I made my first design as 2 gain block composite, with a transconductance amp or (trans-amp) as the first stage, then a standard discrete op amp as described here previously. The RIAA would be split, with the high frequency rolloff (2K approximately) in the first stage, and the 500, and 50 Hz inflection points in the second stage. I use this approach, even today, as it frees the second stage from any potential slew rate induced distortions.
Both gain blocks, at first, had global negative feedback. In fact, as an engineer, it was 'necessary' to reduce CCIF distortion that could creep in a post equalized gain stage.
Well, the unit WAS BETTER than the Levinson JC-2, subjectively. Still, it was not good enough for HK so it was never put into production. However, I took it one step further. I decided to remove the global negative feedback from the first stage, and only use the low noise differential jfet input stage, and just put the high frequency roll-off cap across the 2 drains of the input stage and finally just cap couple to the 2'nd gain block. This sounded even better! But WHY? And could I even improve it further? In 1980, I built my own personal preamp using this approach.
Both gain blocks, at first, had global negative feedback. In fact, as an engineer, it was 'necessary' to reduce CCIF distortion that could creep in a post equalized gain stage.
Well, the unit WAS BETTER than the Levinson JC-2, subjectively. Still, it was not good enough for HK so it was never put into production. However, I took it one step further. I decided to remove the global negative feedback from the first stage, and only use the low noise differential jfet input stage, and just put the high frequency roll-off cap across the 2 drains of the input stage and finally just cap couple to the 2'nd gain block. This sounded even better! But WHY? And could I even improve it further? In 1980, I built my own personal preamp using this approach.
By the time 1980 came around, I was asked by Peter Madnick, then president of Dennesen Electrostatics (sp) to build a new preamp that they would call the JC-80. I worked part time for about 1 year on this, utilizing the 2 stage design, removing any coupling caps with a folded cascode second stage, and a number of other 'improvements'. The JC-80 was a success, especially in Japan, where we got awards for it.
Years later, with Peter Madnick not with the company anymore, my royalties again fell off and Vendetta Research as born. I started with just a complementary folded cascode jfet input stage, like many designs described on this website, over the years, servoed it, and refined the layout, and called it the SCP-1. This little gain stage ran rings around my previous attempts to make an 'open loop' input gain stage, and I have stuck with it ever since. This is probably the most successful gain stage that I have ever made. For example, Dave Wilson still uses it. Later, I added a discrete second stage to make SCP-2, that many people are familiar with. More later.
Years later, with Peter Madnick not with the company anymore, my royalties again fell off and Vendetta Research as born. I started with just a complementary folded cascode jfet input stage, like many designs described on this website, over the years, servoed it, and refined the layout, and called it the SCP-1. This little gain stage ran rings around my previous attempts to make an 'open loop' input gain stage, and I have stuck with it ever since. This is probably the most successful gain stage that I have ever made. For example, Dave Wilson still uses it. Later, I added a discrete second stage to make SCP-2, that many people are familiar with. More later.
Now, if you have gotten this far, we as still back to 22 years ago. It is relatively to talk about the past, because it is less subtle than present changes to make a better phono stage, and first, we should go back and see WHY what I 'evolved' to over 15 years of solid state phono design really did work better.
To understand this we need to know more about the dynamic performance of moving coil phono cartridges. To get a visual 'handle' on this, we have to go back to my 1978 IEEE paper, "Omitted Factors in Audio Design' linked here, somewhere. To do this paper, I had to look at the REAL BANDWIDTH of a MC phono cartridge, especially when mistracking, which it does a good deal of the time. You will find, if you do this, with the right test equipment that requires transient storage and response at least to 1 Meg Hz, that phono cartridges can 'spit' garbage out to 500KHz, and 200KHz being typical. This is HARD on many feedback amplifiers. Why is this often overlooked? Well, the phono stage essentially INTEGRATES the input signal and tames it down, so to speak, but that does not imply that the INPUT STAGE is very happy doing it. Also, ticks and pops put out a tremendous transient that will virtually overload any input stage, so you had better be able to recover virtually instantly, or the tick or pop will be extended in duration and more audible. This is part of what many inexpensive phono playback systems suffer, and cause so much complaint. More later.
To understand this we need to know more about the dynamic performance of moving coil phono cartridges. To get a visual 'handle' on this, we have to go back to my 1978 IEEE paper, "Omitted Factors in Audio Design' linked here, somewhere. To do this paper, I had to look at the REAL BANDWIDTH of a MC phono cartridge, especially when mistracking, which it does a good deal of the time. You will find, if you do this, with the right test equipment that requires transient storage and response at least to 1 Meg Hz, that phono cartridges can 'spit' garbage out to 500KHz, and 200KHz being typical. This is HARD on many feedback amplifiers. Why is this often overlooked? Well, the phono stage essentially INTEGRATES the input signal and tames it down, so to speak, but that does not imply that the INPUT STAGE is very happy doing it. Also, ticks and pops put out a tremendous transient that will virtually overload any input stage, so you had better be able to recover virtually instantly, or the tick or pop will be extended in duration and more audible. This is part of what many inexpensive phono playback systems suffer, and cause so much complaint. More later.
