Author Topic: DMM Noise comparison testing project (Read 217238 times)

dr.diesel · « **Reply #225 on:** December 15, 2015, 11:05:28 am »

Quote from: TiN on December 15, 2015, 04:13:24 am

I think there is need to revamp my python knowledge to make it more useful though.
Would be nice if you just follow existing format for now.
What you will run to get GPIB data? I can try to make an script for 7510, but haven't read it's docs yet.

No problem I will use existing format, poll at 1 sec. I don't have a Windows box so can't use the EZGPIB tool, I will just use my own and fetch over TCP.

Output of python script on my k2010 and k2450:

Code: [Select]

| Multimeter 		| Range		| NPLC	| 	Counts	|    Sq.sum	|  Average	|    SD		|    SD,uV	|
| k2010 		| 10 		| .01 	|	120	| 0.00427245	| 3.560375E-05	| 9.487653E-05	|  94.877	|
| k2010 		| 10 		| .1 	|	120	| -0.00019567	| -1.630594E-06	| 1.470821E-05	|  14.708	|
| k2010 		| 10 		| 10 	|	360	| -0.00070481	| -1.957792E-06	| 5.162154E-07	|   0.516	|
| k2010 		| 10 		| 1 	|	120	| -0.00023459	| -1.954949E-06	| 1.273443E-06	|   1.273	|
| Multimeter 		| Range		| NPLC	| 	Counts	|    Sq.sum	|  Average	|    SD		|    SD,uV	|
| k2450 		| 2 		| 1 	|	121	| -0.00503126	| -4.158066E-05	| 3.858809E-06	|   3.859	|
| k2450 		| 2 		| 10 	|	121	| -0.00469228	| -3.877919E-05	| 3.558100E-06	|   3.558	|
| k2450 		| 2 		| .1 	|	121	| -0.00864380	| -7.143638E-05	| 2.372225E-05	|  23.722	|
| k2450 		| 2 		| .01 	|	121	| -0.04565281	| -3.772959E-04	| 5.115046E-05	|  51.150	|

Vgkid · « **Reply #226 on:** December 16, 2015, 02:29:09 am »

Lookingforward to your K7510 update Dr.diesel.

Dr. Frank · « **Reply #227 on:** December 16, 2015, 10:01:17 am »

Quote from: deadlylover on December 15, 2015, 09:43:17 am

I think it is done more simply than the 3458A, some clues are in the 24hr range error specification. From 10V>1V>0.1V the range error is 0.1>1>10 ppm. I don't know if the difference are from a worse ADC, or different Auto Cal scheme, or both.

If you have any free time with the 3458A, can you check the error from applying 100mVDC to the 10V range and then reading it from the 100mV range? I get a difference of about 2-5ppm of reading even after an Auto Cal.

Hello,

I won't do that test again .. just a look into the hp journal 4/89 and its specifications is necessary to explain the problem.

The 3458A can do accurate 10:1 transfers, i.e. from one range to the other, on the order of 0.3ppm.

That means, 10V and 100mV gain should agree within 1ppm.
That is specified much more conservatively, i.e. the transfer 10V > 1V > 100mV is 1ppm each. The +0.3 and + 3 ppm / range can probably be assigned to offset errors.

Therefore, if one encounters bigger differences than 2 ppm between 100mV readings on the 10V and 100mV ranges in practise, it is also very probably due to such offset voltage effects.
If you have the possibility your measurement setup to cancel offsets, then you may get better results.

This is not a trivial aspect, as you often can not do precise zero measurements in your circuitry..

If I re-calibrate the 1V and 100mV ranges of my 5442A standard, I indeed have the possibility to output 0V first, null the offset, and then measure the 100mV output.
So I always achieve transfer accuracies < 1ppm even in the 100mV range..

Same scheme applies, when I measure 1V on the 10V range and 100mV on the 1V range, to check the foregoing AutoCal.
Each time I find, that the transfer are precise to << 1ppm, provided I had cancelled offsets before.

You may also experiment in that direction, if you own a stable, programmable DC voltage source.

Frank

Edit, because of an additional aspect!

A direct transfer 10V => 100mV range is about 10 times more unprecise than the path 10V => 1V => 100mV.

That's due to the linearity specification of typ. 0.02ppm of F.S. for the 3458A.
That gives about 0.3ppm for each 10:1 transfer, equivalently on a measurement at 1/10 of F.S., but for a 100:1 transfer, that's already 3ppm!

