Author Topic: DIY low frequency noise meter and some measurement result of voltage references  (Read 67545 times)

0 Members and 2 Guests are viewing this topic.

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2326
  • Country: de
Re: DIY low frenquency noise meter
« Reply #25 on: May 13, 2016, 06:22:39 am »
Did a quick simulation of a equivalent filter:

(2*620 R in parallel = 310 R single)

with best regards

Andreas
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 5793
  • Country: de
Re: DIY low frenquency noise meter
« Reply #26 on: May 13, 2016, 07:04:20 am »
With noise measurements it's not the -3 dB point's that really count, but the integrated curve. The filter is not perfect and will still give some contribution from outside the -3 dB points. For white noise one uses a equivalent noise bandwidth. With the often dominating 1/f noise it gets more complicated. So to compare data one needs more than just the band limits, but also the type of filter, especially at the lower end.

Usually the RMS values are a little better to measure, as they don't fluctuate as much as the peak - peak values.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2326
  • Country: de
Re: DIY low frenquency noise meter
« Reply #27 on: May 13, 2016, 07:13:05 am »
And a possible dimensioning when you do not want to change the electrolytics.

With noise measurements it's not the -3 dB point's that really count, but the integrated curve.

Unfortunately there is no "standard" for measuring. Some use 2nd order other 4rth order cirquits.

The most important thing is: you have a cirquit where you can compare different references.
(And sort out the "stinkers").

@Zlymex: for the LM399 / LM329 the current should not play a large role above  0.5-1mA.
All current above around 250uA is shunted away from the zener element.

With best regards

Andreas
« Last Edit: May 13, 2016, 07:19:48 am by Andreas »
 

Offline zlymex

  • Frequent Contributor
  • **
  • Posts: 556
  • Country: cn
Re: DIY low frenquency noise meter
« Reply #28 on: May 13, 2016, 10:57:46 am »
I did the simulation before, I just took that -6dB fall off for granted. That's inevitable isn't it for a second order filter? I mean Linear did the same thing for their filters in AN124f where they choose 1300uF-1.2k and 165uF-10k for 0.1Hz HPF.

The question is, should I modify the parameters so that the fall off at 0.1Hz and 10Hz be -3dB?
 

Offline zlymex

  • Frequent Contributor
  • **
  • Posts: 556
  • Country: cn
Re: DIY low frenquency noise meter
« Reply #29 on: May 13, 2016, 11:07:35 am »
And a possible dimensioning when you do not want to change the electrolytics.

With noise measurements it's not the -3 dB point's that really count, but the integrated curve.

Unfortunately there is no "standard" for measuring. Some use 2nd order other 4rth order cirquits.

The most important thing is: you have a cirquit where you can compare different references.
(And sort out the "stinkers").

@Zlymex: for the LM399 / LM329 the current should not play a large role above  0.5-1mA.
All current above around 250uA is shunted away from the zener element.

With best regards

Andreas
Thanks very much for the simulation. I didn't see this second page until now because I've just got up in the morning and my mind is still in a half sleep stage. I'm sorry I'm unable to respond this thread in the next 8 hours time.

True for the zener current. I just got that 1k resistor and 12V battery at hand. That slightly larger current than required had some heating effect.
 

Offline Alex Nikitin

  • Frequent Contributor
  • **
  • Posts: 997
  • Country: gb
  • Femtoampnut and Tapehead.
    • A.N.T. Audio
Re: DIY low frenquency noise meter
« Reply #30 on: May 13, 2016, 07:24:26 pm »
I did the simulation before, I just took that -6dB fall off for granted. That's inevitable isn't it for a second order filter? I mean Linear did the same thing for their filters in AN124f where they choose 1300uF-1.2k and 165uF-10k for 0.1Hz HPF.

So they've cheated  ;) .

The question is, should I modify the parameters so that the fall off at 0.1Hz and 10Hz be -3dB?

I suppose yes. Below is the simulation I've done to get -3dB points at right frequencies.

Cheers

Alex
 
The following users thanked this post: zlymex

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2326
  • Country: de
Re: DIY low frenquency noise meter
« Reply #31 on: May 13, 2016, 07:58:44 pm »
I mean Linear did the same thing for their filters in AN124f where they choose 1300uF-1.2k and 165uF-10k for 0.1Hz HPF.

Ok good point: that explains (partly) why for the LTC6655 (5V) I get different results against the data sheet.
Datasheet value is 0.25ppmpp = 1.25uVpp.
whereas I got 2.2uVpp for the MSOP and 2.7uVpp for the LS8-package.

