Hi Edmond,
Here is square waves test at 5k, 100k, end 200kHz.
It is still with my input stage, GNF, and input filter. What do you say about it?
Damir
Attachments
Hi Damir,
It doesn't look good.
edit: In the meantime I tried to make the high impedance low offset version working,
but I'm still experiencing troubles: above 100kHz, let alone square waves, it goes berserk.
So I give up and try to improve the original version of AB2.
Regarding to offset, you might use a DC servo to zero both voltage and current offset.
What about that?
Cheers,
E.
It doesn't look good.
edit: In the meantime I tried to make the high impedance low offset version working,
but I'm still experiencing troubles: above 100kHz, let alone square waves, it goes berserk.
So I give up and try to improve the original version of AB2.
Regarding to offset, you might use a DC servo to zero both voltage and current offset.
What about that?
Cheers,
E.
Last edited:
Hi Edmond,
Do you mean to use a DC servo for AB2 as OPS for tube preamp, or AB2 used in all bipolar amp with GNF? I am interested in both, thermal and bias stability is annoying with standard bias spreader even if TT transistors are used.
Damir
mods
Looking at some modifications of my circuit, I got the impression that not everybody fully understand every detail of the AB2 circuit. Of course, those mods do work (more or less), but as they deviate too much from Class-I and AB2, I think it's more appropriate to post those circuits on a more general thread about sliding bias circuits. For example this one: http://www.diyaudio.com/forums/solid-state/145013-variable-operating-biass-output.html#post1841397
In order to ease the distinction between circuit which belongs to Class-I (and look-alikes) and those which do not, what about an (short) explanation of the basic principles of Class-I and AB2?
Any interest?
Cheers,
E.
Looking at some modifications of my circuit, I got the impression that not everybody fully understand every detail of the AB2 circuit. Of course, those mods do work (more or less), but as they deviate too much from Class-I and AB2, I think it's more appropriate to post those circuits on a more general thread about sliding bias circuits. For example this one: http://www.diyaudio.com/forums/solid-state/145013-variable-operating-biass-output.html#post1841397
In order to ease the distinction between circuit which belongs to Class-I (and look-alikes) and those which do not, what about an (short) explanation of the basic principles of Class-I and AB2?
Any interest?
Cheers,
E.
servo
Hi Damir,
I mean a servo for the preamp version. Integrated with an all bipolar amp is another story. It depends on the input stage whether you need a servo or not.
Cheers,
E.
Hi Damir,
I mean a servo for the preamp version. Integrated with an all bipolar amp is another story. It depends on the input stage whether you need a servo or not.
Cheers,
E.
This a great to explain all what is in those circuits, I am very interested.
Damir
Hi Homemodder,
Probably you mean Kendall Castor-Perry.
Anyhow, his paper has been published in Jan's Linear Audio Vol. 2.
Perhaps Jan is willing to scan 24 pages and send them to you
A patent of the same circuit can be found here: http://www.diyaudio.com/forums/ever...-linear-audio-publication-41.html#post2799107
Please, look here: http://www.diyaudio.com/forums/soli...erview-negative-feedback-336.html#post2433940
Thanks, the same to you.
Cheers,
E.
Hi, I am also re introduce my cross management in other thread that ever mentioned at "Variable operating biass output ?" thread, it is fun.
May right or wrong, as I know AB2 is simple cascoded current feedback with that cross connection for smaller current gap, and midle transistors for keeps output not turned off.
Correcting error with rejecting error using NFB. Output is current, thats why additional buffer or driver will increase its gain, but decrease its stability. I see no canceling error method except two different Re values, so almost pure rejecting.
Don't worry I like to learn.
I think that not-turned-off is not reject crossover distortion but little help, and there are many type of non switching, class-i, kendall, nakayama, edmondAB2, kenpeter, nelsonpass, may be many more, i think i have one
.This one reject crossover distortion into 5mV scale.
Attachments
Someone else also stole your baby, as early as 1978: Patent US4145666 - Multistage amplifier circuit - Google Patents
The same topology is described (in the context of mos ic op amps) in:
Rudy G.H. Eschausier and Johan H. Huijsing. "Frequency Compensation Techniques for Low-Power Operational Amplifiers". Kluwar Academic Publishers, 1995.
Pdf easy downlod: Multistage amplifier circuit
What the use of make simple R-C circuit patented? Anyone will have to place it if needed, even if he never reads books or paper about it has special function or patented. RC could be used in AB2 to cancel its timing error by added a bit HF gain.
BTW: I have many good stuff in my side, also I have very good performance amplifier. and I like duck toy that quack when pressed or hit.
What the use of make simple R-C circuit patented? Anyone will have to place it if needed, even if he never reads books or paper about it has special function or patented. RC could be used in AB2 to cancel its timing error by added a bit HF gain.
BTW: I have many good stuff in my side, also I have very good performance amplifier. and I like duck toy that quack when pressed or hit
.
TMC
Hi Ontoaba,
The virtue TMC (a simple RC network) wasn't obvious to everybody. Only a very few people recognized its potential: about 15dB distortion reduction at 20kHz. In the past, even an expert as D. Self dismissed it as being 'not exiting'. In this regard TMC is very patentable. In the early nineties, I also brought TMC to several expert to their attention, but nobody was interested. The majority (stupidly) thought that such a simple RC thingy could never make such a large difference. Later I brought TMC to a larger audience (this forum) and in the beginning only Bod Cordell saw its virtues. Even now some nuts stubbornly think that TMC is BS. This is why TMC is worth a patent, because it does have commercial value and it's not obvious.
What do you mean by 'timing error'?
BTW, again we are drifting away from the main subject: Class-I and AB2.
Tomorrow I hope to drop my explanation about AB2. It'll take some time as I'm not a writer and English is not my native language.
Cheers
E.
Hi Ontoaba,
The virtue TMC (a simple RC network) wasn't obvious to everybody. Only a very few people recognized its potential: about 15dB distortion reduction at 20kHz. In the past, even an expert as D. Self dismissed it as being 'not exiting'. In this regard TMC is very patentable. In the early nineties, I also brought TMC to several expert to their attention, but nobody was interested. The majority (stupidly) thought that such a simple RC thingy could never make such a large difference. Later I brought TMC to a larger audience (this forum) and in the beginning only Bod Cordell saw its virtues. Even now some nuts stubbornly think that TMC is BS. This is why TMC is worth a patent, because it does have commercial value and it's not obvious.
What do you mean by 'timing error'?
BTW, again we are drifting away from the main subject: Class-I and AB2.
Tomorrow I hope to drop my explanation about AB2. It'll take some time as I'm not a writer and English is not my native language.
Cheers
E.
OK, Edmond is not inventor of TMC, but he brought it to attention here to DIY community, almost forgotten art, and I am grateful for that.
Damir
The mane "timing error" may be fit in canceling error mechanism, generally we could say "lagging". Just added zero (phase lead) around pole by litle HF gain using RC to ground at inverting input for simplest, or use canceling EC style by added HF from Vdrop to input.
Yes I hope so. Did you ever planned or make that crossconnection transistors to not switched off and work inclass AB instead, I just have this idea and not yet tried it.
Nakayama is using voltage feedback style, I am not try it yet, may be later.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.