So, even if you properly cancel all offsets, 2-5 ppm error is to be expected, depending on the linearity of your R6581.

3roomlab · « **Reply #228 on:** December 16, 2015, 10:04:49 am »

i just have to satisfy my logging obsession

... esp after a sparodic few days of troubleshoot (partially) and fix and "re-boot" up the K2015. although the DMM works, there is still a lingering fault in the form of a noise going around the DCV/ohm sense section. in the time of repairing, the fried AD822/LTC1050 is also replaced by OPA2140 + LTC2057. it should be noted the noise is very prominent in the 0.1v/1v/10v range, but lesser in the 100v/1000v range, but even so, it is still able to run out some very interesting logs. (the pic probably need to be open wide in its own window, the newer log are the bottom half, compared to old logs of AD822+ LTC1050)

the most interesting to note is that, after the opamp change, the 1NPLC STDEV has become a rival to it own old 10NPLC reading.

dr.diesel · « **Reply #229 on:** December 16, 2015, 10:56:41 am »

Quote from: Vgkid on December 16, 2015, 02:29:09 am

Lookingforward to your K7510 update Dr.diesel.

Wait no longer! Although unit was only out of the box for 2 hours last night when I started the run.

Code: [Select]

| Multimeter 		| Range		| NPLC	| 	Counts	|    Sq.sum	|  Average	|    SD		|    SD,uV	|
| k7510 		| .1 		| .01 	|	1800	| -0.00946209	| -5.256717E-06	| 7.585452E-06	|   7.585	|
| k7510 		| .1 		| .1 	|	1799	| -0.00845946	| -4.702313E-06	| 4.205452E-06	|   4.205	|
| k7510 		| .1 		| 1 	|	1801	| -0.00984773	| -5.467925E-06	| 4.684346E-07	|   0.468	|
| k7510 		| .1 		| 10 	|	1801	| -0.00976237	| -5.420526E-06	| 1.885766E-07	|   0.189	|
| k7510 		| 1 		| .01 	|	1801	| -0.00793555	| -4.406191E-06	| 7.629865E-06	|   7.630	|
| k7510 		| 1 		| .1 	|	1799	| -0.00784621	| -4.361427E-06	| 3.954875E-06	|   3.955	|
| k7510 		| 1 		| 1 	|	1801	| -0.00977324	| -5.426563E-06	| 4.579941E-07	|   0.458	|
| k7510 		| 1 		| 10 	|	1801	| -0.00979378	| -5.437968E-06	| 1.943267E-07	|   0.194	|
| k7510 		| 10 		| .01 	|	1801	| 0.00635690	| 3.529649E-06	| 1.290502E-05	|  12.905	|
| k7510 		| 10 		| .1 	|	1799	| -0.00104714	| -5.820696E-07	| 4.829904E-06	|   4.830	|
| k7510 		| 10 		| 1 	|	1801	| -0.00204352	| -1.134657E-06	| 7.491898E-07	|   0.749	|
| k7510 		| 10 		| 10 	|	1801	| -0.00203306	| -1.128850E-06	| 4.907266E-07	|   0.491	|
| k7510 		| 100 		| .01 	|	1801	| -0.09902185	| -5.498160E-05	| 1.346015E-03	|1346.015	|
| k7510 		| 100 		| .1 	|	1799	| -0.20888873	| -1.161138E-04	| 1.807134E-04	| 180.713	|
| k7510 		| 100 		| 1 	|	1801	| -0.31506582	| -1.749394E-04	| 5.533520E-05	|  55.335	|
| k7510 		| 100 		| 10 	|	1801	| -0.32145543	| -1.784872E-04	| 1.786285E-05	|  17.863	|
| k7510 		| 1000 		| .01 	|	1801	| 1.33690515	| 7.423127E-04	| 1.674120E-03	|1674.120	|
| k7510 		| 1000 		| .1 	|	1799	| -0.16155031	| -8.980006E-05	| 2.495570E-04	| 249.557	|
| k7510 		| 1000 		| 1 	|	1801	| -0.35094636	| -1.948619E-04	| 8.040639E-05	|  80.406	|
| k7510 		| 1000 		| 10 	|	1802	| -0.35127405	| -1.949357E-04	| 5.129073E-05	|  51.291	|

dr.diesel · « **Reply #230 on:** December 16, 2015, 11:25:54 am »

^ Mmmm, not sure why code tags didn't preserve the formatting. I have all of the raw data but would like to ensure my scripting, and would really like to kill more of my 2.4Ghz home automation devices, some of which are with feet of the DMM.

dr.diesel · « **Reply #231 on:** December 16, 2015, 04:43:15 pm »

Dangit.

