Noise on LiPo batteries?

[alanambrose]

Hi,

Has anyone measured the noise (or is in a position to measure the noise) on LiPo batteries? I've seen the old Nist paper which suggests that AA NiCd are lowest noise. But Nist didn't test LiPo batteries in that study - and on the 'source resistance is responsible for the battery noise' theory, which the Nist paper suggests - then LiPo batteries should be good ... and maybe even lower noise than NiCd.

Anyone?

Alan

[Cerebus]

And further, what's the likely noise performance of EDLC super-capacitors as a power source? If the ESR Johnson noise is the key they ought to perform rather well.

[DuPe]

is there a reference to the NIST paper that you are referring to?

[Zeranin]

Hi,

Has anyone measured the noise (or is in a position to measure the noise) on LiPo batteries? I've seen the old Nist paper which suggests that AA NiCd are lowest noise. But Nist didn't test LiPo batteries in that study - and on the 'source resistance is responsible for the battery noise' theory, which the Nist paper suggests - then LiPo batteries should be good ... and maybe even lower noise than NiCd.

Anyone?

Alan

I have also read a paper on the noise of various battery chemistries, can't remember if it was a NIST paper. From memory, they had to use correlation methods to even measure the noise at all, it was so low. I have always used a battery as a low noise DC voltage source, for testing noise performance of very-low-noise equipment. Out of interest, what is you application that calls for such a low noise source?

[DuPe]

Are we talking about the attachment?
If yes, measurement is going to be a real challenge:
They are talking about 200dBV/sqrtHz precision

[alanambrose]

Yes, that's the one - 1995. It also has an interesting but brief description of the 'cross correlation fft' method....

The measurement methods don't look that difficult - assuming you have access to the previously-mentioned 'cross correlation fft'. A note here for instance:

http://dsp.stackexchange.com/questions/736/how-do-i-implement-cross-correlation-to-prove-two-audio-files-are-similar

and reference [3] in the original paper:

http://tf.nist.gov/general/pdf/90.pdf

Alan

[mmagin]

Does the protection add to the noise?  I assume there's at least a mosfet in the path to provide the over-discharge cutoff.

I guess if you cared you could reimplement that with a really nice relay.

[Kleinstein]

A MOSFET used as a switch should not give much noise. Usually this are low R_on MOSFETs, so a few mOhms of resistance. A relay might be more trouble with not so reliably contact resistance.

[Cerebus]

A MOSFET used as a switch should not give much noise. Usually this are low R_on MOSFETs, so a few mOhms of resistance. A relay might be more trouble with not so reliably contact resistance.

Just to put some numbers on it, the Johnson noise of 1 mOhm is -228 dBv/sqrtHz so we are in the territory where we might care about it mucking up our measurements. In practical terms it's about -40 dB-bee's-dick, i.e. we don't care, but for measurements meant to establish the theoretical limits we might.

[alanambrose]

Are you saying they were not just using a little 2 op amp test rig feeding a SR785 or similar? Sure looks that way, but maybe you have other ideas. I have not investigated fully but the x-correlation method on two similar sources to eliminate op amp noise with lots of samples sounds cute, no? I think this is a standard function in the SR785 and the Agilent equivalent unit. Ref [3] suggests that 'simple' method also.