Nice to have you back, John.
Totally off topic, yesterday I had my carpets cleaned, and
afterword I was talking to the guy who turns out to be a rabid
audiphile. He worked for a few years at the local laser factory
with a bunch of the people who had been at Threshold, and
even a couple who dated back to ESS days.
Years before that he worked at Ampex. "Yeah", sez he, "there
was a brilliant audio guy there who worked on tape deck
electronics who also did some work for the Grateful Dead. I
can't quite remember his name......"
😎
Back on topic: Other performance issues aside, it's my
experience that JFETs are quieter when there's RF in the
neighborhood.
Totally off topic, yesterday I had my carpets cleaned, and
afterword I was talking to the guy who turns out to be a rabid
audiphile. He worked for a few years at the local laser factory
with a bunch of the people who had been at Threshold, and
even a couple who dated back to ESS days.
Years before that he worked at Ampex. "Yeah", sez he, "there
was a brilliant audio guy there who worked on tape deck
electronics who also did some work for the Grateful Dead. I
can't quite remember his name......"
😎
Back on topic: Other performance issues aside, it's my
experience that JFETs are quieter when there's RF in the
neighborhood.
Nelson Pass said:
Hi Nelson,
Yes, Ampex had some brilliant people back then. Another one was Ray Dolby. I wonder if John ever ran into him there?
I think JFETs are just naturally less likely to rectify RFI, since one is looking into an effectively reverse-biased PN junction with a JFET, as opposed to a forward-biased PN junction with a BJT.
Bob
Nelson, that could well have been my friend Ron Wickersham, inventer (patent holder) of the PZM mike and a subwoofer eq with double integration. And then, it could have been me.
Ron worked in the Video department and was hired away by the Grateful Dead in 1969. In 1970, Ron hired me to work on GD projects under the company named Alembic. The Wall of Sound was partially designed by Ron and me, but Bear (Owsley) deserves the most credit.
Ron worked in the Video department and was hired away by the Grateful Dead in 1969. In 1970, Ron hired me to work on GD projects under the company named Alembic. The Wall of Sound was partially designed by Ron and me, but Bear (Owsley) deserves the most credit.
john curl said:
Good day Mr Curl,
First, do you really mean that the laws of physics, discovered some 35 years ago, are no longer valid?
Second, it was far from clear which of the two designs of Glen you was referring to. The 500W version, published in part on this forum with unspecified input circuitry around the LTP's, or the 12W version, which is published on his own website?
Indeed, the latter has unnecessary high resistors at the inputs, which should be lowered by a factor of ten (sorry Glen) to reduce the Johnson noise. Why didn't you tell him this? Only after this has been done (lowering the resistors), one should ask whether to use BJTs or JFETs.
Third, the input transistors in his 12W design are operating at 2.5mA. So it's unfair to compare them with noise figures of JFETs operating at a much higher current, that is, 5-15mA.
Cheers, Edmond.
QUOTEOriginally posted by estuart
Whether jfets produce less noise than bjts
depends on the source resistance of system
...normally in the order of 100..1000 Ohms.
Under these conditions BJTs are the preferred choice as input devices. Please also see the attachment.
Cheers, Edmond. /QUOTE
==================================================
Hi .
When occationally visiting here this topic, I may as well use this
( worship, a very popular smiley in nelson pass forum )
to set my position right where it should be.
See my attached figure, refered to by edmondo
------------
1. That formula is just about what I would use, too 😀
( if somebody prefer, or have to, design by ear - this would be almost as good a method
we do not all have to money or want to put priority to buying fancy test gears for detecting nVolt / uVolt noises
.. we have to eat food as well as study & pay rent, too )
2. The diagram with curves a, b , Noise in BJT/JFET vs. Rsource.
This is a comparison we often can find in Datasheet PDFs of JFET input Op-Amps.
In a portion of curve, a range, there is no difference in noise level .. for bjt/jfet.
The actual proportion of type of noise may differ, but the final resulting noise is just about the same for bjt/jfet.
In this range, we so, have maximum options of devices to chose from = both BJT and JFET input transistors.
3. My short rule of Thumb, for 'normal input differential' say:
... Rsource 15-30 kOhm, any low noise BJT or JFET can be used
... Rsource 1Ohm-5kOhm, preferably use BJT low noise transistor(s)
... Rsource 50kOhm- or more, use some good lownoise JFET
😎
This is good information, not only to audio elite debatators.
