DMM Noise comparison testing project

[OldNeurons]

ManateeMafia,

I did some runs on 2001, but had some issues (unrelated to script) and still need little more time.
Hope to post summary and details this weekend.

Hello TiN,

Have you solved your problems? Can we expect to see your data?

[TiN]

50/50
Let me work on it this weekend, almost got all server stuff updated to fresh system.
Want finish and start working already on new storage RAID, so all data and links would be kept permanent.

[barnacle2k]

Any updates on this TiN?
Would be a shame to let that mountain of data rot.

[TiN]

Sorry for delay, really busy lately.

Cooked little dirty python script to collect data from ManateeMafia's EZGPIB script CSV-format and calculate RMS values automatically.

Code: [Select]

import csv
import sys
import math
import glob
from math import exp, expm1
from datetime import datetime

#with open('Keithley2002_DCVOLT1000_NPLC1.csv','r') as i:
with open('output.txt','wb') as o:
    for filename in glob.glob('*.csv'):
        i = open(filename, 'r')
        reader = csv.DictReader(i)
        #o.write ("date,Keithley2002\r\n")
        sum = 0
        sqsum = 0
        cnt = 0
        for row in reader:
            #d = datetime.strptime(row['date'], '%d-%m-%Y %H:%M:%S')
            #date = d.strftime('%Y-%m-%d %H:%M:%S')
            sum = sum + float(row['Keithley2002'])
            sqsum = sqsum + pow(float(row['Keithley2002']),2)
            cnt = cnt + 1
            avg = sum / cnt
            rms = math.sqrt(sqsum / cnt)
            #print ("Counts %d, RAW:%s SUM:%.8f AVG: %.8f" % (cnt,row['Keithley2002'],sum,avg))
        print ("%s ; Counts %d ; RAW:%s ; SQSUM:%.12f ; AVG: %E ; RMS: %E ;" % (filename,cnt,row['Keithley2002'],sqsum,avg,rms))
        i.close()
        o.write ("%s ; Counts %d ; RAW:%s ; SQSUM:%.12f ; AVG: %E ; RMS: %E ;\n" % (filename,cnt,row['Keithley2002'],sqsum,avg,rms))

It works but need some manual inputs and finetuning.
Then will need to bring other's data format to this one (which should be pretty easy) and import results into nice excel spreadsheet with graphs.

Here's example output from my Keithley 2002 EZGPIB data:

Code: [Select]

