Sigurd, it is easy with your noise graph to see why it is difficult to get really low noise from jfets below 100 Hz or so. It is not impossible, but the 1/f corner is usually higher than a good bipolar. Add RIAA eq, and it can become a measurement nightmare.
However, if you lower the operating current and temperature of quality Toshiba fets and put them in an isothermal environment, some Toshiba devices can be among the quietest in the world today down to 1/1000 Hz. I have an associate who makes this happen for a living and sells his creations to Stanford U (for example), and all over the world.
Zinsula, your design is interesting. Can anyone make a bipolar version of Zinsula's design? (Then it can be integrated)
However, if you lower the operating current and temperature of quality Toshiba fets and put them in an isothermal environment, some Toshiba devices can be among the quietest in the world today down to 1/1000 Hz. I have an associate who makes this happen for a living and sells his creations to Stanford U (for example), and all over the world.
Zinsula, your design is interesting. Can anyone make a bipolar version of Zinsula's design? (Then it can be integrated)
Scott - yes, it is a simple circuit and nothing to write home about, but that was not the purpose either. I wanted to show how many
JFETs would be needed to make a design with approx 0.25 nV/sqrt(Hz), and also to show that SPICE can be very accurate and useful for doing noise analysis.
The design is neither DC-coupled nor complementary, so it is not my style actually.
The BF862 is actually an RF-JFET and not made for audio at all,
but it can be used for our applications, too. It's low capacitance is extraordinary.
It has a "brother" in the BF861A/B/C which does not have as low noise as the BF861 but has other merits instead.
Price for these JFETs is low, very low.
Attached is a freq resp graph up to 10Ghz.
I am sorry to read that you are leaving the Blowtorch thread. It has been great having you around!
Not many here have the experience and design skills that you do.
I read that you will not post any of your audio designs here, but why not open up a thread of your own?
Regards,
Sigurd
JFETs would be needed to make a design with approx 0.25 nV/sqrt(Hz), and also to show that SPICE can be very accurate and useful for doing noise analysis.
The design is neither DC-coupled nor complementary, so it is not my style actually.
The BF862 is actually an RF-JFET and not made for audio at all,
but it can be used for our applications, too. It's low capacitance is extraordinary.
It has a "brother" in the BF861A/B/C which does not have as low noise as the BF861 but has other merits instead.
Price for these JFETs is low, very low.
Attached is a freq resp graph up to 10Ghz.
I am sorry to read that you are leaving the Blowtorch thread. It has been great having you around!
Not many here have the experience and design skills that you do.
I read that you will not post any of your audio designs here, but why not open up a thread of your own?
Regards,
Sigurd
scott wurcer said:
Attachments
Sigurd, while I applaud your effort, please understand that it is 40 years old. Even the parts that you chose were essentially available at the time, as the 2n4416. Ampex used to buy them for video cameras with preselected 100Hz noise maximums and we used to parallel them, just like you, in 1968, and we then realized that this was possible and almost practical. I suspect that you will find REAL PARTS slightly worse than the parts that you have modeled. Just a word of warning, due to long experience.
John,
yes, I know. It is frustrating for me not to have more skills to get a low noise amp without parallelling loads of JFETs.
With today's parts,
is there ANY other way to get this low noise?
While I am at it, I attached the en for BF861 vs 2Sk170 vs 2SK147. Observe the log Y-scale.
I do miss the 2Sk147....
Sigurd
yes, I know. It is frustrating for me not to have more skills to get a low noise amp without parallelling loads of JFETs.
With today's parts,
is there ANY other way to get this low noise?
While I am at it, I attached the en for BF861 vs 2Sk170 vs 2SK147. Observe the log Y-scale.
I do miss the 2Sk147....
Sigurd
john curl said:
Attachments
Scott's hobby circuits put forth here were impractical and thoughtless. WHERE is anyone going to get the 2SK146's and the 2SJ73 devices he cited? His circuit topology was upside down as well, leading to extra nonlinear input capacitance with the same number of parts used. This is why I asked him to rethink his input, here. Sorry, if that is out of line. I would just like to see this one thread give accurate and meaningful information, on a consistent basis to audiophiles and designers alike.
John - I agree with your recommendation.
People might want to know what grades of the SK170 to buy.
And for what usage each grade can be used for?
I have a mix of 25% GR, 50% BL, and 25% V.
Sigurd
People might want to know what grades of the SK170 to buy.
And for what usage each grade can be used for?
I have a mix of 25% GR, 50% BL, and 25% V.
Sigurd
john curl said:NOW THAT I HAVE YOUR ATTENTION! Buy all the 2sk170's that you can, while they still make them. They are VERY GOOD PARTS. Cheap too! Do it now, don't wait, and have to cry later. Your 'pitiful cries' of "feed me fets" will not work in future. You will pay big time for them.
I think that it is an honest appraisal of a circuit and it is the most direct and honest answer that I could give. I realize that many here think that criticizing me and my circuit design is OK for some reason, I don't know why, but I seem to have to take it without personal regard. Let's only hope for a level playing field. I'm willing to play fair, is everyone else up to it?
Mike, the last guy who criticized me about this, previously said that I didn't know anything about low noise design. Was this accurate, was this fair, was this in the spirit of cooperation and respect?
I quote from post #5925: "I know a former visiting professor of electron device physics and modeling at Stanford and TAMU universities occasionally reading and posting on this form, that based on your contributions here, would not give you a pass mark."
Now what would you call this? Constructive criticism? If I knew the professor, I might be able to answer back, but I don't know who it is. Read the entire post for yourself and please tell me what is fair and equitable?
I quote from post #5925: "I know a former visiting professor of electron device physics and modeling at Stanford and TAMU universities occasionally reading and posting on this form, that based on your contributions here, would not give you a pass mark."
Now what would you call this? Constructive criticism? If I knew the professor, I might be able to answer back, but I don't know who it is. Read the entire post for yourself and please tell me what is fair and equitable?
- Status
- Not open for further replies.