Of course LT will not change its cirquit so they would have to update their datasheets with higher 1/f noise levels.

The question is, should I modify the parameters so that the fall off at 0.1Hz and 10Hz be -3dB?

Its not necessary from my side. But in comparison with other amplifiers you will have perhaps -10 or -20% deviation.
TI obviously uses -3 dB corner frequency definition (at least for characterisation of OP-Amps).
But with a 2nd Order high pass and a 4rth order low pass.
http://www.ti.com/lit/ug/slau522/slau522.pdf
http://www.ti.com/tool/tipd122?keyMatch=0.1-10hz%20filter&tisearch=Search-EN-Everything

So if you measure TI devices you will probably need a different amplifier ;-)  :-//

On the other side: Electrolytics have large tolerances (up to +80/-20%)
so for the lower limit (high pass) you will probably be already on the safe side.
But for the low pass 620R||620R + 2*100uF in series the upper 10Hz limit may be much too low if you do not have measured the capacitors before soldering in.

Edit: Alex was faster. (is the cirquit a "defeat device"?)

With best regards

Andreas
« Last Edit: May 13, 2016, 08:46:40 pm by Andreas »
 
The following users thanked this post: zlymex

Online alanambrose

  • Frequent Contributor
  • **
  • Posts: 266
  • Country: gb
Re: DIY low frenquency noise meter
« Reply #32 on: May 14, 2016, 12:16:37 am »
A very nice design and build - I think this wins the 'most functionality out of a smallish number of components competition' :)

Alan
“A foolish consistency is the hobgoblin of little minds"
 
The following users thanked this post: zlymex

Offline zlymex

  • Frequent Contributor
  • **
  • Posts: 556
  • Country: cn
Re: DIY low frenquency noise meter
« Reply #33 on: May 14, 2016, 03:07:49 am »
Thanks very much every body for the comments and suggestions especially for Andeas providing the probe and Alex providing the simulation.
Now I modified the schematics(uploaded in my first post) according to the parameters from Alex's which I've confirmed in my simulation.
I also modified my meter so that it reflect most of the changes in the schematics. It turns out that I've already use 1k for R1 owning to C1 in my first implementation is 1500uF. Now the floor noise of my meter is increase by about 11%. I will use purple letter for chart marking from now on.
« Last Edit: May 15, 2016, 03:05:09 am by zlymex »
 

Offline DuPe

  • Regular Contributor
  • *
  • Posts: 67
  • Country: de
Re: DIY low frenquency noise meter
« Reply #34 on: May 14, 2016, 05:18:14 pm »
Can someone help me to understand the virtual ground concept within this schematic?
Virtual ground is done by U2A. LMC6064 is able to provide 16mA max.
Assymetry introduced by the voltmeter (I took murata DMS-40PC series as example) is roughly 70mA for the low power version.
And RS3 only contributes ~20mA to the balance
How does this work?
Cheers
Peter
« Last Edit: May 14, 2016, 06:02:28 pm by DuPe »
 

Offline zlymex

  • Frequent Contributor
  • **
  • Posts: 556
  • Country: cn
Re: DIY low frenquency noise meter
« Reply #35 on: May 14, 2016, 08:49:31 pm »
Can someone help me to understand the virtual ground concept within this schematic?
Virtual ground is done by U2A. LMC6064 is able to provide 16mA max.
Assymetry introduced by the voltmeter (I took murata DMS-40PC series as example) is roughly 70mA for the low power version.
And RS3 only contributes ~20mA to the balance
How does this work?
Cheers
Peter
My LED meter draws only 18mA current, therefore the U2A provides only about 2mA extra.
If that murata DMS-40PC has to be used here, a transistor booster to the U2A is necessary. However, I'm not quite sure about the loop stability.
« Last Edit: May 14, 2016, 09:04:56 pm by zlymex »
 

Offline DuPe

  • Regular Contributor
  • *
  • Posts: 67
  • Country: de
Re: DIY low frenquency noise meter
« Reply #36 on: May 14, 2016, 09:16:03 pm »
Can someone help me to understand the virtual ground concept within this schematic?
Virtual ground is done by U2A. LMC6064 is able to provide 16mA max.
Assymetry introduced by the voltmeter (I took murata DMS-40PC series as example) is roughly 70mA for the low power version.
And RS3 only contributes ~20mA to the balance
How does this work?
Cheers
Peter
My LED meter draws only 18mA current, therefore the U2A provides only about 2mA extra.
If that murata DMS-40PC has to be used here, a transistor booster to the U2A is necessary. However, I'm not quite sure about the loop stability.
Thanks zlymex, this makes things fit together.