Keithley2002_DCVOLT.1_NPLC.01.csv ; Counts 25173 ; RAW:+0.001932E-03 ; SQSUM:0.000000108820 ; AVG: 1.866774E-06 ; RMS: 2.079154E-06 ;
Keithley2002_DCVOLT.1_NPLC.1.csv ; Counts 3087 ; RAW:+0.006889E-03 ; SQSUM:0.000000172460 ; AVG: 7.467759E-06 ; RMS: 7.474385E-06 ;
Keithley2002_DCVOLT.1_NPLC1.csv ; Counts 29111 ; RAW:+0.002913E-03 ; SQSUM:0.000000296672 ; AVG: 3.187090E-06 ; RMS: 3.192345E-06 ;
Keithley2002_DCVOLT.1_NPLC10.csv ; Counts 3171 ; RAW:+0.003173E-03 ; SQSUM:0.000000030584 ; AVG: 3.101818E-06 ; RMS: 3.105641E-06 ;
Keithley2002_DCVOLT1000_NPLC.01.csv ; Counts 36037 ; RAW:+0.00042E+00 ; SQSUM:0.423976726800 ; AVG: 7.061392E-04 ; RMS: 3.430020E-03 ;
Keithley2002_DCVOLT1000_NPLC.1.csv ; Counts 27187 ; RAW:-0.00007E+00 ; SQSUM:0.004349389700 ; AVG: 2.049807E-04 ; RMS: 3.999756E-04 ;
Keithley2002_DCVOLT1000_NPLC1.csv ; Counts 3085 ; RAW:+0.00017E+00 ; SQSUM:0.000122373500 ; AVG: 1.851572E-04 ; RMS: 1.991664E-04 ;
Keithley2002_DCVOLT1000_NPLC10.csv ; Counts 3146 ; RAW:+0.00018E+00 ; SQSUM:0.000123968200 ; AVG: 1.930960E-04 ; RMS: 1.985070E-04 ;
Keithley2002_DCVOLT100_NPLC.01.csv ; Counts 37259 ; RAW:+0.008344E+00 ; SQSUM:1.047291303500 ; AVG: -1.429550E-04 ; RMS: 5.301737E-03 ;
Keithley2002_DCVOLT100_NPLC.1.csv ; Counts 27343 ; RAW:+0.000114E+00 ; SQSUM:0.000865851267 ; AVG: 1.351317E-04 ; RMS: 1.779502E-04 ;
Keithley2002_DCVOLT100_NPLC1.csv ; Counts 24049 ; RAW:+0.000233E+00 ; SQSUM:0.001122555603 ; AVG: 2.130842E-04 ; RMS: 2.160506E-04 ;
Keithley2002_DCVOLT100_NPLC10.csv ; Counts 3155 ; RAW:+0.000228E+00 ; SQSUM:0.000136496793 ; AVG: 2.070285E-04 ; RMS: 2.079991E-04 ;
Keithley2002_DCVOLT10_NPLC.01.csv ; Counts 37220 ; RAW:+0.0000020E+00 ; SQSUM:0.005388871941 ; AVG: 1.830596E-04 ; RMS: 3.805053E-04 ;
Keithley2002_DCVOLT10_NPLC.1.csv ; Counts 6671 ; RAW:+0.0000119E+00 ; SQSUM:0.000001456198 ; AVG: 1.244638E-05 ; RMS: 1.477457E-05 ;
Keithley2002_DCVOLT10_NPLC1.csv ; Counts 32311 ; RAW:+0.0000139E+00 ; SQSUM:0.000003083435 ; AVG: 9.409851E-06 ; RMS: 9.768822E-06 ;
Keithley2002_DCVOLT10_NPLC10.csv ; Counts 3080 ; RAW:+0.0000012E+00 ; SQSUM:0.000000008606 ; AVG: 1.401396E-06 ; RMS: 1.671587E-06 ;
Keithley2002_DCVOLT1_NPLC.01.csv ; Counts 29560 ; RAW:+0.00000191E+00 ; SQSUM:0.000022246224 ; AVG: 5.251152E-06 ; RMS: 2.743317E-05 ;
Keithley2002_DCVOLT1_NPLC.1.csv ; Counts 3086 ; RAW:+0.00000786E+00 ; SQSUM:0.000000169236 ; AVG: 7.365016E-06 ; RMS: 7.405402E-06 ;
Keithley2002_DCVOLT1_NPLC1.csv ; Counts 6726 ; RAW:+0.00000300E+00 ; SQSUM:0.000000052482 ; AVG: 2.784019E-06 ; RMS: 2.793349E-06 ;
Keithley2002_DCVOLT1_NPLC10.csv ; Counts 3080 ; RAW:+0.00000317E+00 ; SQSUM:0.000000026442 ; AVG: 2.924023E-06 ; RMS: 2.930037E-06 ;
MM's K2002:

Code: [Select]

