John Curl's Blowtorch preamplifier

Status
Not open for further replies.
Jump to Latest
#18,082
Scott, let me assure everyone. You make one of the most linear and quiet IC's in the audio world. I am using it in a new design. I am also using an AD825 in another design. Both of these IC's you gave me as samples over the years. I am grateful.
I only want as much perfection and understanding that I can obtain, either from you, as an IC designer, or by other 'want to be' IC designers. That is what drives me forward, not giving you a hard time. Just look at the people who give me a hard time around here. Feel grateful that you do not incur their insults. :geezer:
 
#18,084
Re: Re: Re: Ouput Trannies

Wavebourn said:
Yes, but which one of negative connections should be more negative?

Also, it would be great to have pins for both capacitors. In such case a cathode follower, for example, could be used externally to drive own feedback loop in class A. 😉
Click to expand...

Or a control pin to set Iq. Even the old CA3094E has such pin.

edit:
>Yes, but which one of negative connections should be more negative?
Good question, depends on the application. What about one more pin connected to the substrate?
How many extra pins do we have already, four? 😉
 
#18,085
Re: Re: Ouput Trannies

Edmond Stuart said:



PS: How about the MOSFETs?
Click to expand...


Hello Edmond

The Mosfets have a confirmed delivery date of 6/5/09 next Wednesday, you might be in Crete by then. Where abouts are you going on the island.

With your simulated AD797 (MC9 schematic) , apart from the input device MAT02 what models did you use for the NPN , PNP devices . The reason I ask is because my simulation using NPN (KSC1845) and PNP (KSA992) gives me 2.2ppm THD @20Khz at 3Vrms gain of 10 .

Regards
Arthur
 
#18,088
john curl said:
[snip]I have hundreds of examples in a book by my side called: 'The Experts Speak'. However, this is not a universal truth, except for the fact that usually younger people come through with the most 'breakthrough' ideas before the age of 30. Look an Einstein, etc. I did my best ideas myself, and laughed with my friend of 40 years, just yesterday, when he asked what I was up to, and I said that I was building amps with the same topology that he and I developed for Alembic, in 1970! [snip]
Click to expand...

You are totally right John, and it is interesting if you compare this to the 'career' of distiguished artists. You will find that in science and engineering, as you noted, people have their best performnces below the age of, say, 30. But great artists, be they paintes, sculpters, writers, mostly have their best works when they are above, say, 50.

So what is it in our brains that deteriorates above 30 and that we would need for engineering and science, and what is it that improves above 30 or 40 and that we need for artistic, emotionally related stuff?
I find that question at least as interesting as capacitor DA 😉

Jan Didden
 
#18,089
John,
there are many good OPamps out there. The rather old AD797 is good but do not forget the new ones from National or LT.

BTW,
what SNR rel 0,5mV do you get with your AD797 MCRIAA amplifier?

The AD797 has a special place in my heart as it was with it and an extreme PS and a stepped attenuator that I started my high end audio life.



Sigurd

john curl said:
Scott, let me assure everyone. You make one of the most linear and quiet IC's in the audio world. I am using it in a new design. I am also using an AD825 in another design. Both of these IC's you gave me as samples over the years. I am grateful.
I only want as much perfection and understanding that I can obtain, either from you, as an IC designer, or by other 'want to be' IC designers. That is what drives me forward, not giving you a hard time. Just look at the people who give me a hard time around here. Feel grateful that you do not incur their insults. :geezer:
Click to expand...
 
#18,090
and the answer is...

janneman said:


You are totally right John, and it is interesting if you compare this to the 'career' of distiguished artists. You will find that in science and engineering, as you noted, people have their best performnces below the age of, say, 30. But great artists, be they paintes, sculpters, writers, mostly have their best works when they are above, say, 50.

So what is it in our brains that deteriorates above 30 and that we would need for engineering and science, and what is it that improves above 30 or 40 and that we need for artistic, emotionally related stuff?
I find that question at least as interesting as capacitor DA 😉

Jan Didden
Click to expand...

I believe the answer for engineering here is tolerance for risk... those younger years represent a period of presumed invincibility, the mileage accumulated over decades of insults, failures, misunderstandings, etc. i believe contribute to a more conservative and less innovative perspective on problem resolution and design goals.

As for artists, I'd guess they don't give a damn what anyone thinks after say 45 or 50 years of lack of recognition...

I don't think there is any deterioration after 30; just a gradual rationalization process towards a more thoughtful and conservative methodology.

I know I got my patents in my thirties... now I'm too old (58) and/or lazy to jump through all the hoops to pursue innovation for profit...😉

John L.
 