Cheers
Peter
 

Offline splin

  • Frequent Contributor
  • **
  • Posts: 719
  • Country: gb
Excellent work zlymex. It would be interesting to see the noise floor with a few values of R1, up to say 1M, to measure the actual input noise current.
 

Offline zlymex

  • Frequent Contributor
  • **
  • Posts: 556
  • Country: cn
Excellent work zlymex. It would be interesting to see the noise floor with a few values of R1, up to say 1M, to measure the actual input noise current.
Thanks. Did you mean to change C1 as well togther with R1?
I can at least do 22uF-100k pair and 2.2uF-1M pair with no difficulty.
 
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 5793
  • Country: de
The interest is likely getting input current noise data on the OPs. So it's about changing R1 to a large value (like 1 M or 10 M) and keep the input open, so C1 does not matter.

Input current noise data for OPs seem to be not that reliable, so measured data are of interest. It might be good to know anyway how much input current noise the circuit has, just to know the background. Though I don't think it will not be much of a concern with just 1 K at the input.
 

Offline zlymex

  • Frequent Contributor
  • **
  • Posts: 556
  • Country: cn
Okey, I managed to run the circuit at R1=10 Meg and C1 open. It is not as easy as I thought because the voltage created by Ib on R1 is very large and easily saturated the output.
Anyway, the reading is somewhat between 27uVpp and 32uVpp, see chart attached.
The noise of a 10M resistor is roughly 6uVpp @10Hz bandwidth, therefore that 30uVpp is largely contributed by the noise current of the opamp which will be less than 32uVpp/10M/1.4 = 2.3pApp per opamp, smaller than specified 10pApp.

I also measured the Ib = 150pA
« Last Edit: May 17, 2016, 09:05:35 am by zlymex »
 

Online alanambrose

  • Frequent Contributor
  • **
  • Posts: 266
  • Country: gb
“A foolish consistency is the hobgoblin of little minds"
 

Offline zlymex

  • Frequent Contributor
  • **
  • Posts: 556
  • Country: cn
Hey zlymex,

Is this the low leakage capacitor you're using?

http://www.mouser.co.uk/ProductDetail/United-Chemi-Con/EKMG350ELL222ML25S/?qs=sGAEpiMZZMtZ1n0r9vR22WE9Am08kdJyz%252bhaf7%252bY9CY%3d

Regards, Alan

Yes, the type is KM, also rated at 105 deg C, but not low leakage type.

I choose this KM type because that is the thing immediate available to me, and I selected the one from a batch of five(according to minimum leakage).
Recently, I tested more capacitors of 470uF from Nitsuko(25V EL(M) type), Nippon Chemi-Con(35V SME type), Nichicon(25V VX(M) type), they are all very good in leakage current, but again should be selected in about 1 to 3 ratio. After apply 10V for a day, leakage current of about 50% of those capacitors is below 2nA.

Edit: someone suggested to me to use low leakage type(0.002CV leakage current) such as ELNA RLB, Nichicon KL or UKL
« Last Edit: May 31, 2016, 11:48:02 am by zlymex »
 

Online alanambrose

  • Frequent Contributor
  • **
  • Posts: 266
  • Country: gb
Thanks zylmex for the capacitor info. Re Nichicon UKL...

>>> After 1 minute's (for case size 10 × 12.5 or smaller) or 2 minutes' (for case size 10 × 16 or larger) application of rated voltage at 20°C, leakage current is not more than 0.002CV or 0.2 (µA) whichever is greater.

For 2,200uF / 50V that's ~220µA - so the spec isn't that helpful. Will order a few in to test. The Chemicon equivalents btw seem to be their LLA series.

Regards, Alan

« Last Edit: June 01, 2016, 08:02:19 pm by alanambrose »
“A foolish consistency is the hobgoblin of little minds"
 

Offline zlymex

  • Frequent Contributor
  • **
  • Posts: 556
  • Country: cn
.......
For 2,200uF / 50V that's ~220µA - so the spec isn't that helpful. .............
That's true. However, the actual leakage is much smaller than specified for all the capacitors I tested. The KM type I use is specified as 0.03CV, 15 times worse than the low leakage type(0.002CV), so we can expect better performance for low leakage type.
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 5793
  • Country: de
Leakage testing takes some time. So it is not practical to do the leakage test on every cap produced. Thus leakage specs are very conservative so that even the worst samples will meet the specs - for most applications the leakage is not critical. So there is not that much demand for caps with low leakage specs.
 

