Author Topic: DIY low frequency noise meter and some measurement result of voltage references  (Read 137089 times)

n109 and 1 Guest are viewing this topic.

Offline dr.diesel

  • Super Contributor
  • ***
  • Posts: 2214
  • Country: us
  • Cramming the magic smoke back in...
@VintageNut

I don't recall seeing your name in the DMM Noise Comparison thread here.  If you've not seen it make sure and check it out.   :-+

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 534
  • Country: 00
@VintageNut

I don't recall seeing your name in the DMM Noise Comparison thread here.  If you've not seen it make sure and check it out.   :-+

Ok, I am reading it now. Thanks
working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16544
  • Country: us
  • DavidH
To measure noise, the peak to peak reading is a little difficult, as it is by definition sensitive to singular events. So it will always be not that well reproducible. Today the better way to measure noise is using FFT and than measure the noise spectrum. A fast DMM like the 7510 is a good option here.
For a given frequency band (e.g. 0.1 -10 Hz as the typical for low frequency noise) one can also use the RMS value (= Std deviation) - this number is much better reproducible than the peak to peak value. To get the right bandwidth this would be more like 10 s windows with something like a 20 Hz sampling rate.

I have always gotten great results using the standard deviation technique to make low frequency RMS noise measurements.  While the sin(x)/x response of most DC voltmeters limits the frequency response (see below), aliasing does not; take away half of the samples and the uncertainty goes up but the standard deviation does not change.  Sampling RF voltmeters take advantage of this to make GHz+ wideband RMS measurements while undersampling by millions of times.

Good peak to peak measurements can be made after amplification using a resetable analog peak to peak detector as described in one of Jim William's application notes or with a fast sampling voltmeter or DSO.

Quote
Doing measurements on very long times, there will be a mixture of noise and drift. To get rid of the drift one might want to add a lower frequency limit, by using a software high pass filter before calculation RMS values.
So the lone run gives a number for about the 0.03 mHz - 500 Hz range. The upper band limit depends on filters used in the DMM.

Besides any filtering, a DMM which integrates the input over a length of time like with a VFC, integrating, or delta-sigma converter, will have a sin(x)/x frequency response which is what is responsible for the 50 and 60 Hz normal mode rejection.  Also do not assume that the sample rate reveals the integration time because some ADCs like the recently discussed LTC2508-32 return multiple correlated samples over the integration time.
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
And here they come...



Bit longer than my KX references, but same width/format.



Bottom side.



Stackup cutout ;)

Now need to get some caps, opamps and get the thing going!  :popcorn: :bullshit:
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 
The following users thanked this post: dr.diesel, Muxr

Offline gamalot

  • Super Contributor
  • ***
  • Posts: 1300
  • Country: au
  • Correct my English
    • Youtube
It's a little bit weird there are no pads for emitter pins of those transistors, rotate them 90deg CCW and have all 3 pads for each will be better.  :)

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Got bit of time to assemble test module, using just one AD8248ARZ.

I think it's not working :)







Reported values are way too low to be real. Tried 10VDC from HP 3245A and 10VDC from EDC MV106. It settles to final figure in matter of seconds. I thought cap need more time to charge?
I'm using Nichicon HE 1000uF 50V for now.

Tried +7/-7V from ultralow noise LDOs (my X1801 preamp power supply card, isolated) and today with paid of 9V batteries with same result.
Scope captures with 1Meg TCA-1MEG adapter and direct BNC cable. Attenuation on scope set to 2000:1.





Seems there is no way around, but to build divider for generator, to make 0.1-1mV peak-peak waveforms to test with known signal.
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline bktemp

  • Super Contributor
  • ***
  • Posts: 1616
  • Country: de
Could connecting the case of the capacitor to ground affect the operation?
The case is often not completely isolated from the capacitor, but has a potential slightly above the negative terminal. For example if you connect a capacitor to 12V, you will measure a couple of 100mV up to a few volts between case and negative terminal.
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Ok, amplifier works with DC coupling and 1Mohm load, so I think I'll need to add LT1028 on the output, so it can drive 50 ohm inputs of my scope directly.