Keithley2002_DCVOLT.2_NPLC.01.csv ; Counts 28500 ; RAW:+0.012987E-03 ; SQSUM:0.000001055413 ; AVG: 3.677938E-06 ; RMS: 6.085394E-06 ;
Keithley2002_DCVOLT.2_NPLC.1.csv ; Counts 28500 ; RAW:-0.000484E-03 ; SQSUM:0.000000002879 ; AVG: 8.280295E-08 ; RMS: 3.178376E-07 ;
Keithley2002_DCVOLT.2_NPLC1.csv ; Counts 28500 ; RAW:+0.001461E-03 ; SQSUM:0.000000050083 ; AVG: 1.319546E-06 ; RMS: 1.325635E-06 ;
Keithley2002_DCVOLT.2_NPLC10.csv ; Counts 1190 ; RAW:+0.001523E-03 ; SQSUM:0.000000001753 ; AVG: 1.209540E-06 ; RMS: 1.213718E-06 ;
Keithley2002_DCVOLT1000_NPLC.01.csv ; Counts 28750 ; RAW:+0.00110E+00 ; SQSUM:0.060372603700 ; AVG: 1.076401E-03 ; RMS: 1.449109E-03 ;
Keithley2002_DCVOLT1000_NPLC.1.csv ; Counts 28500 ; RAW:-0.00038E+00 ; SQSUM:0.004084380200 ; AVG: 1.501102E-04 ; RMS: 3.785652E-04 ;
Keithley2002_DCVOLT1000_NPLC1.csv ; Counts 28500 ; RAW:+0.00016E+00 ; SQSUM:0.000835754000 ; AVG: 1.568912E-04 ; RMS: 1.712446E-04 ;
Keithley2002_DCVOLT1000_NPLC10.csv ; Counts 1190 ; RAW:+0.00020E+00 ; SQSUM:0.000025156800 ; AVG: 1.392437E-04 ; RMS: 1.453966E-04 ;
Keithley2002_DCVOLT200_NPLC.01.csv ; Counts 28500 ; RAW:+0.001165E+00 ; SQSUM:0.522674665140 ; AVG: 3.491382E-05 ; RMS: 4.282460E-03 ;
Keithley2002_DCVOLT200_NPLC.1.csv ; Counts 28500 ; RAW:+0.000166E+00 ; SQSUM:0.000580846443 ; AVG: 3.007274E-05 ; RMS: 1.427606E-04 ;
Keithley2002_DCVOLT200_NPLC1.csv ; Counts 28500 ; RAW:+0.000166E+00 ; SQSUM:0.000487367848 ; AVG: 1.260648E-04 ; RMS: 1.307694E-04 ;
Keithley2002_DCVOLT200_NPLC10.csv ; Counts 1190 ; RAW:+0.000117E+00 ; SQSUM:0.000014781988 ; AVG: 1.099160E-04 ; RMS: 1.114533E-04 ;
Keithley2002_DCVOLT20_NPLC.01.csv ; Counts 28250 ; RAW:+0.0000094E+00 ; SQSUM:0.000002496170 ; AVG: 9.400000E-06 ; RMS: 9.400000E-06 ;
Keithley2002_DCVOLT20_NPLC.1.csv ; Counts 28250 ; RAW:+0.0000094E+00 ; SQSUM:0.000001865081 ; AVG: 5.507855E-06 ; RMS: 8.125305E-06 ;
Keithley2002_DCVOLT20_NPLC1.csv ; Counts 28250 ; RAW:+0.0000014E+00 ; SQSUM:0.000000234585 ; AVG: 2.560676E-06 ; RMS: 2.881647E-06 ;
Keithley2002_DCVOLT20_NPLC10.csv ; Counts 1190 ; RAW:+0.0000021E+00 ; SQSUM:0.000000005694 ; AVG: 2.013445E-06 ; RMS: 2.187497E-06 ;
Keithley2002_DCVOLT2_NPLC.01.csv ; Counts 28000 ; RAW:+0.00000480E+00 ; SQSUM:0.000080723282 ; AVG: 8.961264E-07 ; RMS: 5.369334E-05 ;
Keithley2002_DCVOLT2_NPLC.1.csv ; Counts 28000 ; RAW:-0.00000018E+00 ; SQSUM:0.000000021068 ; AVG: 1.519464E-07 ; RMS: 8.674316E-07 ;
Keithley2002_DCVOLT2_NPLC1.csv ; Counts 28250 ; RAW:+0.00000111E+00 ; SQSUM:0.000000040999 ; AVG: 1.182998E-06 ; RMS: 1.204699E-06 ;
Keithley2002_DCVOLT2_NPLC10.csv ; Counts 1190 ; RAW:+0.00000106E+00 ; SQSUM:0.000000001203 ; AVG: 9.911765E-07 ; RMS: 1.005444E-06 ;

Here MM's 3458A_ID111 data:

Code: [Select]