Just found line sync turned off, which it wasn't initially. Either my script testing killed it or the built in auto cal, so the above results are pretty much worthless.

Teaser set currently running at a shorter 5 min duration:

Code: [Select]

| Multimeter 		| Range		| NPLC	| 	Counts	|    Sq.sum	|  Average	|    SD		|    SD,uV	|
| k7510 		| .1 		| 10 	|	301	| -0.00024168	| -8.029273E-07	| 5.626927E-08	|   0.056	|
| k7510 		| .1 		| .1 	|	301	| -0.00025192	| -8.369386E-07	| 3.502203E-07	|   0.350	|
| k7510 		| .1 		| 1 	| 	301	| -0.00022957	| -7.627042E-07	| 6.791389E-08	|   0.068	|
| k7510 		| .1 		| .01 	|	300	| 0.00098528	| 3.284266E-06	 | 5.638453E-07	|   0.564	|

TiN · « **Reply #232 on:** December 16, 2015, 04:49:57 pm »

Now we talking.
Can you also drop CSV (you can find link in my article about noise testing) data so I can run median/sd to check zero calibration as well

Also it would be proper to have meter warm up at least overnight...

dr.diesel · « **Reply #233 on:** December 16, 2015, 04:52:57 pm »

I'm going to let this 5 min teaser set finish (which I'll post), then start the 30min per Range/NPLC run overnight again, then will upload the data to your site.

Has been powered on since I got it yesterday, so plenty warm by now!

dr.diesel · « **Reply #234 on:** December 16, 2015, 06:08:07 pm »

5min run:

Code: [Select]

| Multimeter 	| Range	| NPLC	|  Counts |    Sq.sum	|  Average	|    SD		|    SD,uV	|
| k7510 	| .1 	| .01 	|   300   | 0.00098528	| 3.284266E-06	| 5.638453E-07	|   0.564	|
| k7510 	| .1 	| .1 	|   301   | -0.00025192	| -8.369386E-07	| 3.502203E-07	|   0.350	|
| k7510 	| .1 	| 1 	|   301   | -0.00022957	| -7.627042E-07	| 6.791389E-08	|   0.068	|
| k7510 	| .1 	| 10 	|   301   | -0.00024168	| -8.029273E-07	| 5.626927E-08	|   0.056	|
| k7510 	| 1 	| .01 	|   301   | 0.00116343	| 3.865231E-06	| 1.363790E-06	|   1.364	|
| k7510 	| 1 	| .1 	|   301   | -0.00021776	| -7.234496E-07	| 5.733036E-07	|   0.573	|
| k7510 	| 1 	| 1 	|   301   | -0.00028301	| -9.402422E-07	| 1.024698E-07	|   0.102	|
| k7510 	| 1 	| 10 	|   301   | -0.00028457	| -9.454107E-07	| 8.580283E-08	|   0.086	|
| k7510 	| 10 	| .01 	|   301   | 0.00220821	| 7.336241E-06	| 1.301336E-05	|  13.013	|
| k7510 	| 10 	| .1 	|   301   | -0.00016204	| -5.383527E-07	| 2.435271E-06	|   2.435	|
| k7510 	| 10 	| 1 	|   301   | -0.00031698	| -1.053088E-06	| 7.528228E-07	|   0.753	|
| k7510 	| 10 	| 10 	|   301   | -0.00034107	| -1.133112E-06	| 5.040502E-07	|   0.504	|
| k7510 	| 100 	| .01 	|   301   | 0.16086693	| 5.344416E-04	| 7.598755E-04	| 759.875	|
| k7510 	| 100 	| .1 	|   301   | 0.09894463	| 3.287197E-04	| 1.358261E-04	| 135.826	|
| k7510 	| 100 	| 1 	|   301   | 0.10250599	| 3.405514E-04	| 3.750361E-05	|  37.504	|
| k7510 	| 100 	| 10 	|   301   | 0.10177610	| 3.381266E-04	| 2.422995E-05	|  24.230	|
| k7510 	| 1000 	| .01 	|   301   | 0.06277022	| 2.085389E-04	| 1.766823E-03	|1766.823	|
| k7510 	| 1000 	| .1 	|   301   | -0.04796741	| -1.593602E-04	| 2.381594E-04	| 238.159	|
| k7510 	| 1000 	| 10 	|   302   | -0.05574608	| -1.845897E-04	| 5.981283E-05	|  59.813	|
| k7510 	| 1000 	| 1 	|   301   | -0.05423848	| -1.801943E-04	| 7.543913E-05	|  75.439	|