Also I got magical wet slug 1300uF 30V cap (the correct type, XTV), initial check with 100Kohm in series on K2400 shown leakage ~5.2nA at 10VDC after night of soaking. Regular 75 ohm coax used. Looks promising so far.
« Last Edit: December 06, 2016, 04:42:09 am by TiN »
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 14072
  • Country: de
The LT1028 is not really needed to drive the output. With it's high current noise it is even not that suitable to work with source of more than about 1 K (e.g. the resistors from combining the amps). There should be no need for an OP with an extra powerful output (e.g. more than 20 mA) - this might even be a danger to the 50 Ohms input.

Before adding a second amplifier, I would test for current noise as there might no be much use in this if current noise is too high.
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Without buffer, AC coupled output cannot drive the 50 ohm load.

Here's actual schematics in sim:



LTSpiceIV circuit file.

AD8428 model using spice from AD site.

Symbol for LTSpice IV
SPICE file
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 14072
  • Country: de
I don't think a 11 µF cap to GND is a good diode idea at the output.

For AC coupling at the output with 11 µF and 100 K the LT1028 is not a good choice. It has way to much current noise. An OP07 or similar would be a better choice here.

The OPs output should also be insulated from a possible capacitive load. So add a series termination resistor. If DC / LF Gain needs to be 1 and independent of the 50 Ohm, add DC feedback from behind - just like an amplifier made to drive a capacitive load.
« Last Edit: December 06, 2016, 05:51:00 pm by Kleinstein »
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Diode? You meant C5/C6?

I have few LT1464A's, fits the case better?



Not everyday you have chance to see 707$ USD/1pcs capacitor, so here are the photos. Regular radial can caps and SMA connector are nearby for size comparison purpose.



After 10 more hours leakage current stabilized around 3.2 nA at 10VDC from K2400.
« Last Edit: December 06, 2016, 06:02:56 pm by TiN »
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 
The following users thanked this post: Muxr

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 14072
  • Country: de
The LT1494 is OK.

For lowest noise I would consider a larger resistor for R1 and maybe a smaller for R3, so that the lower frequency limit is set by R3*C3,C4 and not by R1*C1. This reduces the noise due to R1 for the lowest octave a little.
 

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 534
  • Country: 00
Hello Tin. Be aware that the 2400 operating at 3nA is not in its "sweet spot". I had occasion recently to use a 2400 to force 1nA and measure capacitor behavior. The 2400 did not respond well at 1nA. A model 236/237/238 responds much quicker on the same device forcing 1nA and letting the capacitor settle to a stable reading.

The length of time to stabilize the capacitor leakage that you observed may be the instrument and not the device.
working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Surely it's not sweet spot of 2400, but I have no time yet to power up 4142B and write program for it to run. I have more of these caps coming in few weeks, so will test and compare them using 41421A (1nA lowest range, 20fA resolution) once I get 'em.

YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline doktor pyta

  • Frequent Contributor
  • **
  • Posts: 488
  • Country: pl
@Tin
What is the reason of clamping input voltage to vcc+0.6V and vee-0.6V instead of simple clamping to +/-0.6V (no problem with leakage current of Q1 and Q2)?
AFAIK the gain of the amplifier will be high and 0.6V will saturate the amplifier.
 
The following users thanked this post: TiN

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
doktor pyta
Right, I somehow missed that fact. Good point, will update it to correct in next rev.

I also got greedy and got more of magical wet slug caps. Two of which are reverse polarity, meaning chassis is positive, not negative.



Also built a test box from cast enclosure with triax ports, so I can make semi-permanent setup to measure components using Agilent 4142B.
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline zhtoor

  • Frequent Contributor
  • **
  • !
  • Posts: 337
  • Country: pk
Great Work.

could you also please test noise on "popcorn" transistors like
PN2222, BC557, 2N3904 etc... in zener connection mode and also
some "standard" 5.1V to 6.8V Zeners?

regards.

 

Offline zhtoor

  • Frequent Contributor
  • **
  • !
  • Posts: 337
  • Country: pk
Re: DIY low frenquency noise meter
« Reply #118 on: February 18, 2017, 03:43:39 am »
Some measurement result
Note also that the results here is only the noises that I tested using my DIY meter on particular voltage references that I possess.