HP3458A__ID111_DCV.1_NPLC.01.csv ; Counts 43366 ; RAW:1.934592131E-06 ; SQSUM:0.000000075188 ; AVG: -1.572004E-07 ; RMS: 1.316741E-06 ;
HP3458A__ID111_DCV.1_NPLC.1.csv ; Counts 37203 ; RAW:5.277066847E-07 ; SQSUM:0.000000007471 ; AVG: 1.015909E-07 ; RMS: 4.481187E-07 ;
HP3458A__ID111_DCV.1_NPLC1.csv ; Counts 26937 ; RAW:-1.160974125E-07 ; SQSUM:0.000000001669 ; AVG: -1.766057E-07 ; RMS: 2.489070E-07 ;
HP3458A__ID111_DCV.1_NPLC10.csv ; Counts 4909 ; RAW:1.409000080E-07 ; SQSUM:0.000000000104 ; AVG: -1.177160E-07 ; RMS: 1.453049E-07 ;
HP3458A__ID111_DCV.1_NPLC100.csv ; Counts 444 ; RAW:-9.322971815E-08 ; SQSUM:0.000000000006 ; AVG: -1.081460E-07 ; RMS: 1.118537E-07 ;
HP3458A__ID111_DCV.1_NPLC1000.csv ; Counts 54 ; RAW:-1.465289721E-07 ; SQSUM:0.000000000004 ; AVG: -2.635645E-07 ; RMS: 2.857508E-07 ;
HP3458A__ID111_DCV.1_NPLC50.csv ; Counts 1050 ; RAW:-1.359746644E-07 ; SQSUM:0.000000000017 ; AVG: -1.218833E-07 ; RMS: 1.284038E-07 ;
HP3458A__ID111_DCV1000_NPLC.01.csv ; Counts 35418 ; RAW:-1.760964367E-03 ; SQSUM:0.247285785299 ; AVG: 1.443854E-04 ; RMS: 2.642333E-03 ;
HP3458A__ID111_DCV1000_NPLC.1.csv ; Counts 31322 ; RAW:1.761266916E-04 ; SQSUM:0.003887161804 ; AVG: 1.327557E-04 ; RMS: 3.522829E-04 ;
HP3458A__ID111_DCV1000_NPLC1.csv ; Counts 21212 ; RAW:1.585166738E-04 ; SQSUM:0.000154840219 ; AVG: 4.540910E-05 ; RMS: 8.543800E-05 ;
HP3458A__ID111_DCV1000_NPLC10.csv ; Counts 4696 ; RAW:-1.109616452E-04 ; SQSUM:0.000026941396 ; AVG: -6.692175E-05 ; RMS: 7.574361E-05 ;
HP3458A__ID111_DCV1000_NPLC100.csv ; Counts 517 ; RAW:-7.045183825E-05 ; SQSUM:0.000001151042 ; AVG: -4.487046E-05 ; RMS: 4.718461E-05 ;
HP3458A__ID111_DCV1000_NPLC1000.csv ; Counts 53 ; RAW:-4.579369486E-05 ; SQSUM:0.000000095233 ; AVG: -4.064273E-05 ; RMS: 4.238933E-05 ;
HP3458A__ID111_DCV1000_NPLC50.csv ; Counts 1019 ; RAW:-7.221313421E-05 ; SQSUM:0.000002797342 ; AVG: -4.867849E-05 ; RMS: 5.239450E-05 ;
HP3458A__ID111_DCV100_NPLC.01.csv ; Counts 31954 ; RAW:-1.759903619E-04 ; SQSUM:0.012203052480 ; AVG: 2.890893E-04 ; RMS: 6.179760E-04 ;
HP3458A__ID111_DCV100_NPLC.1.csv ; Counts 31054 ; RAW:1.408164788E-04 ; SQSUM:0.003746873958 ; AVG: 3.253337E-04 ; RMS: 3.473568E-04 ;
HP3458A__ID111_DCV100_NPLC1.csv ; Counts 21556 ; RAW:3.696494398E-05 ; SQSUM:0.000056098800 ; AVG: 3.553526E-05 ; RMS: 5.101439E-05 ;
