Bonsai maybe you ve done some measurements, dont you find higher order harmonics actually being higher relative to the total THD obtained when using Low feedback on your designs. I use mostly current feedback amps and did not look at this intently till now, but clearly the pattern as JLH predicts happens. Local vas feedback is not a good option with CFB amps though because of stability.
Precisely because it allows me to experiment between lower loop gain, lower VAS stage local gain and the high loop gain options. High loop gain is not a prerequisite for success - I think the sound of both my amps attest to that.
How would we know ...
Yes, higher order harmonics are greater with low feedback. I do not want to get into a feedback debate because on hi feedback, low feedback, no feedback I am basically agnostic. For me what's important is that if I decide to in any of the directions, I can do it and get good results.
OT/ BTW I am just waiting for PCB's and parts for two new amps -sx and nx that will use CFB. But, these are very stripped down designs targeting simplicity. Still a few weeks of work, but I will blog it as I go along ok my website /end OT
Dadod I havent had time for proper inspection, so far I just changed all the semiconductors and looked at a quick sim, The figures were worse but Im using models made through private company and I find them more accurate than most others so I was expecting this anyway. Stability wise things looked good though, TMC puts the performance of this circuit on another level.
Low Power version
Here is low power version and very good to start with this amplifier.
Challenge, now, it to design good PCB according to discution we had in other thread starting somewere here: http://www.diyaudio.com/forums/solid-state/209658-has-anyone-seen-front-end-before-88.html
I know that here, in thst forum, are very skilled layout designers capable to do good job. I will tray to do my best also and suggestion and a layouts are wellcome.
dado
Here is low power version and very good to start with this amplifier.
Challenge, now, it to design good PCB according to discution we had in other thread starting somewere here: http://www.diyaudio.com/forums/solid-state/209658-has-anyone-seen-front-end-before-88.html
I know that here, in thst forum, are very skilled layout designers capable to do good job. I will tray to do my best also and suggestion and a layouts are wellcome.
dado
Attachments
Thanks to catalin http://www.diyaudio.com/forums/solid-state/135122-wonderfull-sziklai-21.html here is a variation on the TCM. What do you think if it could work, simulation is OK? TCM capacitors and resistor could go lower.
dado
dado
Attachments
Did you have the opportunity to go through more simulations with your last proposal? What is the THD obtained with the new feedback combination?
Any prototype realized?
Moreover, to have an idea of the C-multiplier effect, did you run simulations using rippled power supply or... simulate using a power supply schematic?
Any prototype realized?
Moreover, to have an idea of the C-multiplier effect, did you run simulations using rippled power supply or... simulate using a power supply schematic?
Thanks for the interest,
I simulated it long time ago, I need a time to find out what I did. If you are very interested I could do more tests for you? I never built that one, but I built AudioCube TT amp(from the same thread). As that one use separate power regulator(C multiplier) the result should be even better as here the C multipliers are close to the amp on the same PCB. Let me know if you are interested and I'll do more simulations
BR Damir
That one was with four output pairs, never made that amp. One guy was interested and started to design the PCB, but then no more news from him.
Dadod, you are clever person who searching for perfection. But you must try to design more simple. The implementation of simple design have more chance to success. You should to considered the difficulties of the implementation.
I already simplified your design. Btw, thank you for sharing your design.
I already simplified your design. Btw, thank you for sharing your design.
I am not searching for perfection, just following D. Self rule, to make “blameless” amp, but not Self copy.
This amp looks complicated because it integrate cap multipliers, or power regulator on the same PCB, with all needed protection. Most complicated part here is to make good layout and then everything looks simple. One example of the “blameless” principle is how LTP cascode was biased. If you want high power amp and want to use jfet or high beta small transistor(normally low voltage) then you need cascode. Standard way to bias cascode is resistor divider referenced to the ground. Better if it is referenced to the LTP tail, but then the cascode current disturbs the tail current. Small transistors are cheap and way not use one to bootstrap cascode to the LTP tail, and next step to replies lower bias resistor with constant voltage like zener or LED. Next step would be to replace upper resistor with the CCS, but then that is “complicate”. I did that in my first TT amp, but I think that was over the top.
Thank you in your interest. Could you show how did you simplified my design?
I try to simplified the IPS. I have already draw modular PCB of 3EF OPS like Ostripper. My target is making simple amp to build but high performance. I am not finishing the simulation. May be you can help me.
No baker clamp, the clipping behavior is good.
Attachments
Last edited:
The result is:
137W/8Ohm, 1kHz, 0.000227%
137W/8Ohm, 20kHz, 0.000225%
It is good enough for me.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.