The problem with SYN08's design is not solved. It may be unimportant, but you would have to prove that to me. Of course, Scott thinks that his 797 can drive anything, but I am more careful with what I design than to believe that. That is why I decided to offset his output stage with a current source, even though my nominal low frequency load is 1600 ohms and my high frequency worst case load is 800 ohms. This a a far cry from 36 ohms at low frequency and 3 ohms or so at high frequency worst case. My gain is 30, not 70, so I will operate the 797 at an even easier voltage swing than SYN08. This is reality, fellow engineers, and why I get awards for good design of audio equipment.
Maybe "the Agitator" can name a preamp.
Good Night Sirs! sweet dreams, I have to make money, another day another dollar!
Kamskoma
Good Night Sirs! sweet dreams, I have to make money, another day another dollar!
Kamskoma
scott wurcer said:
32?
The results of that will be interesting. I wonder if SYN08 is currently soldering away............Just don't use that AD specified 100 ohm feedback resistor!
Cheers.
Glen
Paralleling Toshiba has been done in a big way over the years. Usually, for low noise instrumentation made by the Italians, for some reason. There are papers in the 'IEEE Transactions in Instrumentation and Measurement' perhaps 5 years ago that go into it in some depth.
Sometimes I wonder why you guys are taken the liberties to put unpleasant comments in this thread. I don’t think that a couple of circuits built by yourselves after reading Horowitz and Hill gave you the right for personal attacks and for assessment ability or disability of someone else. Have an exaggerated sense of self-importance?
Attachments
Let us sum up where we are. A complementary paralleled Toshiba j-fet circuit has been put forth. It IS quiet. It is even slightly quieter than the SCP-2 phono stage. However it is based on the same approach that was originally used in the SCP-2 input stage, except 1 ohm resistors are in series with the paralleled sources rather than 5 ohms. This gives a theoretical 2 ohm advantage to the newer design. That is the only noise difference between the two input stages. This level of performance is exceptional, but not unheard of. Doug Self did 0.2 nV/rt Hz of so, by paralleling a lot of low noise bipolar transistors once, years ago.
The potential problem seems to be in the loading of the IC op amp. This appears to be a genuine oversight, as almost 2uF is the nominal load. This is difficult to drive, even with a power amp. The SCP-2 uses 0.1uF to do the same job, or a 20 times easier load. This 'could' make a difference.
It might be possible to rearrange the feedback and/or add a high current buffer in series with the IC op amp to improve the situation. However, it makes little or no difference to me, in any case.
Next?
The potential problem seems to be in the loading of the IC op amp. This appears to be a genuine oversight, as almost 2uF is the nominal load. This is difficult to drive, even with a power amp. The SCP-2 uses 0.1uF to do the same job, or a 20 times easier load. This 'could' make a difference.
It might be possible to rearrange the feedback and/or add a high current buffer in series with the IC op amp to improve the situation. However, it makes little or no difference to me, in any case.
Next?
john curl said:
My computer simulation shows that everything is wrong with this pre-amp . Lots of noise from the 330R resistors in the I-mirrors and much of distortion. Wrong schematic as usual?
Another day, another silly question.
Could someone explain to me the local feedback network around U239 that is made from R5, R6, C9, and C10?
Could someone explain to me the local feedback network around U239 that is made from R5, R6, C9, and C10?
CG said:
Lag compensation. Don't know about the 797 but the alternative opamp - ths4031 is not stable for gains below 2.
john curl said:
Philips did a .2nV simple op-amp for the Ortophon MC cartridge in the 80's for an AES paper. It never got released, I still have two samples somewhere. The rbb limit on noise for a bipolar and paralleling transistors is not a profound insight you dwell on it too much.
It was, 40+ years ago, when I discovered low Rbb' in the 2N4403 and 2N4405 transistors.
It was the breakthrough that I was looking for, in order to make an effective pre-preamp for the MC cartridge.
It was the breakthrough that I was looking for, in order to make an effective pre-preamp for the MC cartridge.
john curl said:
If you're claiming to have discovered rbb and have not been credited, you certainly deserve an important mention somewhere in the history of analog design. I honestly mean that. OTOH it would be nice here to move on past talking about 40 years ago.
Scott, I did NOT invent Rbb'. I just found LOW Rbb' in devices that did not note it on the spec sheet. I have found several other devices from various manufacturers over the decades that are even better, without ANYTHING about it on the engineers spec sheet.
When the Levinson JC-1 was first introduced at the AES in 1973, I was told that it was impossible to achieve such low numbers as 0.4nV/rt Hz without a transformer. Well, I did it then, and I do it now, AND that is 2.5 times quieter than your AD797. You have a ways to go, Scott.
When the Levinson JC-1 was first introduced at the AES in 1973, I was told that it was impossible to achieve such low numbers as 0.4nV/rt Hz without a transformer. Well, I did it then, and I do it now, AND that is 2.5 times quieter than your AD797. You have a ways to go, Scott.
- Status
- Not open for further replies.