#18,092
Sigurd, if you can show me a better IC, I will consider it. The only real problem that I have with the AD797 is the .5ma Iq in the output stage. I had the same reservation with the LT1028.
Wavebourn, you are impractical. I only use IC's when I must make something compact and inexpensive. My discrete designs are open loop, transconductance amplifiers, except for power amps. You should know better than pre-judge.
 
#18,094
I would like to bring out, for others looking in, what all this is about.
First, you might think of an IC op amp as a transconductance amp or TA, combined with a unity gain follower of some kind. The 2 blocks make an OP AMP.
However, BY DEFINITION, the TA is HIGH out impedance. How high, 1meg, 100meg 1gig, 10gig ohms? Depends on the design, type of cap compensation and cascoding.
This is important, because with a follower made of BIPOLAR devices, the transfer function actually changes depending on output loading, and beta characteristics of the devices.
Most here look at a complementary Darlington follower, or some variation, as mostly effected by changes of Vbe of the output devices, BUT this is not true with very high drive impedance from the TA. Then, the follower actually transforms (mostly) to a BETA MULTIPLIER. Now, maybe, you can see why beta non-linearity is so important to me. Anything the beta does (with current), so goes the open loop transfer function, and subsequent overall linearity.
I found an extreme example of this, decades ago, with an HA909 IC op amp. I could actually IMPROVE the overall distortion by reducing the open loop gain by 40 dB. Try that, sometime!
 
#18,098
Well, not easy to get as low noise as the AD797, but when you mention the AD825, I think that one should consider a bunch of other OPamps that sonically will give the AD825 a match. OPA627B for ex.
Or AD8610, LME49710, LT1468, THS4031, AD8599 etc etc.

My point is more that there are lots of OPamps out there, and one has to try many of them in different applications to find what the best one is for each application.




Sigurd

john curl said:
Sigurd, if you can show me a better IC, I will consider it. The only real problem that I have with the AD797 is the .5ma Iq in the output stage. I had the same reservation with the LT1028.
Click to expand...
 
#18,099
john curl said:
I would like to bring out, for others looking in, what all this is about.
First, you might think of an IC op amp as a transconductance amp or TA, combined with a unity gain follower of some kind. The 2 blocks make an OP AMP.
However, BY DEFINITION, the TA is HIGH out impedance. How high, 1meg, 100meg 1gig, 10gig ohms? Depends on the design, type of cap compensation and cascoding.
This is important, because with a follower made of BIPOLAR devices, the transfer function actually changes depending on output loading, and beta characteristics of the devices.
Most here look at a complementary Darlington follower, or some variation, as mostly effected by changes of Vbe of the output devices, BUT this is not true with very high drive impedance from the TA. Then, the follower actually transforms (mostly) to a BETA MULTIPLIER. Now, maybe, you can see why beta non-linearity is so important to me. Anything the beta does (with current), so goes the open loop transfer function, and subsequent overall linearity.
I found an extreme example of this, decades ago, with an HA909 IC op amp. I could actually IMPROVE the overall distortion by reducing the open loop gain by 40 dB. Try that, sometime!
Click to expand...


All 3 blocks (differntial input, transconductance amplifier, current amplifier) form a device that is not optimal for audio. Beta dependences of output devices are multiplied. Dependences of beta in diamond buffers don't cancel each other, they inevitably increase an order of the resulting transfer function (where your 7'th harmonic is buried). Now, driven by current source they have input resistance variation even when loaded on a stable like a rock resistance; I don't even speak of non-linear and frequency dependent loads. Being interested in output voltage of the opamp, look what happens with output voltage of the transconductance amplifier loaded on non-linear resistance. Plus parasitic things like non-linear capacitances...

Is it optimal all together? No. Even when each of 2 blocks are optimal themselves, even though a Diamond Buffer is optimal to drive heavy load and save an electricity when idling. If efficiency criterium is out the diamond buffer have to be replaced by some different topology especially designed for class A.

Now, a differential stage. It may be a nice differential stage, with huge CMRR rail to rail on all frequencies, but is it needed in the whole amp that has low output resistance? No. You can always use lower feedback resistance from such an output to avoid flaws of a differential input.

AD797 is very well designed opamp. Should we blame on Scott for not designing what it is not, an audio building block? No. His opamp is a very good device, a peace of the art of an electronics design.
 
#18,100
john curl said:
Anything the beta does (with current), so goes the open loop transfer function, and subsequent overall linearity.
I found an extreme example of this, decades ago, with an HA909 IC op amp. I could actually IMPROVE the overall distortion by reducing the open loop gain by 40 dB. Try that, sometime!
Click to expand...

In the simplest case non-DC impedance is -i/(2*PI*f*C) where C is the comp cap. In this case even at 1kHz we are already 100X lower than the output loading. Beta is still doing nothing interesting at 1kHz either. The DC open-loop transfer function has little to do with anything. :soapbox: :smash:
 
Status
Not open for further replies.
Top Bottom