NO
It WAS hard with early opamp stuff. And it is boring now to discuss uA741, 4580 and similar opamp pearls.
This is not about feedback. We are able to make feedback amps that handle MHz signals easily and have very low noise and very low distortion.
I know the paper you speak about. Anyway, it might be nice to specify amplitude of those 500kHz - 1MHz components, that you mention here above.
What PMA is implying is that garbage at 500KHz or so, is at such a low level, that it could not be important. However take a square wave at 5KHz or so with a 1us rise time or so. What kind of output at 500KHz does it have? Answer, -40dB from peak. What about a 500 Hz square wave? -60 dB from peak. etc, etc. The point is that it is an ILLUSION that it is not important, IF high frequency garbage builds a fast transient, which I also measured with real phono cartridges playing real records.
Last edited:
PMA John
Just to say that I am really enjoying this and please do carry on I think that both have a point.
I agree that the gain stages should reproduce the signal from the pick up but a 500KHz was never put on the record intentionally by the musicians and if even this was the case the engineer working on the lathe would do his best to get rid of it so not to risk to damage the lathe.
To get back to the Blow torch where are the diagrams and circuit description
Sorry to ask but I have just stated to look at he treads and this and the No 1 treads are a bit long.
I would really appreciate if same kindred spirit point me to the right post.
Pat have you tried LT1028 to me it sound much better than AD979 especially when it is warm (quite hot actually)
Hum this may be a good one
Why the LT1028 is much better hot (70C) ?
Is this measurable?
What is the difference?
Jfet running hot at larger bias sound better?
Al
Just to say that I am really enjoying this and please do carry on I think that both have a point.
I agree that the gain stages should reproduce the signal from the pick up but a 500KHz was never put on the record intentionally by the musicians and if even this was the case the engineer working on the lathe would do his best to get rid of it so not to risk to damage the lathe.
To get back to the Blow torch where are the diagrams and circuit description
Sorry to ask but I have just stated to look at he treads and this and the No 1 treads are a bit long.
I would really appreciate if same kindred spirit point me to the right post.
Pat have you tried LT1028 to me it sound much better than AD979 especially when it is warm (quite hot actually)
Hum this may be a good one
Why the LT1028 is much better hot (70C) ?
Is this measurable?
What is the difference?
Jfet running hot at larger bias sound better?
Al
Bks, what IS, IS. This garbage is being generated by phono cartridge mistracking all the time. That is WHY it has to be considered. Would you like to drive a car that almost fell apart when it hit a pothole? The roadbuilders did not make the pothole, but they are out there, and your automobile had better handle it well. Same with phono stages, they must handle mistracking cartridges. If they don't, sound quality will suffer, just like PMA seems to complain about. '-)
Last edited:
The schematic diagram as never published. There were some approximate simulations on part 1 of this thread.
Don't raly need to get in to cars do we ?
So we agree that the garbage is generated by the cartridge as garagge it shuld be?
A) ingnored
B) reproduced
Not shure of which one either of you will pick.
As I said I am wery late on this tread at start it was like I discovered a quite god pre
Apart from the wiring maybe.
Now I find that it is essentially useless
Whatsa bit bout experts do you need alphabet after your name to qualify?
Hi Joshua_G
Tanks for the info supose wil have to start reading the tread from the begining to make any sense of it.
still you saved me quite a bit of work going trough undrds of pages of B thig about.
Al
Ps I drive a VR4 manual would not this qualify me as expert?
So we agree that the garbage is generated by the cartridge as garagge it shuld be?
A) ingnored
B) reproduced
Not shure of which one either of you will pick.
As I said I am wery late on this tread at start it was like I discovered a quite god pre
Apart from the wiring maybe.
Now I find that it is essentially useless
Whatsa bit bout experts do you need alphabet after your name to qualify?
Hi Joshua_G
Tanks for the info supose wil have to start reading the tread from the begining to make any sense of it.
still you saved me quite a bit of work going trough undrds of pages of B thig about.
Al
Ps I drive a VR4 manual would not this qualify me as expert?
5us rise time of transient is NOT fast, and amplitudes are low. Comments on opamp integrators are pointless, in case we do not speak about time constants and measured signals exactly. The passive RC network will show the same integrating effect, a question of square repetition frequency and RC time constant value only. You are permanently misleading inexperienced readers. Those who understand circuit theory and are experienced in ICs have already given up this useless arguing.
It is quite usual, in instrumentation, to have sensors that measure derivative of signal. Then you need integrating amplifier to get proper output. Such sensor is Rogowski coil, e.g. Why do I speak about it? Because I produce integrating amplifiers for Rogowski coil sensors, with BW up to 2MHz. And yes, if you put square wave at the input of the integrating amp, the result is triangle. Just as it MUST be. This is not any opamp fault, this is just maths.
It is quite usual, in instrumentation, to have sensors that measure derivative of signal. Then you need integrating amplifier to get proper output. Such sensor is Rogowski coil, e.g. Why do I speak about it? Because I produce integrating amplifiers for Rogowski coil sensors, with BW up to 2MHz. And yes, if you put square wave at the input of the integrating amp, the result is triangle. Just as it MUST be. This is not any opamp fault, this is just maths.
- Status
- Not open for further replies.