HP3458A__ID111_DCV100_NPLC10.csv ; Counts 4696 ; RAW:-8.625151540E-06 ; SQSUM:0.000001555449 ; AVG: -1.120778E-05 ; RMS: 1.819969E-05 ;
HP3458A__ID111_DCV100_NPLC100.csv ; Counts 517 ; RAW:6.512869530E-06 ; SQSUM:0.000000010226 ; AVG: 1.896423E-07 ; RMS: 4.447513E-06 ;
HP3458A__ID111_DCV100_NPLC1000.csv ; Counts 53 ; RAW:5.280705025E-07 ; SQSUM:0.000000000970 ; AVG: 3.742990E-06 ; RMS: 4.277578E-06 ;
HP3458A__ID111_DCV100_NPLC50.csv ; Counts 1019 ; RAW:-1.179357455E-05 ; SQSUM:0.000000048509 ; AVG: -2.418380E-06 ; RMS: 6.899579E-06 ;
HP3458A__ID111_DCV10_NPLC.01.csv ; Counts 44629 ; RAW:5.282965900E-05 ; SQSUM:0.000033473378 ; AVG: 2.829167E-07 ; RMS: 2.738679E-05 ;
HP3458A__ID111_DCV10_NPLC.1.csv ; Counts 40296 ; RAW:-3.522582372E-06 ; SQSUM:0.000001245022 ; AVG: -4.628066E-06 ; RMS: 5.558499E-06 ;
HP3458A__ID111_DCV10_NPLC1.csv ; Counts 21586 ; RAW:1.409056517E-06 ; SQSUM:0.000000034038 ; AVG: 1.080365E-06 ; RMS: 1.255721E-06 ;
HP3458A__ID111_DCV10_NPLC10.csv ; Counts 4910 ; RAW:-2.113584272E-07 ; SQSUM:0.000000000902 ; AVG: 1.639140E-07 ; RMS: 4.285837E-07 ;
HP3458A__ID111_DCV10_NPLC100.csv ; Counts 367 ; RAW:-5.283960681E-08 ; SQSUM:0.000000000004 ; AVG: -1.665335E-08 ; RMS: 1.080092E-07 ;
HP3458A__ID111_DCV10_NPLC1000.csv ; Counts 54 ; RAW:-5.283960681E-08 ; SQSUM:0.000000000000 ; AVG: -5.707982E-08 ; RMS: 7.202534E-08 ;
HP3458A__ID111_DCV10_NPLC50.csv ; Counts 1059 ; RAW:1.409056182E-07 ; SQSUM:0.000000000033 ; AVG: -1.184525E-07 ; RMS: 1.768822E-07 ;
HP3458A__ID111_DCV1_NPLC.01.csv ; Counts 39630 ; RAW:-3.519855741E-06 ; SQSUM:0.000000382711 ; AVG: 2.404746E-07 ; RMS: 3.107589E-06 ;
HP3458A__ID111_DCV1_NPLC.1.csv ; Counts 26059 ; RAW:3.520460482E-07 ; SQSUM:0.000000007892 ; AVG: 5.182273E-08 ; RMS: 5.503157E-07 ;
HP3458A__ID111_DCV1_NPLC1.csv ; Counts 26998 ; RAW:2.640389526E-07 ; SQSUM:0.000000000876 ; AVG: 6.759768E-08 ; RMS: 1.801146E-07 ;
HP3458A__ID111_DCV1_NPLC10.csv ; Counts 4909 ; RAW:-2.763607045E-07 ; SQSUM:0.000000000155 ; AVG: -1.511198E-07 ; RMS: 1.779424E-07 ;
HP3458A__ID111_DCV1_NPLC100.csv ; Counts 365 ; RAW:-5.984881499E-08 ; SQSUM:0.000000000005 ; AVG: -1.088564E-07 ; RMS: 1.139828E-07 ;
HP3458A__ID111_DCV1_NPLC1000.csv ; Counts 54 ; RAW:-4.224622234E-08 ; SQSUM:0.000000000000 ; AVG: -7.455024E-08 ; RMS: 7.752065E-08 ;
HP3458A__ID111_DCV1_NPLC50.csv ; Counts 1059 ; RAW:2.640388897E-08 ; SQSUM:0.000000000010 ; AVG: -8.820412E-08 ; RMS: 9.886552E-08 ;


