The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Quad 8 .... whats the skinny? Dual-Channel Preamps
Old 25th August 2004
  #31
Gear Maniac
 
europa78's Avatar
 

Quote:
Originally posted by De chromium cob
Like Paul said, they're not very desirable, so you should let me take it off your hands.....heh
Indeed. We should kill this thread if we know what's good for us.
Old 27th August 2004
  #32
Gear Maniac
 

The following intrigued me, could you elaborate?

Quote:
Originally posted by ToneLux
...I never liked the 990 design, as it had too many internal compensation things to keep it stable.
By my count, the 990 has 5 components wholly devoted to high frequency stability. I'm not sure it's fair to count the emitter resistors and inductors used in the input diff amp, but even with them included there are only 9 "compensation things".

I've seen some schematics that purport to be of the 2520, and if they're anything to go by, that circuit appears to need 8 components to keep it stable. An RC network across the output of the first diff amp, a cap that effectively makes the HF signal path single-ended, an RC snubber across each driver transistor, and the usual compensation cap. Am I understanding the 2520 circuit correctly, or am I missing something?

Nitpicking of parts count aside, I'd be very interested to hear your thoughts on op-amp topology and compensation in general, and which design choices you believe translate most directly to good sound.
Old 27th August 2004
  #33
Gear Maniac
 
axtrak's Avatar
 

Re: Quad 8 .... whats the skinny?

Quote:
Originally posted by Wiggy Neve Slut
Hey there...

have been chatting to a cpuple of mates and they have started raving about the old Qaud 8 desks and rep/eq modules.

I have tried a few a while back and they often pop up here in Australia and look to be nice and all with their switched freqs etc... but can not remember what they sound like.

I know they have quite a 'rock and roll' 70's hertiage and Dean Jensen did a lot of deisgn work for them in their heyday. Apparently Steely Dan's frost album was cut on one on LA.. dont know where tho!

So.. opinions pls... i am loooking for new flavours other than my favourite neves....

PEACE
Wiggy
Hi,

i worked for 5 years in a studio with a quad eight coronado.

As i can remember, the sound was very coloured, but nice. The several channels sounded very diffrent from each other, as it was not completely recapped.

The Mic Pre´s were ok and the EQ´s sounded nice, but their Q was often to broad. Our´s was highly modified with switchable Q on one Channel, and as i can remember it had new custombuild opamps and there was a system 4 VCA Automation.

The busses were really noisy.

All in all, i think today, it´s a cool console to use with a DAW, because it has character, but due to it´s age these things need a lot of work. I wish that i´d have kept some channelstrips, but they sold it about two years ago...
Old 29th August 2004
  #34
Here for the gear
 

I received a cassette from the original manager of Electrodyne & Langevin which tells the history of the audio industry as it was born out of the motion picture industry. You would not believe how "connected" so many of the earlier companies were.

The companies born out of Electrodyne's parent company were Quad Eight, Sphere, Langevin, Rhichenback, Gauss, Saki, Optimation and Altec. Most original drawings and spec's came out of very early Western Electric drawings. The earliest Electrodyne consoles were tube based using the small tubes with the wires hanging out which were soldered directly in the signal path.

We have Electrodyne to thank for how consoles are layed out and designed today. They invented the modular approach to building console. They were the first to combine a mic pre-amp, line amp, eq section, aux routing and attenuation to a single Intergrated Circut (as they called it at the time), today we just call it a channel strip. They invented the active combinging network that would allow multiple channel strips to be combined (read subgrouped). They also made it so the console no longer had to be impedence dependant. You could combine balanced and unbalanced signals with many combinations of ins and outs and you didn't have to worry about impedence anymore, which was not true of the broadcast consoles of the time.

Their noise floor of -127dbm set the industry standard. There original mic pre's had 100db of gain! They needed this as they were dealing with very noisy stage sets. Think of the original Wells Fargo stage set where you'd have a team of horses, a stage coach, outdoor noise and actors sitting atop the coach and this one poor microphone (over the heads of the actors) trying to pic up all this dialog as well as the stage coach and other noises.

I could go on and on but I'll wait and give it all to you in the article (which keeps getting bigger and bigger).

