time-nuts@lists.febo.com

Discussion of precise time and frequency measurement

View all threads

Re: +1/f of transistors

MM
Mike Monett
Fri, Apr 15, 2022 4:37 PM

Here is a different approach to the 1/f problem.

Marco Reps has an interesting article on low noise amplifiers:

Ultralow Noise Tester: 9V Battery vs. 7805 vs. LTZ1000
https://www.youtube.com/watch?v=XpbDMo8an5w

He uses the zero drift op amp ADA4523: $3.03 at Mouser and Digikey
The datasheet is at https://octopart.com/search?q=ADA4523-1BCPZ

It shows 4.2nV/root(Hz) from 0.1Hz to 1Khz. Cross-correlation could bring
this down by 20dB.

The results could be combined with a low-noise bipolar to extend the
frequency range as high as needed.

The ADA4523 was released in 2020, so it was not available in 2014 when
Gerhard ran his famous "Noise Measurements On Chemical Batteries".

MRM

Here is a different approach to the 1/f problem. Marco Reps has an interesting article on low noise amplifiers: Ultralow Noise Tester: 9V Battery vs. 7805 vs. LTZ1000 https://www.youtube.com/watch?v=XpbDMo8an5w He uses the zero drift op amp ADA4523: $3.03 at Mouser and Digikey The datasheet is at https://octopart.com/search?q=ADA4523-1BCPZ It shows 4.2nV/root(Hz) from 0.1Hz to 1Khz. Cross-correlation could bring this down by 20dB. The results could be combined with a low-noise bipolar to extend the frequency range as high as needed. The ADA4523 was released in 2020, so it was not available in 2014 when Gerhard ran his famous "Noise Measurements On Chemical Batteries". MRM
R(
Richard (Rick) Karlquist
Fri, Apr 15, 2022 7:13 PM

When I was working on a cold atom Rb standard a few years ago,
I settled on the LT1028 with a 3.5 Hz (yes Hertz) noise corner and
noise floors of less than 1nV/sqrtHz and 1pA/sqrtHz.  GBP>50 MHz.
Specified on the data sheet.  In production now for ~40 years!  I can
remember when it first came out.

What's not to like about this?  Are low flicker discreet transistors
still  needed?

Rick N6RK

When I was working on a cold atom Rb standard a few years ago, I settled on the LT1028 with a 3.5 Hz (yes Hertz) noise corner and noise floors of less than 1nV/sqrtHz and 1pA/sqrtHz. GBP>50 MHz. Specified on the data sheet. In production now for ~40 years! I can remember when it first came out. What's not to like about this? Are low flicker discreet transistors still needed? Rick N6RK
G
ghf@hoffmann-hochfrequenz.de
Fri, Apr 15, 2022 8:22 PM

Am 2022-04-15 21:13, schrieb Richard (Rick) Karlquist:

When I was working on a cold atom Rb standard a few years ago,
I settled on the LT1028 with a 3.5 Hz (yes Hertz) noise corner and
noise floors of less than 1nV/sqrtHz and 1pA/sqrtHz.  GBP>50 MHz.
Specified on the data sheet.  In production now for ~40 years!  I can
remember when it first came out.

What's not to like about this?  Are low flicker discreet transistors
still  needed?

Ooohps, that one was sent while uploading the 2nd picture. Weird.
The first post is very redundant. Very redundant, that's what it is.

I would not call that a 3.5 Hz corner frequency, esp. for the
current. And also, I do not like the noise hump at > 200 KHz that was
undisclosed for > 35 of the 40 years. I also do not like the new
prices, but that also goes for the AD797.

ADA 4898-2 seems to deliver most for the $.
Two amplifiers for half the price.

TI guarantees not much for their LT1028 clone, but then I also
got LT1028 from a local Maxim distributor, many years ago.
Must be quite exotic.

Gerhard

Am 2022-04-15 21:13, schrieb Richard (Rick) Karlquist: > When I was working on a cold atom Rb standard a few years ago, > I settled on the LT1028 with a 3.5 Hz (yes Hertz) noise corner and > noise floors of less than 1nV/sqrtHz and 1pA/sqrtHz. GBP>50 MHz. > Specified on the data sheet. In production now for ~40 years! I can > remember when it first came out. > > What's not to like about this? Are low flicker discreet transistors > still needed? Ooohps, that one was sent while uploading the 2nd picture. Weird. The first post is very redundant. Very redundant, that's what it is. I would not call that a 3.5 Hz corner frequency, esp. for the current. And also, I do not like the noise hump at > 200 KHz that was undisclosed for > 35 of the 40 years. I also do not like the new prices, but that also goes for the AD797. ADA 4898-2 seems to deliver most for the $. Two amplifiers for half the price. TI guarantees not much for their LT1028 clone, but then I also got LT1028 from a local Maxim distributor, many years ago. Must be quite exotic. Gerhard