[TiN]

Initial data processed, now doing graphs for each range.

Starting off 0.1(0.2)VDC:

[TiN]

Summary results in range DCV 100mV (200mV)



Summary results in range DCV 1V (2V)



Summary results in range DCV 10V (20V)



Summary results in range DCV 100V (200V or 300V in case of HPAK 34970A)



Summary results in range DCV 1000V



Article with scripts and EZGPIB setup

Excel-file with all data and graphs

[Dr. Frank]

Hi TiN,

well done, great job! 

Very interesting to see these direct comparisons.

Frank

[3roomlab]

i noticed in some of the values ... some results from some benches seem consistently higher in just a bit of noise
just to speculate ...

what if ambient noise added alot of undesirable result to the figures? (example manatee's 3458a vs Dr franks) ... or maybe a noisy cell phone tower nearby? HV power line? a bad PC PSU?

if so, what if the contributor switched off everything and re-collect the results? or used something like a RF-sniffer to "probe" out and eliminate some equipment that is inputting some spurious readings?
(or is there a way to measure ambient noise ? like dosimeter measure ambient radiation? and note down that this measurement is taken with this amount of ambient noise during the measurement process?)

just a thought

[ManateeMafia]

I have no doubt that my tests were made in less than perfect conditions.

I think Dr. Frank has a  more suitable environment compared to my home. I did cover the connections to reduce air flow across the inputs. I should add the ability to track the internal temp of the 3458A and perform an ACAL DC as needed. I performed the tests in a spare room but I had no way to monitor external influences.

Running the acquisition at night time is the best time for me as I am not playing with the thermostat and interfering in other measureble ways.

I could disconnect wi-fi and my phone repeater. All CFL's will have to be disconnected too.

[TiN]

All data is still considered beta and initial, so yes, some results and tests are off.
I do hope eventually we will get statistics to work for us, and when let's say five different members run same test on same multimeter model, we should get pretty close results. If one of datasets would be off to much - we would sort that out. This way we can have confidence and some degree of trust in data.

But since that will take time and more wide participation, we will try do best to make everything EASY to test and maintain. That's main goal for now.

Good points tho, I'll add more information regarding testing environment into project article. Perhaps even try some comparison tests on same meter with different conditions (noisy switcher PSU on same breaker, CFL lights all around meter, cell phone laying on top of meter, etc).

[Dr. Frank]

I have no doubt that my tests were made in less than perfect conditions.

I think Dr. Frank has a  more suitable environment compared to my home. I did cover the connections to reduce air flow across the inputs. I should add the ability to track the internal temp of the 3458A and perform an ACAL DC as needed. I performed the tests in a spare room but I had no way to monitor external influences.

Running the acquisition at night time is the best time for me as I am not playing with the thermostat and interfering in other measureble ways.

I could disconnect wi-fi and my phone repeater. All CFL's will have to be disconnected too.

Maybe..

My basement has a very constant temperature, also during the current heat wave outside. Outside: 40°C, basement: 22°C max! Very comfortable, best place to be!

All HF sources as switching PSUs , cell phones etc were abandoned, the concrete ceiling is full of iron armouring, shielding every radiation from outside.

But the most probable reason is, that I configured the 3458A for digitizing, by switching Auto Zero off. I think I mentioned already, that especially the fast readings, NPLC 0.01 = 200µs were done in real time.

cmd:= 'DCV 1;AZERO OFF;DISP OFF;NPLC 0.01;TRIG SYN;TARM AUTO;INBUF ON;NRDGS 10000,AUTO;';

I suppose, that the auto zeroing might create additional noise.. to be checked..
Frank

[TiN]

Okay, I did some testing here regarding one particular issue:

As lot of homelab user instruments are usually not commonly calibrated, there is issue of having possible DC offsets during measurement of input short. These offsets translate to noise RMS offset as well. Meaning that measurement is not noise itself but noise+offset.

I tried to modify script and run on same unit in same conditions, one with just sampling data as it's described before (No REL), and second dataset - performing REL capture of initial DC voltage sample and then capturing data with offset removed by meter's math. Then for next NPLC+range new REL value taken and everything repeat. This is (REL) graph line.

And here's result:











If we decide to go this way, it could be V2.0 script naming, as data would be not comparable to current V1.0

[macboy]

Okay, I did some testing here regarding one particular issue:

As lot of homelab user instruments are usually not commonly calibrated, there is issue of having possible DC offsets during measurement of input short. These offsets translate to noise RMS offset as well. Meaning that measurement is not noise itself but noise+offset.

If you want to eliminate the offset, take the standard deviation of the dataset, not the RMS value. This removes the offset by the definition of standard deviation, leaving only the RMS noise compared to the average value (i.e. offset).

[TiN]

  You right, wasn't thinking right.