BTW, The original op-amp was called the A1000, it was round and about the size of a dime. Electrodyne never did a formal schematic on this op-amp. I have one of the only schematics for this as it was hand drawn on the back of one of the manuals for the 709 input module. It was given to me by accident. Later this op-amp was redesined by John Hall to the square op-amp (about the same size as the API 2520) that was popular in so many of the early Electrodynes and Quad Eight consoles.

I think someone should do a documantary on this company and the early days of the audio industry before all these gentlemen are no longer with us. Most are in there late 70's and early 80's. Some have moved on to greener pastures, but I don't think anyone has ever told the story on how the audio industry was born. The original owner of Electrodyne is still with us as well as Langevin and a couple others. I should contact Alan Slutsky (Standing In The Shadows Of Motown) and talk to him about it. We owe all these guys many thanks as they invented the process which makes are jobs so much easier today.

Stay Tuned...

Larry DeVivo
www.silvertonemastering.com/stone
Old 29th August 2004
  #35
Lives for gear
 
loudist's Avatar
 

Larry, a very worthwhile endevour.
Old 30th August 2004
  #36
Here for the gear
 

qe

I purchased a QE pacifica some years back, and I love it. I've done quite a bit of work (read: troubleshooting) on it, but it keeps chuggin' along.

I also own a QE AM3b two channel compressor which is a very useful squish monster. Also, a pair of 712 graphics which are the best sounding graphic EQs I've ever heard.

In short, QE's rock. And more people are realizing it every day.

The Pacifica (circa 1978) is designed around the AM-10 op amp which most people seem to think is superior to the AM4 series amp found in the 2082 (circa 1972) series consoles. For a time, I owned much of the 2082 series console from the original MGM Grand hotel in Vegas. It was comprised of MM100 and MM62/71 strips as well as some 4x1 mixer modules. The MM series sounded very good to my ear. I have since parted out that console.

However, I liked QE gear so much I put up a website to help other QE fans connect. It can be found at Http://www.quadeight.net .

I'm looking forward to that article!

BTW Tape Op's Larry Crane is also a former QE owner.
Old 30th August 2004
  #37
Mr Addison,
Welcome.
I still haven't got any of my Q8 stuff up and running.
(I bought some of those parted out modules from Mark)
It will be the engine room of my new drum recording room though.
After reading this thread I absolutely can't wait.
Old 31st August 2004
  #38
Here for the gear
 

Quote:
Originally posted by Kent
Mr. Addison,

Any chance that you have the service (or any other documentation) pertaining to the AM3b? Ours is beginning to act strangely and it needs a visit to the bench.

Care to share?

Thanks,

Kent
Hey Kent, I wish I could help you, but I don't have any documentation. Try Ken Hirsch at Orphan Audio, he seems to have a lot of QE docs.

I do have some extra AM-4 cans around here somewhere if that'd help. (AM3b uses Am4s)

Let me know if you find any docs - I might be asking you for them!

Cheers,

Mark
Old 3rd September 2004
  #39
Gear Head
 

MM310

europa78, it sounds like we have the same box.

Unfortunately for me, mine is for sale.

Check the classifieds here.
Old 3rd September 2004
  #40
Gear Maniac
 
europa78's Avatar
 

That's them alright. If I didn't just buy some Telefunken V672's, a couple of 421's, a little Labs/Mercenary STD...

What can I sell this week?

How did you come across those? Any Idea what year?


__________________
Sincerely,

L. Cohen.
Old 3rd September 2004
  #41
Lives for gear
 
MJGreene Audio's Avatar
 

Mark,
I think I spoke with you a while back about my AM3b. It is up for sale if anyone has any interest in it. I haven't fired it up in about 10 years and I have no idea what operating shape it is in.

If anyone is interested please make me an offer,
Michael Greene

[email protected]
Old 3rd September 2004
  #42
Gear Head
 

Quote:
Originally posted by europa78
That's them alright. If I didn't just buy some Telefunken V672's, a couple of 421's, a little Labs/Mercenary STD...

What can I sell this week?

How did you come across those? Any Idea what year?


