Hi all, I'm looking for a QST construction article about a circuit designed for measuring the noise generated by resistors. I recall the design used a paint can as the enclosure. My interest is in the low noise amp design.
Mike
All resistors of a given value generate the same Johnson noise voltage, so there's no practical reason to measure that.
Practical value metal film resistors have almost no shot (Poisson) noise. I have measured excess noise in Gohm range cermet resistors, but it was difficult. I used a Danish butter-cookie can.
-- John Larkin Highland Technology, Inc lunatic fringe electronics
Not so bad if you use the correlation method, which gets rid of the voltage noise of the measuring gizmo. (It doesn't help the current noise of course.)
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
amdx wrote in news:q5116v$jh5$ snipped-for-privacy@dont-email.me:
Close! Not QST, but...
John Larkin wrote in news: snipped-for-privacy@4ax.com:
The old, bulk carbon media resistors were the most noisy.
They quell HV arcs without damage real good though.
ned
d a
Mike
Not exactly what you are looking for but Heathkit's Audio Signal tracer kit had a Noisy Resistor test feature. It put a voltage across the resistor in parallel with the audio probe and you could easily find the noisier resistors. Not calibrated at all, other then your ear.
You could use that concept for testing resistors with an op amp for amplifying the noise using a capacitor for dc isolation if you can't find the article.
John :-#)#
It's not all that hard if you are OK with ~audio frequencies, opamps and largish (10 k ohm) resistors.
You do need a lot of gain.
Some sort of shielding will most likely be needed.
I've seen designs will all sorts of involved grounding schemes. But it works fine if you just nail one end of the resistor to the inside 'ground' of your cookie tin.
I can talk more if you are interested.
George H.
It generates the same noise power not noise voltage.
It's a resistor. Power is V**2/R. And noise is random, so it's only the power spectrum and amplitude statistics that we care about (or that are well-defined).
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
The MIT experiment uses an SRS preamp, the SR560, to measure resistor noise. In AoE III's low-noise chapter, 122-pages, four pages (512-515) discuss in detail the SR560's design principles. We show how to make even more quiet amplifiers.
--
Thanks,
- Win
You should read the 122 page chapter in AoE III, all about noise and low-noise amplifiers. It's easy to make amplifiers be far more quiet than resistors.
--
Thanks,
- Win
If it generates the same power, won't it generate the same voltage?
Resistors of the same value generate the same Johnson noise voltage density. Stray capacitance can round off the high frequency part.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
You can do pretty well with just opamps these days. On the practical front, we used FET opamps with 'no' (or negligible) current noise. This makes it easy to separate the opamp noise from the resistor noise, by just changing the input resistor.
(But your low noise chapter is great.. and well worth my re-reading.)
George H.
A 10K resistor makes a lot more noise than a good jfet opamp.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Am 25.02.19 um 20:51 schrieb John Larkin:
It is not so much more. Good FET opamps run out of steam at 5.x nVrt Hz That is abt. 1K5.
When there was a dual op amp in SO-8 that does 1nV/rtHz I had one less problem.
cheers, Gerhard
Right, it looks better if you plot noise density (V^2/Hz) Say as long as the noise 'experts' are here. How do you disentangle voltage noise and current noise in a bjt amp? I guess I'm asking how you measure the current noise?
George H.
I'll do that, I need to pull your book of the shelf more often.
My curiosity started when I watched a video of a guy using an LM386 as a amp for a parabolic reflector and a microphone. I tried to tell him his circuit was noisy and he could do better with a quiet front end. He argued it was very quiet. I gave him a search with 100 links to how noisy the LM386 is, he didn't budge on his it's very quiet. Hard to tell from his video, but it's seems like his electronics are noisy to me. >
Mikek
Haha, using a speaker amplifier as a low-noise preamp! But I suppose if you have a parabolic reflector to concentrate the noise, sorry, the signal, it might be more noisy than the amplifier.
--
Thanks,
- Win
Short the input to see the voltage noise.
Then use resistors to see the current noise, corrected for Johnson noise.
Maybe you could connect two opamp inputs together to a resistor or a cap. Voltage noises wouldn't correlate, current noise would. Maybe that's what Phil was talking about.
That's a useful concept: use identical gadgets to measure themselves.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
** What about the electronics INSIDE the electret microphone ??
Likely to generate more noise then the LM386 input stage does.
FYI:
Low noise microphones typically have large diaphragms and are true condenser types - not some cheap electret capsule.
.... Phil
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.