Agilent 34461A Blog #2 - Histograms

This blog entry describes my experience with the histogram function on the Agilent 34461A 6 1/2 digit multimeter.

Note:  the symbol table function in the e14 blog editor is not working for me tonight, so I've compromised by using the word "ohm", etc where a symbol would normally have been inserted.

My exploration of the histogram function begins with a plan to build up a histogram of single measurements on batches of same value through-hole carbon composition  resistors all from the same production batch.  All are 1/2 watt 5% tolerance.  The table below lists the resistors I have available for this experiment and the calculated range of expected values based on the marked 5 % tolerance.  Below the table are photgraphs illustrating what the resistors look like.

My hypothesis is that I will get a histogram of values that fall within 5% tolerance of the marked value.  Becuase I don't know much about how carbon composition resistors are made I'm at a loss to predict what the distribution will look like; will it be a normal distribution, will it be clustered tightly around one value, or will it be random within the +/- 5% tolernace range?

The Agilent was set up with a 10 PLC aperture (default) on 2 wire Ohm measurement mode.  I used the mini-grabber accessory clips provided in the standard meter package and attached to the resistors as shown in the photograph below.

The data and screen captures shown below were uploaded to my laptop using Agilent's web server interface and Java control application.

So, let's see what the Agilent 34461A has to say about the 27 k Ohm resistors.

I'd like to mention that the mini-grabbers worked well - most of the time.  There were four occassions out of 35 measuremtns where at least one mini-grabber popped off the resistor due, I think, to too much spring tension.

The data set and a screen capture of the histogram are shown below.

Statistics

  Maximum      +2.73474963E+04    

  Average      +2.70284436E+04  

  Minimum      +2.68477032E+04  

  Max-Min      +4.99793179E+02  

  Std Dev      +1.21381349E+02  

  Samples      +3.50000000E+01  

Readings in Memory        

     1     +2.73474963E+004   OHM

     2     +2.69631261E+004   OHM

     3     +2.68739164E+004   OHM

     4     +2.71572212E+004   OHM

     5     +2.72346633E+004   OHM

     6     +2.71610892E+004   OHM

     7     +2.69721184E+004   OHM

     8     +2.69428587E+004   OHM

     9     +2.70358656E+004   OHM

    10     +2.71179774E+004   OHM

    11     +2.71977400E+004   OHM

    12     +2.71104134E+004   OHM

    13     +2.69775011E+004   OHM

    14     +2.68982629E+004   OHM

    15     +2.69891330E+004   OHM

    16     +2.71619889E+004   OHM

    17     +2.70255919E+004   OHM

    18     +2.69424486E+004   OHM

    19     +2.69425727E+004   OHM

    20     +2.69995757E+004   OHM

    21     +2.70591338E+004   OHM

    22     +2.69333675E+004   OHM

    23     +2.68949665E+004   OHM

    24     +2.69386663E+004   OHM

    25     +2.70410473E+004   OHM

    26     +2.68530799E+004   OHM

    27     +2.71647968E+004   OHM

    28     +2.68477032E+004   OHM

    29     +2.68888014E+004   OHM

    30     +2.71103027E+004   OHM

    31     +2.68996053E+004   OHM

    32     +2.71378899E+004   OHM

    33     +2.70726875E+004   OHM

    34     +2.69801325E+004   OHM

    35     +2.71217843E+004   OHM

My hypothesis seems to be holding up, at least on this first batch of resistors.  The tolerance is in fact tighter than +/- 5 %, more like +1.23% and -0.56%.  The distribution on the histogram only has 35 samples, but I'm imagining a normal-ish looking distribution around 27 k and another a little above 27.1 k.  Can't really tell with so few samples, but the distribution does not look random to me.

For the next batch (27 ohm) I'm going to try controlling the meter remotely using the Java app.

Okay, that didn't work as expected.  Everytime I issue a single measurement command, the meter erases the previous reading and displays the new reading on the histogram.  I'm going to switch back to local control.

Here is the data upload and histogram screen shot from the tests on the thiryfive 39 Ohm resistors:

Statistics

  Maximum      +3.94152503E+01   

  Average      +3.90134987E+01   

  Minimum      +3.86178934E+01   

  Max-Min      +7.97356924E-01   

  Std Dev      +1.95523792E-01   

  Samples      +3.50000000E+01   

Readings in Memory          

     1     +3.90358902E+001   OHM

     2     +3.88118580E+001   OHM

     3     +3.90688944E+001   OHM

     4     +3.90347095E+001   OHM

     5     +3.89345431E+001   OHM

     6     +3.86178934E+001   OHM

     7     +3.89141395E+001   OHM

     8     +3.89478755E+001   OHM

     9     +3.87854830E+001   OHM

    10     +3.89800036E+001   OHM

    11     +3.89409588E+001   OHM

    12     +3.88595993E+001   OHM

    13     +3.89785270E+001   OHM

    14     +3.89142445E+001   OHM

    15     +3.88326800E+001   OHM

    16     +3.87932210E+001   OHM

    17     +3.89549198E+001   OHM

    18     +3.91400952E+001   OHM

    19     +3.89353286E+001   OHM

    20     +3.93903102E+001   OHM

    21     +3.94152503E+001   OHM

    22     +3.88626692E+001   OHM

    23     +3.91126198E+001   OHM

    24     +3.92140419E+001   OHM

    25     +3.93256019E+001   OHM

    26     +3.88831385E+001   OHM

    27     +3.92743014E+001   OHM

    28     +3.89708495E+001   OHM

    29     +3.89907057E+001   OHM

    30     +3.92882155E+001   OHM

    31     +3.92911029E+001   OHM

    32     +3.87861133E+001   OHM

    33     +3.91051478E+001   OHM

    34     +3.88135788E+001   OHM

    35     +3.92679423E+001   OHM

Distribution looks like it might be heading toward normal if another dozen or samples were added.  Tolerance on the batch is +1.06%, -0.98%.

On to the last batch of 82 k Ohm resistors.

Here is the uploaded sample data and a histogram screen capture for the 82 k set:

Statistics

  Maximum      +8.54020421E+04   

  Average      +8.27459595E+04   

  Minimum      +8.16834894E+04   

  Max-Min      +3.71855267E+03   

  Std Dev      +8.35697976E+02

  Samples      +2.90000000E+01   

Readings in Memory         

     1     +8.32284702E+004   OHM

     2     +8.17466775E+004   OHM

     3     +8.54020421E+004   OHM

     4     +8.25844720E+004   OHM

     5     +8.22425058E+004   OHM

     6     +8.32842572E+004   OHM

     7     +8.28901213E+004   OHM

     8     +8.31141028E+004   OHM

     9     +8.24669051E+004   OHM

    10     +8.28375632E+004   OHM

    11     +8.24135370E+004   OHM

    12     +8.27156239E+004   OHM

    13     +8.16834894E+004   OHM

    14     +8.17992854E+004   OHM

    15     +8.29342385E+004   OHM

    16     +8.52172421E+004   OHM

    17     +8.25209963E+004   OHM

    18     +8.22971059E+004   OHM

    19     +8.21039877E+004   OHM

    20     +8.24272434E+004   OHM

    21     +8.21239864E+004   OHM

    22     +8.22688290E+004   OHM

    23     +8.25217364E+004   OHM

    24     +8.27924554E+004   OHM

    25     +8.24654742E+004   OHM

    26     +8.30222581E+004   OHM

    27     +8.26209760E+004   OHM

    28     +8.24595790E+004   OHM

    29     +8.34476653E+004   OHM

Tolerance on this batch is +4.15%, -0.39%.  Histogram looks like it might be normal around 82 k.

For a comparison, I tested the 39 Ohm batch on a Tektronix DMM4050.  The Agilent mini-grabbers fit perfectly over the Tektronix probes.  Set up on the Tektronix was more involved than on the Agilent, but once set up, testing progressed identically to the Agilent.  I haven't figured out the ethernet interface to the DMM4050 , so I took a picture.

Results compare fairly well with the Agilent.  SD on the Tektronix is 0.19233 and 0.1955 on the Agilent.  Average value on the Tektronix shows 38.4887 Ohms and 39.0135 on the Agilent.  The differences can be explained by calibration tolerances on the two instruments and differences in contact resisitance between the two test runs.

I think I will give the thumbs up to the Agilent on the histogram feature.  Except for the oddity with the Java app resetiting the histogram with every remote single measurement activation, it is easier to set up than the Tektronix and I prefer the colour graphic display.

Mark

November 24, 2013 update - Single Resistor Histogram

Following up on requests from DAB and Michael Kellett here are the results of taking 50 single shot measurements on a single 39 Ω resistor.  The mini-grabbers were attached then left alone as I pushed the "Single" button 50 times, about once per second.  Not bad, the range of values is only 15.999 mΩ.