Will be helpful the average diy-person.
When selecting Operational Amplifier chip that will match you application
or if you are buildning your own input stages.
Regards lineup 🙂 only demands his due respect, which is not too big amount
Whether jfets produce less noise than bjts
depends on the source resistance of system
...normally in the order of 100..1000 Ohms.
Under these conditions BJTs are the preferred choice as input devices. Please also see the attachment.
Cheers, Edmond. /QUOTE
==================================================
Hi .
When occationally visiting here this topic, I may as well use this
( worship, a very popular smiley in nelson pass forum )to set my position right where it should be.
See my attached figure, refered to by edmondo ------------
1. That formula is just about what I would use, too 😀
( if somebody prefer, or have to, design by ear - this would be almost as good a method
we do not all have to money or want to put priority to buying fancy test gears for detecting nVolt / uVolt noises
.. we have to eat food as well as study & pay rent, too )
2. The diagram with curves a, b , Noise in BJT/JFET vs. Rsource.
This is a comparison we often can find in Datasheet PDFs of JFET input Op-Amps.
In a portion of curve, a range, there is no difference in noise level .. for bjt/jfet.
The actual proportion of type of noise may differ, but the final resulting noise is just about the same for bjt/jfet.
In this range, we so, have maximum options of devices to chose from = both BJT and JFET input transistors.
3. My short rule of Thumb, for 'normal input differential' say:
... Rsource 15-30 kOhm, any low noise BJT or JFET can be used
... Rsource 1Ohm-5kOhm, preferably use BJT low noise transistor(s)
... Rsource 50kOhm- or more, use some good lownoise JFET
😎
This is good information, not only to audio elite debatators.
Will be helpful the average diy-person.
When selecting Operational Amplifier chip that will match you application
or if you are buildning your own input stages.
Regards lineup 🙂 only demands his due respect, which is not too big amount
Attachments
To shorten discussion: My Vendetta input stage made of jfets, has a self noise of 0.4nV/rt-Hz. That is 10 ohms noise folks!
Hi John,
That doesn't surprise me in the least. Very good achievement BTW.
One day I hope to see exactly what makes it tick (the Vendetta). There aren't too many exciting or interesting products anymore it seems John.
Jfets are also used in low noise test equipment. I guess the A/D stages like good sounding signals.
-Chris
That doesn't surprise me in the least. Very good achievement BTW.
One day I hope to see exactly what makes it tick (the Vendetta). There aren't too many exciting or interesting products anymore it seems John.
Jfets are also used in low noise test equipment. I guess the A/D stages like good sounding signals.
-Chris
In this patent (John Curl's), the MC catridge's impedance is used as a part of voltage divider (11). MC's impedance is low, below 5ohm.
I see this is smart. Mr. Curl, why you abandoned this approach? Isn't that this approach makes full use of the current generated by MC catridge?
I see this is smart. Mr. Curl, why you abandoned this approach? Isn't that this approach makes full use of the current generated by MC catridge?
Attachments
My MC stage from 1977 was a common-base design, similar
in concept to Leach's except that it was single-ended.
It ran the bias through the cartridge, and works just fine,
although it will expose any bad connection ruthlessly.
😎
in concept to Leach's except that it was single-ended.
It ran the bias through the cartridge, and works just fine,
although it will expose any bad connection ruthlessly.
😎
I personally found that common base does not work as well as common emitter, because most MC cartridges don't really like to be loaded down that much.
We (Levinson and I) made a common base version of the JC-1 and called it the JC-1 AC and JC-1 DC. It sounded worse than the original JC-1 with many cartridges. In fact I took a JC-1 AC and put a switch to change it from common base to common emitter. It was obvious in the comparison.
We (Levinson and I) made a common base version of the JC-1 and called it the JC-1 AC and JC-1 DC. It sounded worse than the original JC-1 with many cartridges. In fact I took a JC-1 AC and put a switch to change it from common base to common emitter. It was obvious in the comparison.
three or four max
I'm mathematically challenged
Edit : given MC output impedance of 150 ohm maybe paralleling many more do more good than harm. I don't know how much though, must look at the FET datasheet, and biasing.
I'm mathematically challenged
Edit : given MC output impedance of 150 ohm maybe paralleling many more do more good than harm. I don't know how much though, must look at the FET datasheet, and biasing.