__________________
Sincerely,

L. Cohen.
I bought them at auction a few years back from an engineer in Chicago. I have no idea what "vintage" they are.
Old 9th September 2004
  #43
Lives for gear
 
De chromium cob's Avatar
 

SWWWEEEEEET

Lucky for us (well, ME to be more accurate) people haven't caught on to the value of these babys yet....I just picked up 4 more Q8 MM-61s with AM-3 op amps for $819! And I already have the power supply for them.... Sorry to gloat, but those who have heard these things know how stupid cheap that is.... Drinks on me! heh heh heh heh
Old 24th September 2004
  #44
Here for the gear
 
vintageguy's Avatar
 

Hi there.... QE pacifica strips for sale!

Have a bunch of QE Pacifica strips for sale...
Email for pics if interested....just cleaned and tested.
Just let me know gang.
Old 5th November 2004
  #45
Quote:
Originally posted by SeventhCircle
The following intrigued me, could you elaborate?



By my count, the 990 has 5 components wholly devoted to high frequency stability. I'm not sure it's fair to count the emitter resistors and inductors used in the input diff amp, but even with them included there are only 9 "compensation things".

I've seen some schematics that purport to be of the 2520, and if they're anything to go by, that circuit appears to need 8 components to keep it stable. An RC network across the output of the first diff amp, a cap that effectively makes the HF signal path single-ended, an RC snubber across each driver transistor, and the usual compensation cap. Am I understanding the 2520 circuit correctly, or am I missing something?

Nitpicking of parts count aside, I'd be very interested to hear your thoughts on op-amp topology and compensation in general, and which design choices you believe translate most directly to good sound.
Actually, having slept with the 2520, there are very few parts for stability. The schematic is most likely bogus. Lots of gain, but not a loose cannon like the 990.

My new discrete class A op-amp has only 6 transistors in all and is stable without a feedback cap. WHen you have too much gain, you are asking for termites. A clean simple design will translate in the end. Although the 2520 has a good bit of gain, I would include it with the API 2510 (which I did) mine, the Neve's class A's as stable, good sounding amps.
Old 10th December 2004
  #46
Lives for gear
 
Seamus TM's Avatar
 

Hey there,
I'm bring this thread back to life to see if any of you Q8 guys know anything about the Virtuoso console?

Seamus.
Old 12th December 2004
  #47
Lives for gear
I want to know more about this alleged working Olive!

I too (SOMEWHERE) have a rather cool Olive handout that was printed by Westlake Audio in about 1972.
They were enourmously innovative (the first mention I ever saw of VCA groups, automation, a four band EQ with linear horizontal swiyches and a compressor/gate on every channel).
they predated SSL (which incorporates all of thsoe ideas) by what? 6 years? more?
But althogh I THINK Andre Perry had one in his Montreal studio (not Morin Heights), my impression was always that none ever really WORKED.

if that's NOT true I'd love to know about it and how it sounds. (maybe unlike SSL, it soudned good!)

by the way Paul (and hi) you wouldn't rate the Trident A Range discrete class-A amp as "stable"?... it's certainly terrific sounding, maybe the best of all.
Yes, "a good designer can mix discrete and IC" parts in a circuit (as long as one doesn't ever run music through it) <g>



oh, and one of you Quad Eight fans should buy my Noise Gates (in the classifed section)!
Old 12th December 2004
  #48
Gear Addict
 
BrianK's Avatar
 

I have a Pacifica from 1977. I an quite familiar with it (of course) and would say it's best feature is an open and clear sound, I would say more musical than Never or API.

API definitely had more punch, and Neve definitely has more range of sound than Q8 or API. You can't change a sound much going through a Q8 - more EQ power on a Neve, but I don't need that for my work.

The typical comment that is it "somewhere in between API and Neve" are fairly accurate I think (tastes taken into consideration). I am quite curious about the Tape Op article... more (in general) need to be explored about the history of recording gear before it gets lost...
Old 12th December 2004
  #49
William Wittman
Producer/Engineer
(Cyndi Lauper, Joan Osborne, The Fixx, The Outfield...)

Don't forget Too Much Joy...


And a damn good Producer and Engineer to boot...
Old 12th December 2004
  #50
Gear Maniac
 
BJohnston's Avatar
 

Quad 8 FS?