Statistics
  Maximum         +3.93168063E+01       
  Average         +3.93121523E+01       
  Minimum         +3.93008066E+01       
  Max-Min         +1.59997727E-02       
  Std Dev         +4.87241628E-03       
  Samples         +5.00000000E+01       
Readings in Memory          
     1        +3.93008066E+001      OHM
     2        +3.93022761E+001      OHM
     3        +3.93025170E+001      OHM
     4        +3.93033838E+001      OHM
     5        +3.93039061E+001      OHM
     6        +3.93049939E+001      OHM
     7        +3.93050193E+001      OHM
     8        +3.93059102E+001      OHM
     9        +3.93058261E+001      OHM
    10        +3.93063711E+001      OHM
    11        +3.93069525E+001      OHM
    12        +3.93075867E+001      OHM
    13        +3.93079976E+001      OHM
    14        +3.93091710E+001      OHM
    15        +3.93095270E+001      OHM
    16        +3.93100188E+001      OHM
    17        +3.93105183E+001      OHM
    18        +3.93110841E+001      OHM
    19        +3.93115193E+001      OHM
    20        +3.93125160E+001      OHM
    21        +3.93129599E+001      OHM
    22        +3.93134301E+001      OHM
    23        +3.93139920E+001      OHM
    24        +3.93141193E+001      OHM
    25        +3.93145563E+001      OHM
    26        +3.93150615E+001      OHM
    27        +3.93152292E+001      OHM
    28        +3.93153802E+001      OHM
    29        +3.93157713E+001      OHM
    30        +3.93158292E+001      OHM
    31        +3.93158171E+001      OHM
    32        +3.93160875E+001      OHM
    33        +3.93160494E+001      OHM
    34        +3.93159974E+001      OHM
    35        +3.93163855E+001      OHM
    36        +3.93165920E+001      OHM
    37        +3.93168063E+001      OHM
    38        +3.93165653E+001      OHM
    39        +3.93162551E+001      OHM
    40        +3.93166738E+001      OHM
    41        +3.93163524E+001      OHM
    42        +3.93163196E+001      OHM
    43        +3.93160285E+001      OHM
    44        +3.93162206E+001      OHM
    45        +3.93161061E+001      OHM
    46        +3.93159135E+001      OHM
    47        +3.93157415E+001      OHM
    48        +3.93156908E+001      OHM
    49        +3.93158962E+001      OHM
    50        +3.93158861E+001      OHM

One last test.  Here I remove and attach the mini-grabbers before taking a single shot measurement.

I won't be doing that again; nearly went cross-eyed trying to see the mini-grabbers and the resistor lead.  Like threading a needle.

Here are the results:

Variation is 43 mΩ.  A little higher than repeated static measurements, which is to be expected.

Statistics
  Maximum         +3.93001830E+01       
  Average         +3.92697205E+01       
  Minimum         +3.92571718E+01       
  Max-Min         +4.30111369E-02       
  Std Dev         +9.65337414E-03       
  Samples         +5.00000000E+01       
Readings in Memory          
     1        +3.93001830E+001      OHM
     2        +3.92927014E+001      OHM
     3        +3.92938794E+001      OHM
     4        +3.92780041E+001      OHM
     5        +3.92720320E+001      OHM
     6        +3.92824438E+001      OHM
     7        +3.92862144E+001      OHM
     8        +3.92713947E+001      OHM
     9        +3.92659655E+001      OHM
    10        +3.92638516E+001      OHM
    11        +3.92655769E+001      OHM
    12        +3.92670959E+001      OHM
    13        +3.92626747E+001      OHM
    14        +3.92613892E+001      OHM
    15        +3.92836176E+001      OHM
    16        +3.92744823E+001      OHM
    17        +3.92645096E+001      OHM
    18        +3.92653555E+001      OHM
    19        +3.92627554E+001      OHM
    20        +3.92696779E+001      OHM
    21        +3.92602449E+001      OHM
    22        +3.92646397E+001      OHM
    23        +3.92689527E+001      OHM
    24        +3.92633604E+001      OHM
    25        +3.92615403E+001      OHM
    26        +3.92584853E+001      OHM
    27        +3.92571718E+001      OHM
    28        +3.92591628E+001      OHM
    29        +3.92619207E+001      OHM
    30        +3.92623446E+001      OHM
    31        +3.92621642E+001      OHM
    32        +3.92604018E+001      OHM
    33        +3.92754237E+001      OHM
    34        +3.92707788E+001      OHM
    35        +3.92894296E+001      OHM
    36        +3.92694495E+001      OHM
    37        +3.92773542E+001      OHM
    38        +3.92741571E+001      OHM
    39        +3.92710827E+001      OHM
    40        +3.92685833E+001      OHM
    41        +3.92682162E+001      OHM
    42        +3.92696821E+001      OHM
    43        +3.92666290E+001      OHM
    44        +3.92668182E+001      OHM
    45        +3.92680937E+001      OHM
    46        +3.92663928E+001      OHM
    47        +3.92667144E+001      OHM
    48        +3.92656447E+001      OHM
    49        +3.92638388E+001      OHM
    50        +3.92635403E+001      OHM

Kind regards and thanks for reading.