3roomlab · « **Reply #235 on:** December 16, 2015, 06:57:58 pm »

hi dr.diesel. is there anyway to see/know p-p noise of the 7510?

dr.diesel · « **Reply #236 on:** December 16, 2015, 07:13:52 pm »

You mean what is listed on the spec sheet?

3roomlab · « **Reply #237 on:** December 16, 2015, 08:00:41 pm »

nope not the pdf
i meant the data you collected for 7510, which gave the STDEV. is there a way you can also output the peak-peak noise you have?

dr.diesel · « **Reply #238 on:** December 16, 2015, 08:13:01 pm »

Oh, gotcha. I can't do that and keep TiN's format, but I'll upload the raw data here if you'd like to import and play.

I'm gonna do the 30min runs overnight tonight, takes about 9 hours total, I'll then upload in the AM.

dr.diesel · « **Reply #239 on:** December 17, 2015, 11:19:44 am »

30 min run results, will upload to TiN's site shortly.

Warmed up a couple days, auto zero, no REL/math, 60Hz, 1sec polling, Keithley 4 term shorting plug.

Code: [Select]

| Multimeter 	| Range	| NPLC	|  Counts  |    Sq.sum   	|  Average	|    SD		|    SD,uV	|
| k7510 	| .1 	| .01 	|   1800   |   0.00649502	| 3.608347E-06	| 5.157321E-07	|   0.516	|
| k7510 	| .1 	| .1 	|   1801   |   -0.00123428	| -6.853294E-07	| 2.515242E-07	|   0.252	|
| k7510 	| .1 	| 1 	|   1801   |   -0.00110277	| -6.123113E-07	| 5.430182E-08	|   0.054	|
| k7510 	| .1 	| 10 	|   1801   |   -0.00110202	| -6.118930E-07	| 3.105326E-08	|   0.031	|
| k7510 	| 1 	| .01 	|   1801   |   0.00784606	| 4.356503E-06	| 1.394442E-06	|   1.394	|
| k7510 	| 1 	| .1 	|   1801   |   -0.00105845	| -5.877040E-07	| 3.416894E-07	|   0.342	|
| k7510 	| 1 	| 1 	|   1801   |   -0.00112253	| -6.232788E-07	| 8.595656E-08	|   0.086	|
| k7510 	| 1 	| 10 	|   1801   |   -0.00101890	| -5.657409E-07	| 7.448410E-08	|   0.074	|
| k7510 	| 10 	| .01 	|   1801   |   0.01771527	| 9.836353E-06	| 1.266939E-05	|  12.669	|
| k7510 	| 10 	| .1 	|   1801   |   -0.00054103	| -3.004068E-07	| 2.496712E-06	|   2.497	|
| k7510 	| 10 	| 1 	|   1801   |   -0.00143662	| -7.976804E-07	| 7.432814E-07	|   0.743	|
| k7510 	| 10 	| 10 	|   1801   |   -0.00134081	| -7.444814E-07	| 5.072808E-07	|   0.507	|
| k7510 	| 100 	| .01 	|   1801   |   0.89856873	| 4.989277E-04	| 7.725466E-04	| 772.547	|
| k7510 	| 100 	| .1 	|   1801   |   0.56770246	| 3.152151E-04	| 1.102306E-04	| 110.231	|
| k7510 	| 100 	| 1 	|   1801   |   0.59237881	| 3.289166E-04	| 3.473867E-05	|  34.739	|
| k7510 	| 100 	| 10 	|   1801   |   0.58874287	| 3.268978E-04	| 1.637639E-05	|  16.376	|
| k7510 	| 1000 	| .01 	|   1799   |   1.35719201	| 7.544147E-04	| 1.723659E-03	|1723.659	|
| k7510 	| 1000 	| .1 	|   1801   |   -0.31368228	| -1.741712E-04	| 2.448538E-04	| 244.854	|
| k7510 	| 1000 	| 1 	|   1801   |   -0.35386557	| -1.964828E-04	| 7.926890E-05	|  79.269	|
| k7510 	| 1000 	| 10 	|   1802   |   -0.34549073	| -1.917263E-04	| 5.191135E-05	|  51.911	|

TiN · « **Reply #240 on:** December 18, 2015, 04:27:39 am »

I'll process it and add pk-pk readings into script as well. I think it's not a bad idea to have more than one metrics.

rigrunner · « **Reply #241 on:** December 19, 2015, 11:15:35 pm »

Just uploaded a little over 30 minutes worth of hp3457a 3V range.

