775 Nanovolt Noise Measurement for a Low Noise Voltage Reference

with Jim Williams, Staff Scientist
video.linear.com/7?...
Frequently, voltage reference stability and noise define measurement limits in instrumentation systems. In particular, reference noise often sets stable resolution limits. Reference voltages have decreased with the continuing drop in system power supply voltages, making reference noise increasingly important. The compressed signal processing range mandates a commensurate reduction in reference noise to maintain resolution. Noise ultimately translates into quantization uncertainty in A-to-D converters, introducing jitter in applications such as scales, inertial navigation systems, infrared thermography, DVMs and medical imaging apparatus.

[@pennyl.8799]

They just don't make analog engineers like him any more

[@Jackandthebeanstock]

Simulation tools have corrupted all engineers to lose their skills, instinct and memory to recall critical fundamental aspects of electronics.

[@hrmyler8463]

A much drier wit than Bob Pease (RIP), quite fun. Only engineers can understand the wit of engineers. I also liked his footnote in the app note: The plastic cup, supplied by Martinelli and Company, also includes, at no charge, 10 ounces of apple juice. He briefly mentions it in the video.

[@LinearTechnology]

@nick1f - The preamplifier has a gain of 10,000. The filter then provides an addition gain of 100 for a total gain of 1,000,000. This translates 0 to 1uV to 1 to 1V.

[@KissAnalog]

Thanks so much!

[@ramanarao32]

oh my God !!! i saw the Guru finally..i felt sad when i learnt that he passed away...i used to keep as application notes as a bible for many years...Together with RAP he might have opened an analog lab for God in the heavens!!

[@0x80O0oOverfl0w]

A very interesting video. Can you make the videos available at 480p in the future? It's very hard to read any of the text in the videos at 240p. Thanks and keep posting videos!

[@mineown1861]

Tech gold .

[@windward2818]

It was interesting how he confirmed testing front end bandwidth by using rise time, and do not think he had easy access to a high accuracy spectrum analyzer effective down to DC.

[@nick1f]

Hi, thanks for explanation. I am very sorry, I made a silly mistake. Now it is all clear. I am very sad to hear about the death of Jim Williams. The world lost a great engineer.

[@kaybhee6]

keen

[@upnicmars]

RIP. :(

[@paulg.3067]

240p? Pffff! Real professionals watch this in 144p, upscaling it in the brain to 1080p! :D

[@Nermash]

Rip Jim...

[@nick1f]

Isn't there a mistake in the presentation? If the gain of the filter is 100, shouldn't the 0 to 1uV translate into a 0 to 0.1V (not 1V)? Isn't the gain of the filter 1000 instead?

[@MrHmm-cv6gs]

He looks like hiachi mishima from takken game,

[@nsfeliz7825]

dilberts boss. looks like dilberts boss.

[@kevinaylward]

Ahmmm, well, all this peak detecting and extra analog pissing about is really all pedestrian and not the way to do it. Pretty much any testing system, today, and for the least 20 years at least, especially for frequencies this low, is gain up the signal large enough for standard data acquisition boards to not dominate the noise, and process it all digitally, with ahmm, a PC. T.he flexibility of doing everything digitally is simple too overriding for any other method to add any value. Hardwired Analog processing is just way too restrictive.

[@pureturnip]

So using amplification and and standard data acquisition you are going to introduce much less than 775 nanovolts to your noise floor and then make your measurement? Probably not. I think Jim has the right idea here.

[@kevinaylward]

Not at all. Jim's idea is wrong today, and probably so for 10 years or more. Of course, once it is gained up with a low noise amolifier, the problem is, essentially, all over with modern data acquistion boards. Feel free for more analog tutorials at www.kevinaylward.co.uk/ee/index.html :-)

[@gino3286]

"is gain up the signal large enough" i guess not a trivial task without introducing further noise. This is a technical challenge. And if i am not wrong is also what Dr.Williams explains at 4:53. A low noise high gain preamp is needed. Easy to say difficult to execute. As i am interested in audio fidelity i see a similar issue with MC cartridges that put out signals as low as some tenths of mV. To amplify that signal with low noise and distortion is a challenge.

[@kevinaylward]

@@gino3286 There are some considerations. The lowest noise would be to use a diff pair of discrete transistors within the feedback loop of a commercial op amp. Available commercial op amps don’t appear to get to the 0.5 nV/rthz achievable with external transistors, that is, thay have low base spreading resistance such as below 10 ohms. The feedback potential divider needs to have something like 10 ohms to ground. For lower closed loop gains this means a good drive capability. Gain should b kept less than about 100 to avoid dc offset, a blocking cap in series with the 10 ohm is problematic. Blocking the dc after gaining up the signal is a better method.

[@gino3286]

@@kevinaylward Hi ! thank you very much indeed. Unfortunately i am very uneducated in electronics. I need a kit to hope to make something ... because i understand that a even a schematic alone is not enough. Also the components lay-out is very important. So beyond my reach. Thanks a lot anyway for the kind and helpful advice. Kind regards, gino