I'm not sure what the 8068 is worth, but I'm in the market for a board. If you can get an idea of what it's worth let me know. I could be interested.

Thanks,
BJohnston
[email protected]
Old 12th December 2004
  #51
Gear Nut
 

Olive Console

Quote:
Originally posted by wwittman
I want to know more about this alleged working Olive!

I too (SOMEWHERE) have a rather cool Olive handout that was printed by Westlake Audio in about 1972.
They were enourmously innovative (the first mention I ever saw of VCA groups, automation, a four band EQ with linear horizontal swiyches and a compressor/gate on every channel).
they predated SSL (which incorporates all of thsoe ideas) by what? 6 years? more?
But althogh I THINK Andre Perry had one in his Montreal studio (not Morin Heights), my impression was always that none ever really WORKED.

if that's NOT true I'd love to know about it and how it sounds. (maybe unlike SSL, it soudned good!)

by the way Paul (and hi) you wouldn't rate the Trident A Range discrete class-A amp as "stable"?... it's certainly terrific sounding, maybe the best of all.
Yes, "a good designer can mix discrete and IC" parts in a circuit (as long as one doesn't ever run music through it) <g>

oh, and one of you Quad Eight fans should buy my Noise Gates (in the classifed section)!
Weren't they a Canadian company out of Montreal?
World's first automated board (sort of - when it worked)
had a protoype at AES show in early 70's

I recall seeing one ages ago - under wraps, not installed, owned by

Brian Mitchell
Studio 306
Toronto, Canada

I could be mistaken, but I think Steve Hasleton (ex TML tech) might know something about them
Old 12th December 2004
  #52
Gear Maniac
 

Paul Wolff said:
Quote:
Lots of gain, but not a loose cannon like the 990.
Paul: Could you explain what you mean by this?

John Hardy
The John Hardy Co.
www.johnhardyco.com
Old 12th December 2004
  #53
Quote:
Originally posted by John Hardy
Paul Wolff said:


quote:
--------------------------------------------------------------------------------
Lots of gain, but not a loose cannon like the 990.
--------------------------------------------------------------------------------


Paul: Could you explain what you mean by this?

John Hardy
The John Hardy Co.
www.johnhardyco.com
Why sure...

Don't get me wrong, I'm not implying that the 990 isn't a great op-amp and doesn't have a great reputation. It is also not a discussion of what op-amp is better, or better sounding, as I have always been the one that doesn't focus on the specs of a circuit if it sounds great. I am also one that keeps it simple.

But, technically speaking, 990 has to much gain, and although it helps to keep the distortion down low, it requires chokes on the input stage to keep the gain lower at high frequencies because the internal individual stages have such a wide difference in gain bandwidth that the input stage has to be slowed down because it will oscillate if the compensation is not there, because there is no way for the higher bandwidth to make it through the rest of the amp to get fed back and be brought under control. This is also required because of the higher junction capacitance of the 100 or so paralleled transistors in the LM394 cause a higher degree of phase shift, which has to be delt with. The LM is good for matching, it helps a little with the distortion, and helps a ton with the offset at higher gains, which is interesting, because if you servo an op-amp, matching the input stage to that degree isn't a huge requirement. Proof of this is the GML op-amp, which sounds great but should be servoed as does the API 2510.

Both the API 2510 and my new TX-240 have similar distortion specs, but have a slightly more basic design, which not only lends itself to stability, but at the same time, low distortion.

The TX-240 can run with 40dB of gain WITHOUT an external compensation cap because the internal stages are so well alike in bandwidth. The first open loop (full gain, no feedback resistor) pole of the TX-240 is around 8Khz, where the 990 is around 200Hz (if my old mind serves me right, I remember doing this test many years ago). The main reason for this is that the 990 has more gain, which effects the overall phase margin. But this also means that you have to be more critical with the compensation of the 990 to keep it happy. (990 has around 125dB open loop gain, the TX-240 has around 75dB)

The 990 is a fine op-amp, but it is like driving a car with the gas to the floor and the brakes to the floor. I prefer a lighter body and less horspower.

