[Phys-L] Re: Digital Multi Meters (DMM)

["Folkerts, Timothy J" <FolkertsT@BARTONCCC.EDU>]

You might add a theoretical estimate of the expected variations.

The standard deviation of a uniform distribution from a to b is sigma=
 =3D (b-a)/12^0.5, or 0.58 in your case.  The standard error is SE =
=3D sigma/n^0.5.  With 2000 samples/sec for 20 sec, this works out to=
 0.58/200 =3D 0.0029.  This is the expected standard deviation for th=
e means you calculated.  Your means show a st dev of 0.0047, which is=
 noticably bigger than the results of the simulation, but in the same=
 ballpark.  The difference is probably not statistically significant.=
 =20

Tim F


-----Original Message-----
=46rom: Forum for Physics Educators on behalf of Ludwik Kowalski
Sent: Tue 9/20/2005 8:11 PM
To: PHYS-L@LISTS.NAU.EDU
Subject: Re: Digital Multi Meters (DMM)
=20
Inspired by two private massages, and by what was posted yesterday by
JohnD (see below), I summarized my own thinking at

Http://blake.montclair.edu/~kowalskil/cf/258sampling.html

1) Is the Monte Carlo approach convincing? All comments and suggestio=
ns
will be appreciated. Thanks in advance.

2) A week from now I will be working in a private research lab trying
to replicate a French excess heat experiment (see unit #252). Two oth=
er
teams (one involving two students), will try to replicate the same
experiment. Our instrument will sample volts and amps at the rate of =
2
kHz; the results will be sent to a computer file.

Ludwik Kowalski
Let the perfect not be the enemy of the good.


On Sep 19, 2005, at 3:04 PM, John Denker wrote:

> ludwik kowalski wrote:
> > Digital Multi Meters (DMM) sample waveforms and display updated
> > avarages.
> >
> > 1) Does anybody know what the sampling frequencies are in cheap DM=
Ms?
> Long ago Rat Shack sold one that used a dual-slope integrator with
> a one-second cycle time.  Ewwwww.
>
> > 2) What are these frequencies in more expensive DMM's?
>
> I have a handheld Fluke (10+ years old) that is good to 100kHz,
> true RMS.
>
> I have (relatively) cheap National Instruments boards lying around =
the
> lab that are good to 20MHz.
>
> > 4) Any potential traps in using DMMs?
>
> More than I know how to count.
>
> Perhaps the biggest one is Dragon Breath:  The instrument injects w=
eird
> noise out of its input into the device under test.
>    There are many types of Dragon Breath.  One kind arises if the
>    nominally-floating (i.e. gauge invariant) probes aren't quite,
>    i.e. there is some nontrivial impedance (usually capacitive) to
>    the rack and/or the "ground" of the power line.
>
> Then there's aliasing at high frequencies.
>
> Then there's eyeball-aliasing, if you only look at it intermittentl=
y,
> as opposed to logging *all* of the data.
>
> Sensitivity.
>
> Linearity.
>
> Multiplicative calibration errors.
>
> Additive DC offsets.
>
> Non-infinite input impedance.
>
> Thermocouple effects, if the thing you're measuring isn't at room
> temperature.
>
> etc. etc. etc.
_______________________________________________
Phys-L mailing list
Phys-L@electron.physics.buffalo.edu
https://www.physics.buffalo.edu/mailman/listinfo/phys-l