what do you think of this schematic?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
here goes with the attachment, very sorry about image quality but I have had to shrink it a lot to get it to post. I may try to cut it in half if it is too difficult to read.
 

Attachments

  • aem6000_4.jpg
    aem6000_4.jpg
    82.8 KB · Views: 4,410
Hmmmmm:scratch: how can I tell You?

If You have some sponsor to order all the devices.....
I think the input stages are overkill for this output stage. Maybe JFETs are good as input stage, but this is the typical example for "less is more".
With the complementary differetial stage (Q11-14) the input stages are not necessary to use.
My point of view that You can get better quality if You redesign this amplifier and remove Q1-Q8.

Sajti
 
Disabled Account
Joined 2002
A three gain-stage design. The final gain stage is notorious for quiescent current instability.

Outside of a configuration with gain in the output stage, i cannot conceive of any circumstances in which three gain stages are necessary, or required.
 
yes you can get PCB's and yes i have built it. I did not mean this to be a guessing game, it is in fact an old "Australian Electronics Monthly" (now defunct magazine) design by David Tilbrook circa 1986.
I had to crop the edges of the page to get the image to post otherwise I would have left those details on.
The designer ackowledges that it is complicated but he was after best possible performance. My impression is that after years of diy audio as a hobby it is still the best sounding amp I have built and has been in regular use now for best part of 9 years.
My original question still remains though for the following reasons. Most circuits i have built have been based on Australian magazines as pcb's have been offered and also availability of parts for those circuits. The Australian mag circuits I have found to be a little unambitious in terms of sonic performance and are optimised for lowest risk in terms of performance when made by constructors of varying skill. I dont mean that to be a criticism either. We do get Elektor down here at some places but I always found the semi's difficult to source.
I guess i am after new challanges and that is a circuit that is going to be little more challanging but has to give better performance, or, do I build a second one of these and go active?
 
this looks like one of those topologies with lots of gain stages to acheive a high open loop gain and compensate for the lower gain of the output stage.

was the intent of "best possible performance" really to get the lowest THD number possible?

That's what I thought :nod:

Be careful if you build it; sometimes those numerous gain stages conspire to make a wonderful oscillator. ;)

mlloyd1
 
Hi mlloyd1,
yes one of the designers aims was to get lowest THD, remember it was 86 and that was everything back then.
It is still however the best sounding loudspeaker amp i have built( to date), mosfet inputs & outputs, multiple stages, older semis and all.
I dont have a scope but it does not appear to oscillate though i admit my only test is what i percieve to be excessive heat or not & whether or not it seems to be operating okay in general.
As far as acoustic performance what of the designs that seem to appear on this forum with regularity would, in your opinion, better this design?
 
Three gain stages

The second gain stage appears to be to convert the signal from jfet diff pair to a push pull signal to pull the gates of the output mosfets too within a few volts of the rail. The second stage uses a lot of degeneration and appears to have a gain of about 7 for each of the complementary diff pairs. This stage has a very low gain and does the conversion of the signal and DC levels to swing the output mosfets close to the power supply rails. A pretty simple and linear gain stage for two diff pairs and two current sources. The final voltage gain stage also has emitter degeneration and a large enough resistor from the diff pairs to the rails to allow decent quiescent current stability. The final gain stage has 15 pF of base to collect compensation and the 10 ohm degeneration resistors and looks competently designed to me. I see no real basis for the assumption of high frequency stability or bias stability concerns at first look. Zip the schematic and repost post the file so I can be sure of the component values. This would be required to make any of the judgments expressed in the previous post with any degree of credible analysis. It is really not that complicated a design as one would assume at first glance and looks like a reasonable design to me. Gain stages with degeneration may be better than two high gain stages a higher and less linear gain. Three is much more to amplifier design than the number of gain stages. The design of each of the gain stages and sources is of much greater importance than something as simplistic as the number of gain stages.

Gene
 
Disabled Account
Joined 2003
PMM said:
It is still however the best sounding loudspeaker amp i have built( to date), mosfet inputs & outputs, multiple stages, older semis and all.

you said it sounded good? I thought in an earlier discussion on the Mousquito / Zenquito someone told that using FET in and FET out would sure produce horrible sound.

Does your amp sound harsh?
 
I also saw that mentioned re the fet inputs & outputs. i would have to say that i my amp does not sound harsh and this is compared to 6 other loudspeaker amp circuits i have made myself and also other commercially made products that friends, family and I have owned. (though, I do not consider myself an expert in this area)

To put the sound thing in perspective the only electronics magazine on sale in Australia currently published two circuits a couple of years back that were supposedly excellent in measured performance (yes THD was a big factor).
One was a 15 watt class A and the other a 100w class AB with seperate regulated supplies for the input and VAS. I built both and they lie around unused under my house. Complete dissapointments though i did use the class A as a headphone amp till I discovered Headwize. They probably do have exceptionally low THD etc etc but they rendered the music completely lifeless and took away all the excitement. I recently had a seconfd chance to confirm this as the class ab was presssed into service when we moved house and a temporay source of music (not!) was needed during the chaos.

I must be able to find a better sounding circuit surely after 15 years!?
 
last post was for millwood re the monet comment.

mikek,
I would have thought that the primary goal should be to produce an amp that sounds good. sure THD is a criteria but not everything?
The class a and ab amps that I mentioned in a recent reply had exceptional THD according to the publishers and i did construct them adhering rigidly to their specifications in the areas that were to affect THD and still they sounded lousy.
 
Disabled Account
Joined 2002
PMM said:
last post was for millwood re the monet comment.

mikek,
I would have thought that the primary goal should be to produce an amp that sounds good. sure THD is a criteria but not everything?
The class a and ab amps that I mentioned in a recent reply had exceptional THD according to the publishers and i did construct them adhering rigidly to their specifications in the areas that were to affect THD and still they sounded lousy.


if the published distortion figures are good and it doesn't 'sound' right, then it's not being measured right...
 
PMM,

I have heard good things about this Tillbrook amp too, but it is certainly complex. It builds on lessons learned in the ETI5000 design, which sounded horrible, but used similar concepts although the front end was different. In turn it was a refinement of the original Hitachi AN some thirty years ago.

Tillbrook is an interesting character, a flamboyant Professor of Mathematics at Uni of Sydney, IIRC. In the early eighties he built to a THD spec, then after listening to some of his creations he decided to try for a good sound as well. The result was your amp, as I understand it. I'm pretty sure it was his last magazine amplifier.

The 'dull and lifeless' comment you make is the heart of the matter. The Leo Simpson circuit you refer to, the 15W Class A in Silicon Chip, is a nice, polite amp but your phrase describes it well. It is neither a good nor a bad amplifier according to others who have built it here in Melbourne. This quality - I call it 'lifeforce' - comes back almost entirely to the stability regime used in the design. It's very hard to preserve 'lifeforce' AND stability, and this is where great amplifier designs really shine. Anything from Hiraga, Pass, or John Linsley Hood demonstrates immediately what I mean.

Interesting comments. Personally, and I differ here from Mikek, I believe the while THD is a significant measure of an amp's performance, but it is no judge of its sonic qualities. There are probably other measurements we could be taking which predict the sonics more accurately, but we don't know them yet.

Cheers,

Hugh
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.