90 minutes warm up. copper short, NPLC=10, Autozero=on, MATH=off, NDIG=6, 10GOhm, 1 second per reading.

Running 300mV now. Will upload when done.

Edit: 300mV uploaded. Running 30mV now.

3roomlab · « **Reply #242 on:** December 19, 2015, 11:48:52 pm »

hi rigrunner, in my curiosity to see the p-p noise. i DL your csv file but this is what i saw.
i may have converted the file wrongly, but could you check?

rigrunner · « **Reply #243 on:** December 19, 2015, 11:54:01 pm »

Hi 3roomlab,

The data you have is correct. When loaded into librecalc i see it converted:

3roomlab · « **Reply #244 on:** December 20, 2015, 12:27:02 am »

ah i see. this is my log in xls for your sample view (should open as with no problem?), column 1 is time reference, column 2 is data in nV. but for TiN's web format i think 9 decimals of data /w scientific notation? i think you need to unlock the numbers after the 6th decimals

deadlylover · « **Reply #245 on:** December 20, 2015, 12:44:35 am »

Thanks for posting the 7510 results dr.diesel, very interesting stuff!

Its 1plc noise is extremely good, similar to 3458A spec for 10V, but it seems like going to 10plc doesn't give as significant a reduction in STDEV as some other meters. Perhaps it's one of their design compromises for very good 1plc noise performance?

3roomlab, I noted a few pk-pk readings when I was taking measurements from the R6581. Here they are for your amusement.

10V Range, pk-pk
1plc = 6uV
10plc = 1.2uV
100plc = 0.4uV

100mV Range, pk-pk
1plc = 1.18uV
100plc = 0.22uV

rigrunner · « **Reply #246 on:** December 20, 2015, 12:46:40 am »

Quote from: 3roomlab on December 20, 2015, 12:27:02 am

ah i see. this is my log in xls for your sample view (should open as with no problem?), column 1 is time reference, column 2 is data in nV. but for TiN's web format i think 9 decimals of data /w scientific notation? i think you need to unlock the numbers after the 6th decimals

There are no further decimals that i'm aware of?

I sent TiN a sample of the data before i started logging and he said that was ok.

I've just uploaded the log for 30mV range.

3roomlab · « **Reply #247 on:** December 20, 2015, 01:39:48 am »

@rigrunner, i think if you look at manatee mafia's 3458a log in xdev website, you see that the data is 1.2345-Exx. but your data is only zero and one.

@deadlylover ty for the data. reference to your log, it seems we might have similar "specs"

this pic summary of my test logs so far. with alot of help from kleinstein, it has been rather fun so far (hope it wasnt too boring for those who try to help me). i have just changed some caps in zero buffer, IPA washed and spewing new logs looking for improvement (again)

(crazy aim, pp noise below 0.5uV @ 1NPLC?)

** i editted the pic to include some of forumners data 263-270 (my K2015 logs are not to be taken seriously at this point, it is uncalibrated, and not a proven mod, it only proves the changes can provide lower noise)

rigrunner · « **Reply #248 on:** December 20, 2015, 02:15:20 am »

Quote from: 3roomlab on December 20, 2015, 01:39:48 am

@rigrunner, i think if you look at manatee mafia's 3458a log in xdev website, you see that the data is 1.2345-Exx. but your data is only zero and one.

The file contains exactly what was displaying on the 3457a LCD. i.e the only reason there is a 1 and all zeros is because only the single least significant digit was changing whilst logging.
The only other digit available it the 7 MATH HIRES and i do not see a way to get that whilst logging

3roomlab · « **Reply #249 on:** December 20, 2015, 02:26:32 am »

hmmm i see, try math then. result = reading x 1000. so you could at least log in millivolts 0.0001234? but im not familiar with hp, maybe somebody familiar can advice


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: DMM Noise comparison testing project (Read 217238 times)

Share me