Im pretty sure this have already been done many times over but i'll post it anyways.
I just came up with this based on five pairs of IRF540/9540 i have.
First a simulation: http://i.imgur.com/qwiho.png
And a readable schematic: http://i.imgur.com/hq1yk.png
Yes i know, the simulation doesent look all that impressive, but keep in mind i do all my simulations at 20kHz, most amps can do 0.005% THD at 1kHz and still exceed 1% at 20kHz.
However if THD can be knocked down to 0.05% at 20kHz, it should make for a really good amp in my opinion, even if it may be too perfect, making the sound lifeless and flat.
I just came up with this based on five pairs of IRF540/9540 i have.
First a simulation: http://i.imgur.com/qwiho.png
And a readable schematic: http://i.imgur.com/hq1yk.png
Yes i know, the simulation doesent look all that impressive, but keep in mind i do all my simulations at 20kHz, most amps can do 0.005% THD at 1kHz and still exceed 1% at 20kHz.
However if THD can be knocked down to 0.05% at 20kHz, it should make for a really good amp in my opinion, even if it may be too perfect, making the sound lifeless and flat.
Hi Tekko,
It resembles Elektor HeXFET amp, except Elektor one has CFP style output with some gain.
Thanks,
Routhun
It resembles Elektor HeXFET amp, except Elektor one has CFP style output with some gain.
Thanks,
Routhun
The input stage and the VAS in this amp are going to have problems unless the transistors are closely matched for beta. The problem is that the voltages at the collectors of Q1 and Q3 are not well defined. Since those same voltages are at the bases of Q6 and Q7, and thus determine the currents in the 330 ohm emitter resistors, the current in the VAS chain is also undefined. This is a common problem with this topology, using complimentary circuitry with the input stage loaded by an active source. Bob Cordell discusses this in his book, and I think there's some discussion of it somewhere in the book thread in this forum. Randy Sloan's book also has several designs using this topology, and there's been lots of discussion of those designs also in this forum.
I've never had any problems with matching for input stages when using the current mirrors and constant current sources.
I have however never actually built a fully symmetrical circuit like this before.
I have however never actually built a fully symmetrical circuit like this before.
This here i choose to call SymiFet Deluxe, based on the old Elektor HEXFET amp, using current sources for the input stage and beta enhanced VAS: http://i.imgur.com/fBNw2.png
And if the LED/diode type current source isent your flavor, the dual transistor one will also work just fine, if not better.
I do have an old transistor tester that measures the hFE so i can do matching if necessary, assuming that old thing still works and doesent use some long obsolete obscure unobtanium battery format.
And if the LED/diode type current source isent your flavor, the dual transistor one will also work just fine, if not better.
I do have an old transistor tester that measures the hFE so i can do matching if necessary, assuming that old thing still works and doesent use some long obsolete obscure unobtanium battery format.
Old habit. You know the drill 😛
I toss them up on imgur to share on IRC and other instant messaging clients with my "friends" there, and on forums that doesent have a built in image hosting feature.
I toss them up on imgur to share on IRC and other instant messaging clients with my "friends" there, and on forums that doesent have a built in image hosting feature.
And the verdict is: It was all just 100% waste of time!
It does not work at all, all it does is put the negative rail on the inputs of the ltp which puts the speaker output at either the positive or negative rail.
On the initial powerup i had wired the ltps wrong which smoked the current mirror resistors, but i unplugged power fast enuf to avoid burning them, but after fixing that i still had no luck, i replaced two transistors that measured dead short between B and E which did no difference at all!
Only thing that does anything is breaking the contact between the bases of the upper and lower LTP.
I highly doubt closely matched transistors would make any difference either.
It does not work at all, all it does is put the negative rail on the inputs of the ltp which puts the speaker output at either the positive or negative rail.
On the initial powerup i had wired the ltps wrong which smoked the current mirror resistors, but i unplugged power fast enuf to avoid burning them, but after fixing that i still had no luck, i replaced two transistors that measured dead short between B and E which did no difference at all!
Only thing that does anything is breaking the contact between the bases of the upper and lower LTP.
I highly doubt closely matched transistors would make any difference either.
This topology has been made to work, but it usually involves providing some sort of reference voltage to help define the operating conditions of the VAS. I'm sorry that I can't say much more than that, because I only glanced at the various discussions of the subject on this forum. I will say that those who have been able to make it work have been very pleased with the performance, so it might be worth while searching in the forum archives to find some of the solutions.
I did a quick search, here are a few threads on the subject:
http://www.diyaudio.com/forums/solid-state/133864-complementary-input-stages-current-mirrors.html
http://www.diyaudio.com/forums/soli...ror-image-topology-construction-troubles.html
http://www.diyaudio.com/forums/soli...one-amp-ie-yet-another-symetric-topology.html
http://www.diyaudio.com/forums/solid-state/16796-unstable-vas-current-amp-slone-book.html
As you can see, it's a problem that's been extensively discussed. I still think Bob Cordell had the best solution, but I couldn't turn it up with the simple search I did. His comments are worth searching for.
Mike
http://www.diyaudio.com/forums/solid-state/133864-complementary-input-stages-current-mirrors.html
http://www.diyaudio.com/forums/soli...ror-image-topology-construction-troubles.html
http://www.diyaudio.com/forums/soli...one-amp-ie-yet-another-symetric-topology.html
http://www.diyaudio.com/forums/solid-state/16796-unstable-vas-current-amp-slone-book.html
As you can see, it's a problem that's been extensively discussed. I still think Bob Cordell had the best solution, but I couldn't turn it up with the simple search I did. His comments are worth searching for.
Mike
Well the lower LTP putting out -40V on the bases seem more like a issue with the 2N5401's rather than the usual issues discussed about this topology.
This amp will never work properly, like mjurban already mentioned the bias currents for the vas are not defined, rather ask Bob Cordell for his solution as published in his book. I dont think he has shown it here yet.
Your talking about the first circuit i posted that just have the 18k resistors in place of current sources, right ?
Tekko Im afraid the second suffers the same fate. Mjurban has shown the threads where the problem is discussed, Cordell has the simplest solution.
I've replaced all the input stage transistors and now it no longer does what it did before, now its showing signs of life as in a almost square wave oscillation which for me is a victory as yesterday was pretty much a day of fail.
Now to replace the 1.8 ohm resistors i accedently used with something bigger and see about adding some caps to the VAS's and the drivers.
Now to replace the 1.8 ohm resistors i accedently used with something bigger and see about adding some caps to the VAS's and the drivers.
So i assume the issue everyones talking about is that i cant be biased, its either zero bias or near dead short bias.
Thats where im at right now, the amp works and play music but i cant bias it.
Thats where im at right now, the amp works and play music but i cant bias it.
So this is the actual circuit as it stands right now: http://i.imgur.com/96bKv.png
I ran out of 2N5551's so i dident build the current sources. The dc offset on the output is roughly 10mV accoding to the scope.
I have not yet decided if im gonna keep working on this or just scrap this circuit alltogether and stick to the blameless topology i've been using up to this point.
I ran out of 2N5551's so i dident build the current sources. The dc offset on the output is roughly 10mV accoding to the scope.
I have not yet decided if im gonna keep working on this or just scrap this circuit alltogether and stick to the blameless topology i've been using up to this point.