ELKA PAL100 Crossover Distortion

[Ted]

I have an ELKA PAL100 combo here. It has crossover distortion - visible on a scope and sounds terrible. There is not much to the power amp module (see pic). It basically consists of a Sanyo STK086 power amp module on a PCB with a handful of supporting components.
All resistors and caps measure correct. Power rails are +/- 42V and look clean. Ground connection is good. I'm out of ideas.

[Jon Snell]

Sounds like a faulty STK or maybe a rubbing voice coil in your speaker.

[J M Fahey]

Yes, STK modules are intended for Hi Fi (musical instrument market is minuscule by comparison) so have very low distortion and are obviously well biased.
So if yours shows visible distortion, itīs highly suspect.

[Jon Snell]

Just in case you have to make a discrete version, here is the equivalent schematic;

TOA used them in the BA-603/803 powered automatic cassette players, used in stores to play back ground music and there may still be a few left. I used to repair them for HNew Look, years ago. Take a stroll down Tottenham Court Road.

[Mick Bailey]

Check the power rails when the amp is operating to make sure both supplies hold up under any kind of load.

[Jazz P Bass]

With a failed STK module, sometimes a bit can be learned by pulling the module & checking pin to pin with a meter.
The datasheet has an internal component diagram.

[J M Fahey]

Try to get another, shortest cheapest path.
If *absolutely* unavailable , some brave (desperate?) souls have opened them and repaired them at component level.

Power transistor chips are soldered to backing plane so they are not directly replaceable (and NOBODY will sell you unmounted chips anyway) but I have seen repaired modules connected with short wires to *external* flatpack power transistors, think TO218-247 , even TO220.
Look UGLY but work.

For some simpler STK there are commercially available PCBs with exact same footprint and pinout:




Or you vcan try tomount an existing discrete amplifier on the same heatsink/chassis.

Post a couple inside pictures showing chassis, PCB, and actual STK module.

But first try to find an original one.

Although EBay and Alibaba are not much trusted.

FWIW I have designed my own 50/60W "universal replacement" amp on a very small PCB, to replace obsolete TDA1514 (much used by Marshall, Trace Elliot and others in the 80īs), LM3875/76, currently unavailable, and even TDA7294 and LM3886, cheaper and faster than ordering them by mail.

I use TIP122/126 TO220 darlingtons for a VERY compact design which fits on original heat sink space, go figure.

[Ted]

I got a replacement STK086 from ebay from a seller with decent feedback.

The problem didn't go away.

It's basically a big op-amp isn't it? From the circuit diagram of the device, it looks like it attempts to bias its output devices warm enough that it doesn't need to rely on NFB to get rid of crossover distortion. I've checked the values of all components in the circuit. Replaced the 40 year-old PSU electrolytics. There is maybe 30mV of signal visible on supply rails and a couple of volts of 100hz ripple

The amp isn't precious enough to recreate the module from discrete components. I'm on the point of giving up.

[Helmholtz]

it looks like it attempts to bias its output devices warm enough that it doesn't need to rely on NFB to get rid of crossover distortion

Your OP schematic shows an NFB path to inverting input #3.
But as Hfe strongly reduces at close to zero idle current, NFB cannot completely eliminate crossover distortion with too low idle currents.