So no slams here, just the facts, mam.
Old 12th December 2004
  #54
Quote:
Originally posted by SeventhCircle
The following intrigued me, could you elaborate?
I've seen some schematics that purport to be of the 2520, and if they're anything to go by, that circuit appears to need 8 components to keep it stable. An RC network across the output of the first diff amp, a cap that effectively makes the HF signal path single-ended, an RC snubber across each driver transistor, and the usual compensation cap. Am I understanding the 2520 circuit correctly, or am I missing something?
All I can say is that is not the 2520.
Old 12th December 2004
  #55
Quote:
Originally posted by wwittman
I want to know more about this alleged working Olive!



by the way Paul (and hi) you wouldn't rate the Trident A Range discrete class-A amp as "stable"?... it's certainly terrific sounding, maybe the best of all.
Yes, "a good designer can mix discrete and IC" parts in a circuit

(as long as one doesn't ever run music through it) <g>
The Trident class A stuff was very similar to the Neve stuff and was very nice.

There is a way that a designer can mix ICs and Discretes if he knows what they are doing. The critical parts are the Mic Pres, parts of the EQ and Summing. As long as the rest is at line level, you can make a great sounding product and save the customer money at the same time. It's all about tone.

There are many applications where it can be done with success, but you have to select the right IC for the right job. No one complains about the Distressor... Dave uses several different types for different uses. Most just throw in a bunch of 5532s and feel bulky.
Old 13th December 2004
  #56
Gear Head
 

Quote:
Originally posted by Mr. Fancy Pans

Deane Jensen inhertied (if you will) his most famous op-amp design, but that's another story and one I will probably never be able to tell.

Larry DeVivo
www.silvertonemastering.com/stone
From Steve Hogan,
formerly of Jensen Transformers

For Mr. DeVivo to suggest that Deane Jensen's work on the 990 for which he was granted Patent #4,287,479 was not original is rediculous.

Deane was chief engineer of Quad Eight for a time in the early '70s. He hired Bill Whitlock (now president of Jensen Transformers) to replace him as chief engineer at Quad Eight in 1972 because he was starting Jensen Transformers.

Prior to designing the 990 Deane designed an amplifier for Quad Eight that had its transistor count severely reduced by upper management in order to cut costs. While at Jensen he designed the 918 amplifier for Pacific Recorders in San Diego. When the LM-394 supermatch pair came out, he began working on the 990. The 990 took months of measurements, studying of textbooks, and computer modelling which is old hat today but in the mid 1970's was pioneering work --work that was not being done in Audio circuit design for sure.

The 990's internal architecture and compensation was very carefully worked out to not only have lower noise and higher open-loop gain at audio frequencies than previous opamps, but it was unity gain stable as well. I made one of the first 990's in the 1970's in an octal can to replace some Opamp Labs modules. Deane graciously helped me get it working when I accidentally left a trace off my PCB. I saw at that time, and may times thereafter during the years I worked at Jensen the original 990 development breadboard which was about 6 inches square.

The improvements to the DC characteristics of the 990 which I made later (990-B and 990-C) made no changes at all to the open loop gain and phase shaping that make this amp work so well and sound so good but made it work better in DC coupled circuits which were not being used in the 1970's when the amp was originally designed.

I will address in a later post the erroneous comments made by Paul Wolff re the 990 circuit design, which he clearly does not yet understand. For those who want to understand the reason for every part in the 990 circuit, Deane's original article in the AES journal is still available and a phone call to John Hardy or Jensen Transformers will bring you a copy in the mail to read.

The 990 has stood the test of time, and is still the heart of some of the very best-sounding customized mixing consoles (including a Quad Eight console I modified this summer) because of the innovative and thorough engineering done by Deane Jensen. To suggest that he "inherited it" from someone else is pure rubbish!

More later ---
Old 13th December 2004
  #57
See, you can't say anything without everyone getting a boner about it. I actually do understand it, and never, as I said twice, was trying to dis it as a good op-amp. I also said that I prefer lower gain op-amps and that the 990 had too much, which I still feel is too much. Again, I didn't say it was wrong either. So, take a step back, have a beer and understand that I am voicing my opinion about how I like my op-amps, and like I said before, the 990 has proven itself as a wonderful op-amp, but again, not my flavor. I know the history of it, as well as where the transformers came from too. I also didn't say anything about inheriting it from someone else either. Sounds like a mix of two messages. As far as the article about the 990, I have a copy, and also have a copy of the article that Peter Barthelson wrote that addresses the same things that I discussed. I also never said anything against Dean, in fact he and I had many many many great phone discussions about this subject and we had a very positive mutual respect for each other's design work, unlike this reply.

