I tried two locations for the thermistor: 1) screwed to the unused source terminal, and 2) Arctic sliver attachment to the center of the SOT-227 package. While making measurements of the thermistor I also took measurements with an ExTech multimeter with a temperature probe attached to the source terminal, and measurements with a digital oven thermometer probe attached to the underside of the heatsink, between fins, directly beneath the SOT-227. The thermistor attached to the source terminal seemed to track best with the other measurements and responded quickest to changes in heatsink temperature.
Ok, thanks. I will not give up on that location just yet then.
I've increased current from 1.4mA to 2.8mA for TL431 just to be sure.
I've been getting some really strange results. Not sure if it is the thermistor, bad contacts, tl431 playing up.
I need to get more multimeters out of the garage.
And it is no problem that the thermistors tracks only the one puck connected too?
I placed the thermistor as usual in the middle between the two pucks.
And I can tell you the bias point with the TL solution is stable. When warming up, the value even gets a bit smaller, no running away at all.
I placed the thermistor as usual in the middle between the two pucks.
And I can tell you the bias point with the TL solution is stable. When warming up, the value even gets a bit smaller, no running away at all.
And it is no problem that the thermistors tracks only the one puck connected too?
I placed the thermistor as usual in the middle between the two pucks.
And I can tell you the bias point with the TL solution is stable. When warming up, the value even gets a bit smaller, no running away at all.
Either way works.
Mine is working now too, but I'm gonna tweak it a little.
Speaking of thermal runaway: Recently I was making measurements of an IRFP240 with no degeneration and no load resistor. With Vds=32V when I increased Vgs to over 4.5V the fet entered a thermal runaway, rapidly increasing current with no change to Vgs. Fortunately, my bench power supplies limited the current to 3A so no harm was done. The thermal conductivity from junction to case of the hocky pucks is so good that it would require Vds and bias current conditions far beyond what I can test.
That is very good to hear!
🙂)
But some people trying to build the BAF 2015 amp with hockey pucks and Jensen seemed to have a current runaway despite the SOT-227 case.
Meanwhile after the comments of Nelson i suppose they had a kind of motorboating.
🙂)
But some people trying to build the BAF 2015 amp with hockey pucks and Jensen seemed to have a current runaway despite the SOT-227 case.
Meanwhile after the comments of Nelson i suppose they had a kind of motorboating.
That is very good to hear!
🙂)
But some people trying to build the BAF 2015 amp with hockey pucks and Jensen seemed to have a current runaway despite the SOT-227 case.
Meanwhile after the comments of Nelson i suppose they had a kind of motorboating.
I didn't get thermal run away with the BAF2015 amp.
I stopped completing it because it was going to be quite an expensive build. I'll get back to it later.
recycling
Finally got a replacement phone.
Decided to put the bias board from BA2 output stage to use.
I'll start building up the output stage tonight.
Finally got a replacement phone.
Decided to put the bias board from BA2 output stage to use.
I'll start building up the output stage tonight.

More Recycling
Front end PCB.
Very little will be used on the board but it has the circuit topology, I want.
Front end PCB.
Very little will be used on the board but it has the circuit topology, I want.

Starting to assemble output stage
Decided to cancel the idea of recycling the bias board. Going to build one on proto board instead.
The heatsink below is purely for prototype development. The plan is to purchase four MF25-151.5 (2 per side) http://www.conradheatsinks.com/products/flat100_350.html#MF25 for 500mm deep case.
Decided to cancel the idea of recycling the bias board. Going to build one on proto board instead.
The heatsink below is purely for prototype development. The plan is to purchase four MF25-151.5 (2 per side) http://www.conradheatsinks.com/products/flat100_350.html#MF25 for 500mm deep case.

Last edited:
Have a good start building the F4 beast!
I changed from the APT and IXTN170P10P to the IXFN48N60P and IXTN40P50P and the sound is much better.
At the moment k2 is even too much.
Research goes on.....
I changed from the APT and IXTN170P10P to the IXFN48N60P and IXTN40P50P and the sound is much better.
At the moment k2 is even too much.
Research goes on.....
remedy for too much k2 was easy.....
two 100k cascodefeedback resistors on the lower side of the circuit.
Still looking for the best value with my build. At any case it is easy to play here and adjust the portion k2 ones want.
Now I am again fascinated by the sound and cannot imagine what can get better with the lateral TO-220 coming in some days.
But the hard test will be again at my friends house against the Aleph 3 and his Klipschorns, maybe next week.
Maybe I will hesitate a bit to tell you the truth........😀😱😀😱
two 100k cascodefeedback resistors on the lower side of the circuit.
Still looking for the best value with my build. At any case it is easy to play here and adjust the portion k2 ones want.
Now I am again fascinated by the sound and cannot imagine what can get better with the lateral TO-220 coming in some days.
But the hard test will be again at my friends house against the Aleph 3 and his Klipschorns, maybe next week.
Maybe I will hesitate a bit to tell you the truth........😀😱😀😱
remedy for too much k2 was easy.....
two 100k cascodefeedback resistors on the lower side of the circuit.
Look at my Front end circuit again. I think you overlooked something. 🙂
Maybe I will hesitate a bit to tell you the truth........😀😱😀😱
We are all brothers here.
Hehehehe
Front end PCB.
Very little will be used on the board but it has the circuit topology, I want.
I have the same front-end boards and was planning to use them as well. Won't need the relay features, though. Hope my source for the bipolar cascodes supplied legitimate parts...
remedy for too much k2 was easy.....
But the hard test will be again at my friends house against the Aleph 3 and his Klipschorns, maybe next week.
Maybe I will hesitate a bit to tell you the truth........😀😱😀😱
Alternatively, on his speakers you might want more (completely different operating point). That is the beauty of my little trick, you can adjust it on the fly to what sounds best to your ears on your speakers.
Look at my Front end circuit again. I think you overlooked something. 🙂
There was nothing new for me, I did this since 1917...... 😀😀😀
There was nothing new for me, I did this since 1917...... 😀😀😀
I know, cause we are telepathic alien hybrids.
😀
I ordered some TO-220 lateral Hitachi Mosfets SJ79 and SK216 here
Pacific Semiconductors Inc.- Worldwide Distributor-Import-Export
hope that they are no fakes.
I just had an idea for a completely different audio project. Thanks brother.
One in, all in.
I just bought some too.
We need be experienced and respect all councils from Papa that all ok.That is very good to hear!
🙂)
But some people trying to build the BAF 2015 amp with hockey pucks and Jensen seemed to have a current runaway despite the SOT-227 case.
Meanwhile after the comments of Nelson i suppose they had a kind of motorboating.
I am curious about Mr Rothacher Schade 50W single-ended diy.
Hope we get the last informations about this interesting amplifier build & sound.
If his house are finished probably Audiohobby.com show us work at new lab bench.
Like from excellent Labhangout video series ...i stay tuned because is miss me a lot

Output stage assembled and tested. Everything works perfectly.
Next up will be to solder it all up on protoboard and play some music through it.
Next up will be to solder it all up on protoboard and play some music through it.

Last edited:
- Home
- Amplifiers
- Pass Labs
- F4 Beast Builders