SSM and PMI were a buddy buddy thing, I figure sort audio hobby gone crazy. The new part was introduced under the SSM name long after they were both absorbed. If I could ask Derek Bowers he would know the real answers as to dies and fabs for the different parts.
Now I'm up against device low Vp. What to do next.... need minimum 3v ?
THx-RNMarsh
You're trying to fight basic device physics. If you want the ultimate linearity and decent noise for mV level signals try a 6DJ8 or a couple in parallel.
A shielded 6922 (the industrial 6dj8) with proper ps can be a very quiet tube, and very linear, never saw noise measurements through.
Just pulling Dick's leg, I don't know what the point is of his latest demands.
Can these conditions be met and still have ultra low noise?
View attachment jfet_distortion.pdf
THx-RNMarsh
View attachment jfet_distortion.pdf
THx-RNMarsh
Must be some bad karma that I'm excluded from this wave of enlightment,
but I see absolutely no connection between BF862 FET noise corner
measurements yielding different results and Rubiolas search for the reasons
why his bipolar amplifier had such a steep 1 / (f**n) rise at 1 Hz.
His amplifier itself is not new, as he writes himself, it has been in the data
sheets of them olde SSM* and Mat* since decades (exposing my age..)
but his connection between drift and the extra steep 1/f has never been shown
so clearly before.
Also, when I presented my measurements on the noise behaviour of
batteries, Scott directly noted the much too steep rise at very low frequencies
and suggested further investigations in this direction. That was half a year ago
in this thread. That a freshly charged battery shows some drift after being
loaded with a few hundred mA for a noise mesurement is to be expected
and I consider Scott's question answered now.
BTW Me being an oscillator guy, Rubiola is quite a hero for me
regards, Gerhard
PS. Pictures of several of my tries at multiple dual transistors are at
< https://picasaweb.google.com/103357048842463945642/LowNoisePreamplifiers?authuser=0&feat=directlink >
Note the production date of the Mat02. They always where to good to use NOW
And it was a good idea to let AD care about the ultra equal load resistors
of the differential input stage and to use complete op amps and no more dual transistors.
Last edited:
Must be some bad karma that I'm excluded from this wave of enlightment,
but I see absolutely no connection between BF862 FET noise corner
measurements yielding different results and Rubiolas search for the reasons
why his bipolar amplifier had such a steep 1 / (f**n) rise at 1 Hz.
His amplifier itself is not new, as he writes himself, it has been in the data
sheets of them olde SSM* and Mat* since decades (exposing my age..)
but his connection between drift and the extra steep 1/f has never been shown
so clearly before.
Also, when I presented my measurements on the noise behaviour of
batteries, Scott directly noted the much too steep rise at very low frequencies
and suggested further investigations in this direction. That was half a year ago
in this thread. That a freshly charged battery shows some drift after being
loaded with a few hundred mA for a noise mesurement is to be expected
and I consider Scott's question answered now.
BTW Me being an oscillator guy, Rubiola is quite a hero for me
regards, Gerhard
PS. Pictures of several of my tries at multiple dual transistors are at
< https://picasaweb.google.com/103357048842463945642/LowNoisePreamplifiers?authuser=0&feat=directlink >
Note the production date of the Mat02. They always where to good to use NOW
And it was a good idea to let AD care about the ultra equal load resistors
of the differential input stage and to use complete op amps and no more dual transistors.
Not trying to exclude anyone, but if you look to the end of that paper he shows the lf noise, with proper cover, with paper and with no cover, a dramatic difference between them, the 50Hz spur is almost not noticed when the chart rescaled to include the lf niose when there is no cover.
PS I really like the substantial case for your preamp...
Just for one data point I have confirmed exact 1/f behavior of at least 1 of our bipolar processes down to 11uHz (1 cycle per 4 days) with enough care taken in the measurement.
So, a jfet device which is under constant temp conditions is measured and found predictable and exact 1/f behavior down to sub Hz freqs.
AND, on the other hand, if the temp is variable, then lf noise wont be as predictable any more.... such as with a draft of air currents past the DUT.
That is my take-away from this. And, packaging has a large affect on the jfet noise when the device IS within a variable airstream.
THx-RNMarsh
Last edited:
Excellent, it's what's known as doing your homework or in other words thorough engineering. I love tinkering in things that few have examined. BTW your preamps look great.
... and around all preamps is enough nickel-plated brass or aluminium, that is better
than a sheet of paper, especially since all DUTs and the preamp were together in a
2nd layer massive aluminium transport box with just BNC feedthroughs for the noise
measurements. If one reads J.Williams or the LTZ1000 doc, one won't leave a hole
for blowin air...
regards, Gerhard
than a sheet of paper, especially since all DUTs and the preamp were together in a
2nd layer massive aluminium transport box with just BNC feedthroughs for the noise
measurements. If one reads J.Williams or the LTZ1000 doc, one won't leave a hole
for blowin air...
regards, Gerhard
Last edited:
It would not surprise anyone in precision measurements field --- in the DIY audio and using discrete devices --- where such control is harder to get and maintain --- [ Though Gerhard's chassis/approach can be used in HiEnd DIYAudio as well], ---- some device packages are much less susceptable. Even the to92 is MUCH less susceptable to temp changes vs time than the tiny sot23 package. And, for me, that is good news because IF I can find what i want in a larger package , I wont have as much trouble meeting my goals.
THx-RNMarsh
THx-RNMarsh
Last edited:
Thank you for the recommendations --- I have quite a few of these and many others to run up on the noise analyzer and curve tracer for starters. Plus ordering new ones. Was hoping someone has already done this laborious sorting for audio at line levels.
THx-RNMarsh
Last edited:
And all those pseudo problems. As if after 50 years there still would be
unknown Golden Transistors, as if the mass of a sot-23 would play any
role if it is soldered to a board and as if a blob of epoxy could not heal
the suspected mass deficit anyway. Or as if a nV would play any role
in a line stage.
Yes, it's a cargo cult.
Gerhard
unknown Golden Transistors, as if the mass of a sot-23 would play any
role if it is soldered to a board and as if a blob of epoxy could not heal
the suspected mass deficit anyway. Or as if a nV would play any role
in a line stage.
Yes, it's a cargo cult.
Gerhard
- Status
- Not open for further replies.