In most situations, you can improve a design by substituting inferior parts with better parts (e.g. lower noise) BUT the rule is that when people design an amplifier, the "surroundings", the details of the circuit, is made to support the operation of the chosen transistors...
If the amp has been designed with superior transistors, problem may arise when you substitute with lesser transistors. You may end up with adding compensation caps everywhere... (and this FX12 topology is not "simple")
Hello Sir,
I am a diy beginner from India,
Can you please tell the brand of capacitor you are using?
Thanks and regards.
Sumesh
I am a diy beginner from India,
Can you please tell the brand of capacitor you are using?
Thanks and regards.
Sumesh
Hi Mile,
The addition of the two 33p seems to have eliminated the oscillation however the 10p cap made it worse. It plays fine now but I see a little overshoot on the square waves.
Attachments
I had the same effect. But comes from cheap function generator (picture at 20kHz)
regards Olaf
Attachments
My point is that sometimes the overshoot will appear at the input as well. Probably because of feedback. I have seen it before.
Yes I understand. I'm really not an expert, but when I compare with my measurements, I see that they are much more symmetrical. I have the transistors all measured for the same gain. I do not know more what to do.
I think, something more basic has to be solved first. From the DC op, something is still not right. Could be a wrong/broken transistor or LED. Because voltage drop across a LED is more independent, it shouldn't be accepted if one LED has 0.7v and the other has 0.8 something...
The 0.7v should be replaced if it is not reversed...
what do i have to do to fh9 if im going to use 20v+/- supply with it instead of 35? the reason being that i have 25v psu caps readily available and my speakers are high efficiency.
40v caps are cheap. You will be much happier with the sound with the higher voltage PSU vs 20v. I have tried 20v supplies before and it's just not as good. There is probably a way to do it by lowering the resistors that limit current into the CCS LTP and VAS stages but I think changing all those resistors is more complicated than buying some 35/40/50v caps.
thank you bimo. Ok I will order those KSC1845 from digikey. I just figured out that almost all the "out of production" parts in your mod are still available on digikey. Beginners mistake.
2sc1845 = KSC1845FTA (F grade, Hfe 300-600)
2sc3503 = KSC3503DS-ND
2sa1381 = KSA1381ESTU-ND
I think I will just fill in my digikey order since these are not expensive. Kind of a waste of money on my aliexpress orders.
I also managed to find all the various Cordell and keantoken models so I can run and study your fx8 LTspice simulation. I will build up the Prasi layout boards. This will be a good learning experience and I should have a nice amp in the process.
I always prefer lower v sounds more. All the amps ive tweaked to lower v always sounds darker/warmer, im guessing due less excited components from lower voltage? Perhaps less thermal noise? Would love an explanation on this.
And also the loss of headroom can be compensated by more cap values which have now become cheaper with the lower rating. And with high eff speakers 30,40watts is still plenty. Also smaller heatsink, less footprint... too much is in favor of lower voltage for me.
The KSC3503 are D-grades where the KSA1381 are E-grades. Using transistors with such different Hfe ratings may cause problems.
Personally, I would rather use the same grade for both.
I don't know for sure it will cause problems, but I brought this point up a while back in this thread.
The fact remains that the A1381 and C3503 are hard to source unless you use the mismatched KSC replacements. I see a few original Sanyo devices on eBay that appear authentic, but I'm not going to pay the ridiculous price they want for them.
Perhaps Bimo or Mr. Mile can recommend suitable substitutions.