak filter caps cause increased hum. They can cause distortion as ripple modulates signal. Bad enough they lower the voltage in the supply. But I haven't figured out how they will cause the output stage to fail. I can see it making the circuit not work well, but I don't see it killing transistors.

Just my opinion.

[PaulBass]

not only the filter caps its every electrolytic in the amp. they are all chinese junk. I tested some of the film caps and they were in spec

[g1]

I'm not doubtful about the junk caps, I'm sure these are mass produced units that are replaced and not repaired.
But it is also possible they are using cheap STK knock-offs in there that are blowing on their own and not because of the caps.
If you replace the output device and all the caps and it works then that's great, but it doesn't necessarily follow that the caps were the cause of the failed output devices. I suppose it is possible, probably more so from output circuit caps than filter caps.

[J M Fahey]

And as Enzo often says, **how many** are actually blowing?

Of course we get the bad ones on our benches, but 10 bad out 100 is appalling, same out of 5000 is excellent quality.

As of "chinese junk" ... I guess nowadays there´s nothing else on dealer´s shelves , and replacing OEM ones will most probably involve more chinese made ones.

[Enzo]

Plus I am still always interested in what is behind these claims. I would be perfectly happy to accept that bad caps blow up these output chips, but only if I am presented with some sort of reasoning behind HOW they do that.

[nickb]

Plus I am still always interested in what is behind these claims. I would be perfectly happy to accept that bad caps blow up these output chips, but only if I am presented with some sort of reasoning behind HOW they do that.

I have recently seen four cases high ESR caps in the feedback path lead to HF instability. The transistors were saved by virtue of the SMPS over current protection kicking in.

[nickb]

Going back to the original subject. It occurred to me, rather belatedly, that a TDA7293 could be made to work as a substitute. The two biggest problems would be that the heatsink is connected to the -ve supply and so would require isolation and that they require a low thermal resistance to run at any power for any length of time.

On a different tack, an odd thing is that these fakes took a great deal of effort to engineer. They have in injection moulded cover, a heatshink, a leadframe as well as the PCB. What a shame they apparently skimped on the all important transistor dissipation.

[J M Fahey]

Close, but unfortunately no cigar.
These Sanken modules are rated 120V or so, are happy with +/-52V rails, have large die power transistors (the real ones I mean, not the fakes) and have huge thermal contact surface to heatsink.

TDA729x aren´t reliable beyond 45V rails and in fact are better restricted to +/- 42V or so, power chips are way smaller by necessity, not enough space, and contact surface is exactly 2 x TO220 packages side by side.

They are simply not enough and explode trying to do their big brother´s job.

How do I know? ..... well, I tried anyway

As a side note: LM3886 are known as very reliable, simply because they pull 50 or 60W RMS out of that puny package and succeed; TDA729x falsely claims 100W in the datasheet ... and punishes those who believe it.

In my book, they are good up to 80W RMS, period, meaning +/-40V into 8 ohms load or +/-30V or so and 4 ohm load ... and then they work forever.

But 100W RMS? "optimistic" is the understatement of the Year.

[nickb]

Close, but unfortunately no cigar.
These Sanken modules are rated 120V or so, are happy with +/-52V rails, have large die power transistors (the real ones I mean, not the fakes) and have huge thermal contact surface to heatsink.

TDA729x aren´t reliable beyond 45V rails and in fact are better restricted to +/- 42V or so, power chips are way smaller by necessity, not enough space, and contact surface is exactly 2 x TO220 packages side by side.

They are simply not enough and explode trying to do their big brother´s job.

How do I know? ..... well, I tried anyway

As a side note: LM3886 are known as very reliable, simply because they pull 50 or 60W RMS out of that puny package and succeed; TDA729x falsely claims 100W in the datasheet ... and punishes those who believe it.

In my book, they are good up to 80W RMS, period, meaning +/-40V into 8 ohms load or +/-30V or so and 4 ohm load ... and then they work forever.

But 100W RMS? "optimistic" is the understatement of the Year.

Luckily, I never liked cigars anyway. Never understood what the pleasure was in them.

The TDA would only be a candidate to replace the lowest power 70W version. You could mount two operating in a bridged or parallel for more power (lower out Z). That will get you the power but it doesn't solve the supply voltage issue and the replacement would need to be compatible.

[PaulBass]

I'm not doubtful about the junk caps, I'm sure these are mass produced units that are replaced and not repaired.
But it is also possible they are using cheap STK knock-offs in there that are blowing on their own and not because of the caps.
If you replace the output device and all the caps and it works then that's great, but it doesn't necessarily follow that the caps were the cause of the failed output devices. I suppose it is possible, probably more so from output circuit caps than filter caps.

Is there a way to tell if the STK's were knock offs? They looked legit to me. They looked exactly like the ones I yanked out. I bought them from a reputable US seller

[The Dude]

That's kind of what the thread is about. Start at post 1. There's not really a way to tell unless you crack one apart. The originals are no longer in production, so most all of them are fakes.

[J M Fahey]

True.

That said, "another" semiconductor maker *might* make a good quality copy and corner a small but still interesting niche market, there´s tens of thousands (if not Millions) STK modules out there.

No need to be a *full* manufacturer, meaning no need to grow NASA technology own crystals, most large manufacturers sell unmounted dies in bulk for the power transistors and the rest is just an SMT circuit, easy peasy and cheap nowadays.

If I were a young Engineer in China and wanted to start up something with a few friends I would consider it.

Back to the one you bought, only actual use can prove what you have.

[diydidi]

JM Fahey.
Please share your design and pcb if possible.

[nickb]

diydidi: Edit your port to remove your email address (unless you like spam).

[PaulBass]

I'm just curios if anyone has enough experience with these things to tell me whether this might be a counterfeit or not. It was purchased as "Original Sanyo". I'm having serious doubts if that's true. Looking at the actual output devices within the part (on the heat sinks), they look substantially smaller than the ones from an original part and it seems suspect to me. I could certainly be wrong. It's been a while since I cracked one of these open. The reason I ask is because this is the second failure in this amp- both times the STK was replaced. The IC's for each repair came from different vendors. I want to make sure I get a "real" part next time, if that's even possible. I'm thinking NOS versions of these are rare as hen's teeth. BTW, the amp is the GC Acoustic B200H.
For your viewing pleasure:

[ATTACH=CONFIG]45061[/ATTACH]

Thats definitely a fakeroo. I popped one open I bought on Ebay and it looked nice and neat like the STK-130 on the previous page. The problem is it only lasted 2 years before it started distorting. I put in another module and the distortion stopped. I don't know if they are fakes or I keep blowing them up https://www.ebay.com/itm/12174070521...QAAOSw3ydV251k