From what I have read the early production thermaltraks had issues with the integrated diode.
If that diode goes open it can switch the amp into class A bias depending on the bias circuit used, In most amps this should just blow the fuses after a minute or so.
I assume that a lot of the amps that used the thermaltraks were "high current" amplifiers with oversized fuses, power ratings into <3 ohms, and no current limiting on the outputs.
Combine that with faulty thermaltraks the amps will think they are running at max power when they are really nuking every component connected to the output stage.
I'm not excusing onsemi, the Thermaltraks clearly needed more work before they were put onto the market.
Back on the topic of replacement stks I'm mainly interested in trying to make boards with nothing more than the output transistors mounted directly to the heatsink, obviously that doesn't seem possible without compromises of some sort or the use of exotic/non-standard components that may be obsolete in 10 years, but it is fun to experiment.
I do already know all the issues with my design, I was just interested in testing it as is, also if anyone else has tried thermaltraks in Stk alternatives.
Also there's nothing wrong with 100 milliohm emitter resistors as long as the bias is stable and the outputs are well matched, I went low resistance to allow the use of 3 watt emitter resistors.
That’s a fair point, and I largely agree with the concerns around how these devices have been used historically.
One thing I would add is that STD / ThermalTrak–type devices were primarily developed for
very dense, multi-channel layouts, where space is limited and thermal behavior is tightly managed at the system level. When they’re pushed hard, biased aggressively, or dropped into vintage amplifiers with little or no protection, the failure modes can indeed be quite dramatic.
That’s also why, if someone is interested in this general direction, I would strongly recommend
a conservative implementation, rather than treating these devices as a drop-in upgrade.
For that reason, I’ve been working on my
own STD-based discrete STK replacement modules, designed specifically for:
- moderate rail voltages
- adequate emitter resistance
- conservative biasing
- and predictable failure behavior
The goal is not to extract maximum power density, but to achieve
stable operation and long-term reliability in older amplifiers that were never designed around modern protection schemes.
Nem javasolnám ezt a megközelítést minden alkalmazáshoz, de azok számára, akik már az STD-alapú kimeneteket próbálják ki, és egy
szabályozott, szolgáltatásbarát megvalósítást szeretnének , ez egy ésszerű alternatíva lehet.
I should also clarify one important detail.
This STD-based version was originally developed by me specifically as an
STK1050 / STK0050 discrete replacement, not as a general-purpose STD output module. One of the advantages of this approach is its
compact size — it fits into very limited spaces — and the fact that the
bias control network can be kept relatively simple, which is attractive from both a layout and service perspective.
At the same time, this approach has its own downside, and it’s important to be honest about it:
original STD devices are
already becoming difficult to source, and if that trend continues, a failed module years down the line could face the
same long-term serviceability problem that originally motivated STK replacements in the first place.
So while this STD-based solution can make sense technically — especially where space, thermal behavior, and conservative biasing are priorities — it is not a universal or future-proof answer. In that sense, it shares some of the same lifecycle risks as the original STK modules it was designed to replace.
Because of that, I see it as
one possible option, not a blanket recommendation, and something that should be chosen with a clear understanding of both its strengths
and its long-term limitations.