So instead of a pissing match, just chill a bit and don't turn this into an insult contest, because we all know that I am second only to Fletcher when I get wet legs.

Try to keep your opinions your opinions and keep the insults to yourself, or I will start in Baltimore and work my way west. There will be a lot of juicy stories then.

By the way, the patent says exactly what I said, which is the input stage needs chokes to tame the high frequencies, as it is stated from the USPTO:

An operational amplifier circuit having a first amplification stage that includes a differential pair of matches, low-noise transistors coupled in a common-emitter configuration, with a special reactive network in the emitter circuit of each transistor. Each reactive network comprises an inductor and a resistor, arranged in a shunt configuration, whereby at low frequencies the first stage has a relatively high gain and generates relatively low noise, and whereby at high frequencies the first stage has a relatively low gain and the amplifier circuit has sufficient phase margin to ensure stability.

And, the patent has to do with ONE aspect of the opamp, not the entire design of the opamp. It was a remarkable idea to do that and at the same time be able to tame the high gain aspect of the op-amp.

As much as we all want to say that we invented all of our own op-amp designs, we all know that we look and learn as much as we can from others and then take a step back and try to come up with something special, in line with our concepts of what we think sound should be, while trying to fix a common problem that plagues others. I am sure that Dean had a starting point with the 990, just like the rest of us.

It's just like the transformers. Dean did remarkable things with transformers, and with the Comtran design software, took them to lengths that no one had ever taken them, just like Theile figured out the missing variable when designing speaker cabinets. He didn't invent the speaker cabinet, be he sure made them a hell of a lot better, just as Dean didn't invent the transformer either.

God rest his soul.

Take a chill pill. You seem very excitable, defensive and somewhat insulting.
Old 13th December 2004
  #58
Gear Maniac
 

Paul;

It is the emitter resistors that "tame the high frequencies". It is the emitter inductors that reduce the noise voltage.

John Hardy
The John Hardy Co.
www.johnhardyco.com
Old 13th December 2004
  #59
Gear Head
 

Quote:
Originally posted by ToneLux
I also didn't say anything about inheriting it from someone else either. Sounds like a mix of two messages.

So instead of a pissing match, just chill a bit and don't turn this into an insult contest, because we all know that I am second only to Fletcher when I get wet legs.

Try to keep your opinions your opinions and keep the insults to yourself, or I will start in Baltimore and work my way west. There will be a lot of juicy stories then.

Take a chill pill. You seem very excitable, defensive and somewhat insulting.
Paul,

I joined this group today and began at the beginning of this thread on Quad Eight equipment that wandered into some history and then opamp discussions. I try to be very careful with my choice of words, and it would not be practical to address all the issues I feel need to be addressed in this entire thread in one message.

I thought that it was clear from the quote that began my post that I intended to set the record straight regarding Mr. DeVivo's assertion that Deane "inherited" the 990. I never suggested that you said such a thing, nor did I suggest that you ever said anything good or bad about Deane.

In your previous post you made some comments on your understanding of how the 990 works which IMHO are not correct.
I am used to forums where scientific discussions take place without a lot of personal attacks or insults. Your interpretation of the Deane's patent language that "the 990 needs chokes to tame the high frequencies" is misleading and not accurate.
Similar comments about the 990 circuit details you made in your previous post are also misleading and not accurate. I therefore commented that you do not yet understand the circuit. I apologize if you found that insulting. There are a lot of circuits that I don't understand, too.. I just hope to learn more as I go.