1. Panasonic 3200uF/35V capacitor(I was told this is used for airbags), charge to 4.1V, 124nVpp
2. Panasonic NCR18650B lithium battery, charged to full more than six months ago, 4.1V, 115nVpp
3. A Chinese temperature compensated 6.3V zener, 2DW233, powered by 12V battery thru 1k resistor(5.7mA), 336nVpp, much better than a LTZ1000.
This ultra low noise characteristic of the 2DW23x series(from a particular maker) has been confirmed by many Chinese voltnuts before, but I don't believe this until I had my own test.


When current increased to 11.8mA, noise is reduce even further to 236nVpp.

4. Other measurement result is summarize in table below



5. Ordered by noise


6. Some words about 2DW23x
The one I tested is Diamond brand made by Shanghai 17th Radio Factory. I have a lot of other 2DW23x which are much inferior with noise figures ranging from 20uVpp to 100uVpp. The design and construction of this 2DW23x were completely changed although they still share the same datasheet.

Understandably the noise of a zener is inverse proportional to the square root of the zener current in theory,  and in practice I tested that 2DW233 follows this very well. The mystery is, how they achieve this kind of low noise?


I took apart one and took a photo with my card camera plus a magnifier. It seems to me that they are hand made because the die is not centered and wires are irregular.

great work.

could you please check out the following for noise performance?

MAT01AH (dual matched NPN from AD), one of the transistors connected as a zener in series with the other one as a diode,
and measuring noise from 1 to 10ma perhaps?

i suspect that this may rival all others in terms of noise and drift performance (with appropritately tuned current for zero TC).

regards.

 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Hello,

already did similar with a BCV62 (current mirror) with zener connected transistor and diode.
https://www.eevblog.com/forum/projects/low-cost-voltage-reference-experiment/msg633042/#msg633042

As they only guarantee the zener voltage is > 5V it will be usually not in a useful range for tempco compensation.
I got 8-9V on a simple test which is much too high against the ~6.25V needed.
So I think you will have to buy many MAT01 to find a pair with such low breakdown voltage.

You have many ideas.
Why not build your own low noise amplifier and report your measurement results?

with best regards

Andreas
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 14072
  • Country: de
Many "zener" based noise generators use the base - collector junction to as an extra noise zener. Zener-diodes with more than about 7 V are known to be quite noisy.

Measuring noise of zener diodes could be an interesting field for a noise testing equipment. The specs often don't give very much information on the noise. The interesting part is more the low frequency (e.g. < 10 Hz) part.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
they also use a very low current through the "zener" ...
 

Offline zhtoor

  • Frequent Contributor
  • **
  • !
  • Posts: 337
  • Country: pk
Hello,

already did similar with a BCV62 (current mirror) with zener connected transistor and diode.
https://www.eevblog.com/forum/projects/low-cost-voltage-reference-experiment/msg633042/#msg633042

As they only guarantee the zener voltage is > 5V it will be usually not in a useful range for tempco compensation.
I got 8-9V on a simple test which is much too high against the ~6.25V needed.
So I think you will have to buy many MAT01 to find a pair with such low breakdown voltage.

You have many ideas.
Why not build your own low noise amplifier and report your measurement results?

with best regards

Andreas



thanks for the input. a couple of points though.

1. since I am located in Pakistan, it is generally difficult to get hold of quality parts, so my proposals may sound funny.
2. yes noise testing is the key, that is probably the first piece of equipment any serious volt-nutter should have or build.
(point me in the right direction here.)
3. MAT01AH is already guaranteed to breakdown at the usual five to six volt region (zener as opposed to avalance breakdown).
4. i tend to think that if you did your selection from low-noise transistors otherwise, you may get much better long term drift specs,
   and MAT01AH is already pretty lownoise, along with guaranteed long term drift rates. (availability might be an issue).
   so using the MAT01AH's transistors as zener + diode combination and doing a sweep on the operating current while monitoring
   noise may be worthwhile.

regards, and thanks again for your input.
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Incoming....





I need to patch my scope a bit, and prep the specimens for tests. It will be also interesting to try different capacitors I've bought together with this little neat preamp.
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
RSVD for tests.

Keithley 2400 Vsource , +9.5Vset.



YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf