DMM Noise comparison testing project

[Jorn]

Attached is the small script that generates the data for DMM noise floor measurements. Beware, it is an early crude version to be improved...

I use Jupyter Notebook http://jupyter.org/ for python execution. It's perfect for instrument control due to the interactive behavior of the jupyter notebook.

The script uses the pyvisa library because I use a visa compatible GPIB adapter (F82357 from BEIMING technologies).
If a prologix adapter is used consider trying the pyserial library instead of pyvisa.

The script is attached in 3 different versions (in one zip file):
1) The .ipynb is the Jupyter Notebook binary (don't open unless you run Jupyter)
2) The .py is the exported python file. This file should be able to run (with very few changes) in an ordinary Python 3 environment with numpy, pyvisa and matplotlib installed.
3) The last .html file shows the jupyter environment with all the python code mixed with instrument data and graphs - the Jupyter way of programming... 

-jorn 

[alanambrose]

Ah at last I got a couple of hours to debug my 34420A code. Results below:

http://anagram.net/Nuts/MeterNoiseFloor/HP34420A noise floor v2.xlsx

[alanambrose]

OK here are results for a K2182A. This is no filters at all, no averaging, autozero on, line sync on. Also, the front-end block schematic below. The reported temp variation from the internal sensor was <0.2C for each run - so the slight upward sloping on some ranges > 5nplc is curious. Also, the channel asymmetry for 1V and 100mV ranges < 1nplc is interesting - the spec does detail approx 40% more noise on channel 2.

Alan

http://anagram.net/nuts/2182A/K2182A%20noise%20floor.xlsx

[TiN]

I'll just leave this in face of DMM7510 owners who did not do homework 

Comparison of DCV shorted inputs noise on two 3458A in same conditions.

[ManateeMafia]

Do you have time to try it again with a stable reference (3245A) on the inputs? Of course that would leave out readings from the 1kV range.

[TiN]

As of noise test with 3245A actually found recently good article on interwebs about testing LDO noise using 3458A as sampling box.
Output result is provided as 0.1Hz - 10Hz spectrum. I'm interested to write small app in python for it and check how good it works.

[VintageNut]

Hello TiN

I am not following your comment about the DMM7510. I read back to in this thread for the comparison of the 7510 with the 3458A and do not see any large difference. Please give a brief summary/conclusion under what conditions there is a large difference and how much the difference is.

[TiN]

It was more a poke for all (rightfully busy) people with DMM7510's who had no time to run noise test script on the box 

[dr.diesel]

You mean aside from me, or did I miss a new challenge? 

[bingo600]

@TIN

If you are using something linux compatible (linuxgpib) , to measure your 3458 with , could you upload it here ?
Or if you convert JORN's python to linuxgpib & 3458 settings ?

Then i'm ready to test my 58' Opt-02

/Bingo

[TiN]

I had bits and pieces written in python to do same test using Pi, but nothing functional.
So to run these tests I still plug NI USB-GPIB-HS into my Windows WS and run EZGPIB, as listed in the article.
Sorry. 

[VintageNut]

It was more a poke for all (rightfully busy) people with DMM7510's who had no time to run noise test script on the box 

Where is this magical script? I have python 2.7 and pyvisa running on two different Win 7 machines.

[TiN]

It's EZGPIB. Details here.

[VintageNut]

It's EZGPIB. Details here.

Thanks. I downloaded the program. Is there a script for the DMM7510?

[VintageNut]

Here is maybe enough data to give a comparison. My short plug is a crappy banana short. I will try a piece of hookup wire to see if that improves the results.

      DMM7510      
            
Range   NPLC   FILT   STD DEV   READINGS
            
100mV   10   10   24.3nV   60
100mV   10      off   35.9nV   600
100mV   1      off   39.1nV   6000
            
1V   10   10   32.7nV   60
1V   10      off   65.6nV   600
1V   1      off   77.3nV   6000
            
10V   10   10   216nV   60
10V   10      off   344nV   600
10V   1      off   665nV   6000

[VintageNut]

No significant change using a bare solid copper wire.

Here is maybe enough data to give a comparison. My short plug is a crappy banana short. I will try a piece of hookup wire to see if that improves the results.

      DMM7510      
            
Range   NPLC   FILT   STD DEV   READINGS
            
100mV   10   10   24.3nV   60
100mV   10      off   35.9nV   600
100mV   1      off   39.1nV   6000
            
1V   10   10   32.7nV   60
1V   10      off   65.6nV   600
1V   1      off   77.3nV   6000
            
10V   10   10   216nV   60
10V   10      off   344nV   600
10V   1      off   665nV   6000

[VintageNut]

10V range 1000PLC 19 readings, standard deviation  85 nV. I will let this run all week and see what happens.

Here is maybe enough data to give a comparison. My short plug is a crappy banana short. I will try a piece of hookup wire to see if that improves the results.

      DMM7510      
            
Range   NPLC   FILT   STD DEV   READINGS
            
100mV   10   10   24.3nV   60
100mV   10      off   35.9nV   600
100mV   1      off   39.1nV   6000
            
1V   10   10   32.7nV   60
1V   10      off   65.6nV   600
1V   1      off   77.3nV   6000
            
10V   10   10   216nV   60
10V   10      off   344nV   600
10V   1      off   665nV   6000

[dr.diesel]

10V range 1000PLC 19 readings, standard deviation  85 nV. I will let this run all week and see what happens.

Did you change instruments?  7510 has a max NPLC of 15.

[VintageNut]

10V range 1000PLC 19 readings, standard deviation  85 nV. I will let this run all week and see what happens.

Did you change instruments?  7510 has a max NPLC of 15.

NPLC 10, Digital filter 100. This is similar to the 3458A which has no aperture greater than 1 second ( if I recall the number correctly from reading in this forum). Everything beyond that is digital filtering. 

[dr.diesel]

NPLC 10, Digital filter 100. This is similar to the 3458A which has no aperture greater than 1 second ( if I recall the number correctly from reading in this forum). Everything beyond that is digital filtering.

Mmmmmmm.  If that is an accepted standard with the rest of the folks here then I will do the same for another data point.

[HighVoltage]

10V range 1000PLC 19 readings, standard deviation  85 nV. I will let this run all week and see what happens.

Did you change instruments?  7510 has a max NPLC of 15.

Are you sure about Keithley DMM7510 to have a max NPLC of 15
If I remember right, the max NPLC value I could set in DC Volts was 12

[VintageNut]

Yes 15 NPLC.

[HighVoltage]

Hmm, how do you set 15 NPLC on the 7510?
I am getting an error above 12

[VintageNut]

Do you have the latest firmware installed?

[HighVoltage]

Do you have the latest firmware installed?

Yes, latest FW 1.6.1 is installed

Really interesting, that you can set yours to 15.