The chokes in the 990 are not required for stability. They do not "tame the high frequencies". It is actually the job of the 30 Ohm resistors in the emitter circuit to "tame the high frequencies" and they are part of the amplifier's high frequency compensation. They degenerate (reduce the gain) of the first stage at all frequencies.
The chokes in parallel with these 30 Ohm resistors "short out" the 30 Ohm resistors in the audio band, which increases the gain at audio frequencies and lowers the noise that would otherwise be caused by the 30 Ohm resistors. The brilliant trick in the 990 is that the chokes keep the noise low and the gain high in the audio band, but still let the 30 Ohm resistors work to help keep the 990 stable at high frequencies.

In the interest of me learning more, I am interested in the paper by Peter Barthelson that you mention. I couldn't find it in a google search and I would appreciate a link or information about where I can get a copy.

I really don't feel I have anything to be defensive about and I certainly have no desire to be offensive either.

Best wishes,
Old 13th December 2004
  #60
You're parsing words. The chokes are there to tame the high frequencies, but, yes, they are out of the circuit with the high frequencies because they are chokes. But the reality is that they, meaning the whole collection of parts, is there to tame the high frequencies.

I am stating this in laymans terms, so others will be able to understand more about opamp design, since this is a public forum, so please don't point your finger and try to find fault. I can very easily discuss this in poles and zeros, bode plots and degeneration, but it won't get me laid.

The problem with any opamp is the bandwidth of each stage. It just so happens that the input differential stage is the fastest, then the gain stage, then the slow output stage. (This can also be different)

This creates a problem where the overall open loop point of natural high frequency roll off is not the same from stage to stage. This is a problem because when the output is fed back to the minus input, and the point where a certain stage rolls off naturaly and crosses the unity gain point and begins to attenuate, sending it in to the minus input basically inverts this effect and the attenuation now has been inverted and is now actually gain, because when it is at any level above unity, it reduces the gain of the amp by going into the minus input. When it begins to roll off, the gain goes up because it is not there to make it go down.

The result of this is that (basically) the amp reaches a point where it can oscillate at very high frequencies because the slower stages are not allowing the highest frequencies to be fed back to reduce the gain of the high frequencies in the faster stages, which allows hem to go off and do anything they want, from ringing to oscillations. Since you can't hear this, the effect is like having no headroom and the amp sounds "pinched" because the headroom is being used up by an oscillation and the audio is basically riding on top of it.

To fix the differences in the stages, you attemp to compensate each stage on it's own to make them all match so this no longer becomes a problem.

As the open loop gain goes up, the internal stages change because of impedance changes, thus changing the individual stages bandwidths. The output stages always stay the same, because they have no gain. The other stages change depending on the design. Some will take a capacitor and resistor between the collectors of the input differential stage to lower the gain at higher frequencies. Dean used the chokes in the emitter side to do the same thing.

The compensation in the middle stage serves two purposes. It compensates the stage, but is also helps with external loads that can cause problems when they change the phase of the higher frequencies and then get fed back, leaving the amp to have another excuse to oscillate.

All of this has to be done together so the amp is happy.

Now, I have found in my opamp designing that these problems get worse when the amp has higher gain. I took a hard look at the use of the amps, and since 40 dB was the general maximum that people use, I increased the gain to a point where there was still enough feedback available to keep the distortion down.

By doing this, I have a design that has no input stage compensation, and only one cap internally to keep it stable. It is unity gain stable, and the distortion is always below .01%, typically below .005%, but can operate WITHOUT a feedback capacitor all the way through 40 dB of gain.

So, not to sound like a broken record, I prefer opamps with lower gain because they are naturally more stable, and that is why I said the 990 is like a lose cannon, because it has so much gain.

You are taking this as an insult to the design, which I am not saying. I pointed out certain things in the design that show the need to compensate stages because of the high gain. My design theories try to boil things down to the most simple parts because I feel that the more simple it is, the better sounding it is.

The bottom line is that there have been millions of gold records cut on the 990, the API 2520, the API 2510, and when the Tonelux stuff gets in place, it too will have it's success stories. But, at the same time, there have been millions of hits cut on SSLs, Quad 8's nad even Mackeys, so it really doesn't matter.

I guess it must be the song. ****. I thought is was all about me.
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
JKL / So much gear, so little time
9
wshaw / Remote Possibilities in Acoustic Music and Location Recording
27
Herman Munster / Other DAWs
7
Mastering101 / So much gear, so little time
7

Forum Jump
Forum Jump