Has anybody measured the actual noise on a Fluke 732a?
My perception is that they seem very quiet. On an 8.5 digit meter, I am seeing noise SD of 170nV on the 10V output, which is about as good as the meter ever reads on its own. Now, that's over a few minutes and there could obviously be longer-term noise, but the base level seems so low that it would be hard to separate out longer-term noise without a thermally and electrically quiet environment.
I guess this is what you get when you put high-end parts in an oven and run it for a couple of decades.
maybe use the method described in this paper.
best regards.
-zia
Other people have discussed and explained flicker noise so here is a visual to support the understand of flicker noise floor limitations in Voltage Standard measurements.
I used a K2182 Nanovoltmeter to monitor the 10V output voltage of two Fluke 732A zener references. This is similar to what you would do to calibrate a known Voltage reference to an unknown reference. For a real calibration the interconnect leads would be reversed to average the thermal EMFs. During this test the temperature was stable and I did not see any correlation between error voltage and temperature.
The raw data was sampled every 30 seconds and it has a 2.4 uVpp noise level. If that data is post processed with a 10 minute moving average
the noise level only drops to 1.8 uVpp. With an extreme 2 hour moving average the noise is reduce to 1 uVpp. At these time levels the drift of the zeners starts to add to the noise. You can see a slight downward slope to the traces after 100 hours.
With 2 or 4 independent references you can get some noise reduction and you reduce the influence of one bad device.
In this previous post I provided F732A - F732A noise data for 100 hours. It has an attached pdf.
Thanks, those links are perfect, and are consistent with what I am seeing, at least in the short term.
There's no doubting that nanovoltmeter and the differential method.
Interesting that the noise on the 732a roughly matches that of the 3458a.
Hello martinr33 & Voltnuts,
the graphs below were taken with one of my K7510 units hooked on a Fluke 732A in my lab ... - the Y-axis has 1µV/DIV scale
What settings did you use for NPLC?
What is the best settings for the Keithley DMM7510 to detect noise?
May be low NPLC and no filter?
....
What settings did you use for NPLC?
What is the best settings for the Keithley DMM7510 to detect noise?
May be low NPLC and no filter?
As the DMM7510 uses averaging sorter conversions to get the very long integration time, I would suspect to use a PLC setting similar to that for internal averaging (e.g. 1, 3 or 5 PLC). No filter is a good idea, and also using AZ is likely needed for longer times.
However it also depends in which noise frequencies you are interested. In the simple case of using RMS calculation the data rate and sample length set the frequency range. To get close to the common 0.1 -10 Hz range, I would expect about 1 PLC with sync AZ (and thus 24 / 29 reading per second) and than a set of data to cover 10 seconds.
For the 7510, hidden in the 1000+ page reference manual (sitting in front of me) NPLC of 2 or 5 for lowest noise. I usually use 5. I found no difference with the Line Sync on or off in my case. Also be sure to use Auto Z. Math filtering will only mask noise, not reduce it.
How would you describe the difference between reducing noise and masking noise? Is a low-pass filter reducing or masking noise? Or analog averaging (integration)? What about a DSP implementation of that same filter?
How would you describe the difference between reducing noise and masking noise? Is a low-pass filter reducing or masking noise? Or analog averaging (integration)? What about a DSP implementation of that same filter?
Ideally, the measuring device would have zero noise so only the DUT noise would be measured. You can optimize the instrument for minimal noise, but the math averaging functions will mask the noise of the DUT.
Q.: Are there other proposals - better ways - to measure low noise on a Fluke 732A ? How do we know that the measured values with the DMM7510 are not mainly dominated by its own noise floor as it seems that the noise contribution of the 732A should be much lower ... ()
Hello,
if you want to measure noise then you should ever define the kind of noise and the bandwidth.
so what kind of noise do you want to measure?
- wideband noise (uVeff) (e.g. 10Hz-100kHz)
example measurement setup with a AN83 (LT) amplifier (1000:1) and a digital scope (RMS) or AC RMS millivoltmeter.
- 1/f noise (uVpp) (e.g. 0.1 - 10 Hz)
example measurement setup with a LNA (10000:1) and a scope (2mV/div resulting in 0.2uV/Div)
- popcorn noise (uVpp)(usually below 1 Hz)
longer pulses could be measured with a >= 6.5 digit DMM (if below 1uV resolution can be obtained by HPIB/RS232)
shorter can be observed with a 1/f measurement setup
of course you should always determine noise floor of your setup (e.g. measuring 8 AA NiCd/NiMH cells at constant temperature).
with best regards
Andreas
Edit: be aware that AN83 contains a bug regarding frequency response.
To get a flat frequency range you have to change R9 = 5K19
Hi Andreas,
Allready a long time i have the components on stock for building the AN83 amplifier.
You mention this:
To get a flat frequency range you have to change R9 = 5K19I cant find a R9 in this schematic, it is not the alcohol, but maybe my old age
Can you explane please.
Kind regards,
Bram
ok my fault
R9 refers to the PCB of branadic that I use.
In AN83 it is the 5.62 K resistor.
with best regards
Andreas
Re. 732A noise:
When I compared 10V between two Fluke 732A units with Solartron 7081 multimeter, Avg change of difference of voltage was about 0,155uV std dev; 0,918uV p-p/12hours.
7081 at 0.1V level has noise about 25nV std dev, so it seems that one 732A unit (assuming their noise is similar) generates about 0,108uV std dev, 0,68uV p-p/12 hours.
Integration time of single measurement was 51.2s , so this noise was for frequency (0.02 Hz. ..0,00002 Hz).