I am planning to get a over size toroidal transformer with a lot of caps. Do you think I need a regulated supply?
If you really want a regulated supply I'd suggest going for an off-the-shelf SMPSU. Of course it'll need some serious LC filtering. Whether you choose a traditional 50/60Hz unregulated or switching supply, you'll still need LC. Yes, L = inductor, C=cap.
With the focus above all on the PSU, this leads naturally to having a bridge-tied load so then you have only one supply to worry about.
With the focus above all on the PSU, this leads naturally to having a bridge-tied load so then you have only one supply to worry about.
It's been a while, I have been reading some schematics from famous amps, Nelson Pass, YBA( my most favorite), Leach and Apexaudio etc. They all tend to be relatively basic and simple.
1) They don't necessary use complementary LTP as IPS.
2) Even using complementary LTP for IPs, they mostly don't use current mirror but instead a simple resistor as load for the LTP.
3) For symmetrical VAS, most don't even use darlington ( discrete of real darlington) to increase the input impedance.
4) The Nelson Pass schematic I saw, the VAS is not even symmetrical, he use a CCS on one side.
5) Very few use 3EF for OPS!!!
I am confused!!! Here you have Cordell talking about a lot of fancy way to lower distortion, I saw schematics from members here that incorporated a lot of Cordell's ideas like complementary IPS, symmetrical darlington VAS and 3EF. But a lot of the well respected high end amps seems to be a lot simpler.
This is particular true for YBA that I really like!!! I really don't know how to make out of this!!!
1) They don't necessary use complementary LTP as IPS.
2) Even using complementary LTP for IPs, they mostly don't use current mirror but instead a simple resistor as load for the LTP.
3) For symmetrical VAS, most don't even use darlington ( discrete of real darlington) to increase the input impedance.
4) The Nelson Pass schematic I saw, the VAS is not even symmetrical, he use a CCS on one side.
5) Very few use 3EF for OPS!!!
I am confused!!! Here you have Cordell talking about a lot of fancy way to lower distortion, I saw schematics from members here that incorporated a lot of Cordell's ideas like complementary IPS, symmetrical darlington VAS and 3EF. But a lot of the well respected high end amps seems to be a lot simpler.
This is particular true for YBA that I really like!!! I really don't know how to make out of this!!!
Build two or three examples of each type, measure them, listen to them, and decide for yourself which one(s) you prefer. Who cares what somebody else likes? Find out what YOU like. Experimentally.
The good news is, most of the cost of an audio power amplifier is in the chassis, front panel, heat sinks, power transformer, and power supply. If you make the amplifier circuit board a self-contained subassembly, you can swap it out while leaving all of the expensive pieces intact. So the cost of changing the amplifier "guts" is kept quite low. Design topology comparison experiments are inexpensive.
The good news is, most of the cost of an audio power amplifier is in the chassis, front panel, heat sinks, power transformer, and power supply. If you make the amplifier circuit board a self-contained subassembly, you can swap it out while leaving all of the expensive pieces intact. So the cost of changing the amplifier "guts" is kept quite low. Design topology comparison experiments are inexpensive.
A well-executed basic Lin topology with EF2 output tends to be quite adequate for the demands of human hearing, at least as long as we are talking the usual double-digit watts into 8 ohms and normally efficient loudspeakers. THD <0.03% (well-behaved), 20Hz-20 kHz, is quite doable. The simpler the circuit, the more critical individual parts performance is going to be though, so in a basic topology you'd be picking your input stage and VAS transistors with great care (and probably be lamenting the numerous out-of-production parts...). You shouldn't be expecting to be able to skimp on output stage bias current either.
Now a 250 W / 4 ohm amp with the same or better (ideally <0.01%) distortion specs, that's a different kettle of fish.
Audio designers presumably also know how to apply the old "if you can't make distortion negligible, at least make it equal across the audio band" trick reliably.
Now a 250 W / 4 ohm amp with the same or better (ideally <0.01%) distortion specs, that's a different kettle of fish.
Audio designers presumably also know how to apply the old "if you can't make distortion negligible, at least make it equal across the audio band" trick reliably.
Thanks for the reply.
I picked KSA992 https://www.fairchildsemi.com/datasheets/KS/KSA992.pdf for PNP and KSC1845 https://www.fairchildsemi.com/datasheets/KS/KSC1845.pdf for NPN LTP transistors.
I am picked KSA1381 https://www.fairchildsemi.com/datasheets/KS/KSA1381.pdf for PNP and KSC3503 https://www.fairchildsemi.com/datasheets/KS/KSC3503.pdf/datasheets/KS/KSA1381.pdf for NPN for the VAS stage.
I mainly looking for low noise, higher fT, low Cob(which corresponds to lower Cib and higher beta.
What else do I want to look for? I am going to use 300ohm emitter degeneration in the LTP and 100 ohm for the VAS to linearize the stage.
What do you have to look out in choosing power BJT? So far I picked MJW3281 and MJW1302 http://www.onsemi.com/pub_link/Collateral/MJW3281A-D.PDF for NPN and PNP power transistor. I choose by fT= 30MHz, beta = 75.
Let me know what I missed.
Yes, Not only that, I even separate OPS from the IPS/VAS pcb. This is going to be even cheaper and less work to swap. I am planning to even using a switch to switch to switch between IPS/VAS board. All the frequency compensations are on the IP/VAS board, the OPS is just a 3EF, 5 pairs in parallel.
- Status
- Not open for further replies.