Changed python parser and added test data with both 2002 and 2001 to check SD:





Red line = K2002, raw RMS ppm/Range, no REL
Green line = K2002, raw RMS ppm/Range, with REL
Blue line = K2002, same dataset as red line, but with SD ppm/Range instead of RMS
Dashed orange line = K2001, SD ppm/Range, no REL

[3roomlab]

I have no doubt that my tests were made in less than perfect conditions.

I think Dr. Frank has a  more suitable environment compared to my home. I did cover the connections to reduce air flow across the inputs. I should add the ability to track the internal temp of the 3458A and perform an ACAL DC as needed. I performed the tests in a spare room but I had no way to monitor external influences.

Running the acquisition at night time is the best time for me as I am not playing with the thermostat and interfering in other measureble ways.

I could disconnect wi-fi and my phone repeater. All CFL's will have to be disconnected too.

i am not trying to nit pick, i hope i did not sound that way, i apologize if i did. i think many of our surroundings are in a middle of a dense electronic jungle, cant be helped (i for 1 is surrounded by 20 blasting wifi APs, i cant turn those off for sure  ). if i have a 6.5 digit measuring these, im sure my ambient noise would be far worse, my T5 lights up like a radio station

[ManateeMafia]

I didn't take it as a nit-pick. I will setup a 3458A and run the tests again with Dr. Frank's suggested settings for a like-for-like comparison. I will also try and remove anything else nearby and see how much difference it makes.

I hope to get some different readings by this weekend.

[lukaq]

Could someone make script for EZGPIB to automate data gathering for 34401a via RS232?

I have user guide if some needs it for what are the commands

[cncjerry]

Tin, couple of questions:

1) Do you have data for an HP 3457a?

2) I started making some runs just for the heck of it.  According to the instructions I should be using the 10v range which on this unit implies 30v.

3) as far as NPLC, I started with 10 and did hi-res reads.  There is a big difference between the noise at 30v and 3v.  Also, taking NPLC to 100 improved things from two perspectives, the resolution increased and noise dropped.

So to get a fair comparison against the other units being tested, where should I set the 3457a?

Interesting data so far.

Thanks.

Jerry

[TiN]

1. No, I don't. All data is already posted on graphs/article. Will be happy to see your results on 3457A, and any other meters.
2. Yes, use base ranges which you meter have. It's not matter much what range voltage exactly, and end result is ppm/Range, not ppm/Volt.

If you have GPIB interfacing setup, you can try modify 3458A's EZGPIB script, so you can run all NPLC and ranges automatically.

[cncjerry]

Tin, how are you calculating PPM from this data?  I'll admit to not reading the entire thread again, so if it is in there, please just say, "read the thread."

I ran some tests with my 3457a in hi-res.  For example, the STDEV on the various ranges I've tested so far are below:

Range   30v                30V                  3V                3V                300mv           30mv             30mv
NPLC    10                  100                  10                100               10                  10                 100
STDEV  1.32596E-5     4.18793E-6       3.4401E-7    1.7943E-7     1.56618E-7      1.31822E-7    1.67786E-7

Are do you calculate PPM from this?  Also, with the data I have so far, increasing the NPLC from 10 to 100 increases the STDEV except for the 3v scale.  Seems odd.

Thanks

[deadlylover]

Remember the E is multiply by 10^x, so your results look okay. To get PPM I *think* you just divide the STDEV by the range (in volts), then multiply by 1 million.

The 3V range is the primary range on the 3457A, I'm still waiting for a GPIB cable to arrive but I'll use the STAT function to see if it kinda sorta agrees with yours.

I'm sure there are others who want to see the R6581 dogfight with the 3458A. Well the results aren't reeeaally going to surprise anyone, it would be like the Top Gear episode where the McLaren F1 went up against the Bugatti Veyron. 

[TiN]

You can read my article here regarding methodology and calculations.
ppm/Range = SD/range, simple as it is.

[cncjerry]

Stdev/range or (stdev/range) x 1e6?

[ManateeMafia]

Could someone make script for EZGPIB to automate data gathering for 34401a via RS232?

I have user guide if some needs it for what are the commands

I hope to get my hands on a 34401a sometime next week. I thought about putting my 8846a in compatibility mode, but it should be easier with the real thing.

[Jf2014]

hi,
I uploaded the files for the Keithley 2001 ( calibriert 07/2015)  to TiN's FTP server.
ftp://xdevs.com//Jf2014