Offline splin

  • Frequent Contributor
  • **
  • Posts: 719
  • Country: gb
Re: DIY low frenquency noise meter
« Reply #46 on: July 19, 2016, 01:59:00 am »
I mean Linear did the same thing for their filters in AN124f where they choose 1300uF-1.2k and 165uF-10k for 0.1Hz HPF.

Ok good point: that explains (partly) why for the LTC6655 (5V) I get different results against the data sheet.
Datasheet value is 0.25ppmpp = 1.25uVpp.
whereas I got 2.2uVpp for the MSOP and 2.7uVpp for the LS8-package.

Of course LT will not change its cirquit so they would have to update their datasheets with higher 1/f noise levels.

I don't think this is right. I simulated the AN124f filter and it's not far off: as you can see in the picture (normalized to 0dB gain) the -3dB points are at .192Hz and 9Hz which shouldn't reduce the total noise by much more than about 8% compared to .1 to 10Hz.

I think the confusion that has arisen is that zlymex was pointing out above that AN124 .1Hz HPF is incorrect and I'm guessing you assumed that AN124 LPF also had the same problem as zlymex's original LPF with a cutoff at 6.5Hz. It doesn't really matter of course how it arose but it means that there must be some other explanation of the discrepancies between the LTC6655 specs and the measurements.

zlymex measured the LTC6655 noise, in reply 2, as approx .33ppm compared to the .25ppm spec, (33% above spec) but I assume that was using the .16Hz - 6.5Hz filter (as calculated by Alex) - the schematic has been updated since but presumably the results are still from the original circuit. I estimate that the bandwidth reduction should reduce the measured noise by about 15% which means the .1 - 10Hz measured noise would be 33%/.85 = 39% above the spec. The .25ppm spec figure is as usual a typical figure so the part still meets the specs.

Andreas's measurements above, at 76% to 116% above the typical are much worse, and whilst technically within spec are very disappointing for a part which is expressly marketed as being low noise. I estimate that the .14Hz -3dB HPF error compared to .1Hz would only have an error of around 2%, so pretty negligible. So the question is, was Andreas particularly unlucky with the various parts he tested, or did the LT engineer/marketing bod who wrote the datasheet previously work at Vishay  >:D or is Andreas's LTC6655 test setup / layout inducing excess noise for some reason?

To estimate the integrated 1/f noise with differing bandwidths, I used the formulae:

Vnoise rms = Vnw x sqrt(FC x ln (FH/FL))

where:

FC = 1/f cutoff frequency
Vnw = noise density well above FC
FH = HPF -3dB frequency
FL  = LPF -3dB frequency

Vnw is used for frequencies above FC when FH > FC

The only voltage reference datasheets I could easily find showing 1/f noise spectrums were the LTZ1000 and the ADR4520 (bandgap), both of which showed FC to be around 2Hz and the AD584 with FC nearer 150Hz. I used 2Hz in the above estimates; increasing it to 150Hz increases the impact of the LPF errors but decreases the HPF errors. Eg. The AN124 bandwidth errors only change the integrated noise from 92% to 91% (of .1 to 10Hz) when FC is increased from 2 to 150Hz.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 9535
  • Country: us
  • DavidH
Re: DIY low frenquency noise meter
« Reply #47 on: July 19, 2016, 02:20:11 pm »
BTW--- the LTC2057 suppresses 1/f noise.  It's LF noise spec is from DC-10Hz because of this.

Chopper stabilized amplifiers have flat 1/f noise to DC but higher broadband noise.  You can combine a chopper stabilized amplifier with normal amplifier to get the best noise characteristics of both.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2326
  • Country: de
Re: DIY low frenquency noise meter
« Reply #48 on: July 19, 2016, 03:19:52 pm »

 or is Andreas's LTC6655 test setup / layout inducing excess noise for some reason?


Hello,

thanks for the analyzing in detail.

The only thing in the test setup that I can imagine is probably the power supply voltage used.
The datasheet specs VREF+0.5V for all measurements.
Whereas I am typically using 9 or 10.2V for my measurements.
Unfortunately I did not record the power supply voltage.

And the LTC6655 is heating a lot with the 5 mA consumption.

I will do a comparison at different supply voltages (perhaps at the week end).

With best regards

Andreas

 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 5793
  • Country: de
The low frequency noise can vary quite a lot from one sample to another. Its also rather slow an thus expensive to measure. Even with expensive parts like the LTZ1000, there are good ones an bad ones with way more LF noise than typical specs. So it really makes sense to have a system for LF noise measurements.

Besides electronic LF noise, there can also be thermal noise from turbulant air flow and thermal EMF and similar. This can look rather similar to 1/f noise.

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf