Hi,

I'm at a point to purchase a new 8.5dig DVM. Choices are the (good old HP)
Agilent 3485A or the Fluke 8508A. The Fluke is almost 2x as expencive...

My feelings say that the 3485A is still the best one out there, is that
true?

Thanks,
Bram

For measuring low voltages (less than 100 mV, and even up to 1V), I found
the 34420A far superior than the 3458A. But it requires a yearly calibration
plan because the (really) long term drift tends to be higher than other DMM.

Does anybody tried the Keithley 2002 ?

-----Message d'origine-----
De : volt-nuts-bounces@febo.com [mailto:volt-nuts-bounces@febo.com] De la
part de The-al-bundy@home.nl
Envoyé : samedi 14 novembre 2009 16:04
À : volt-nuts@febo.com
Objet : [volt-nuts] Which meter?

Hi,

I'm at a point to purchase a new 8.5dig DVM. Choices are the (good old HP)
Agilent 3485A or the Fluke 8508A. The Fluke is almost 2x as expencive...

My feelings say that the 3485A is still the best one out there, is that
true?

Thanks,
Bram

volt-nuts mailing list -- volt-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.

Hi Bram,
If you want to make high-speed measurements, the Agilent 3458A is the clear
choice (because it can take up to 100,000 readings per second). The same
reasoning would apply if you need to use your DVM to make precision
digitization of waveforms. (For example the 3458A's triggered, digitized
measurements can be used to evaluate the linearity of a triangle waveform.)

For basic DC linearity (e.g., from full-scale to tenth scale or lower), the
3458A is as good as it gets, as far as I know.

Having said that, there might be a particular application where the Fluke
8508A would be your best fit? But I'm not sure about that. Do you have
specific measurement applications in mind?

I've personally always been extremely impressed with the 3458A performance,
so unless I had a specific measurement application (that I'd researched and
found that was addressed better by the 8508A), I'd buy the 3458A.

Best,
Greg

----- Original Message -----
From: "The-al-bundy@home.nl" the-al-bundy@home.nl
To: volt-nuts@febo.com
Sent: Saturday, November 14, 2009 8:03 AM
Subject: [volt-nuts] Which meter?

Hi,

I'm at a point to purchase a new 8.5dig DVM. Choices are the (good old HP)
Agilent 3485A or the Fluke 8508A. The Fluke is almost 2x as expencive...

My feelings say that the 3485A is still the best one out there, is that
true?

Thanks,
Bram

volt-nuts mailing list -- volt-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.

Hi Bram,

I agree that the 8508A and 3458A do seem like good choices. As for
connectivity, I guess you're stuck with GPIB with either of them.

I have a lot of experience with the 3458A, but none with the 8508A. So it
would be nice to hear from someone with 8508A experience for comparison.

Based on my personal experience with the 3458A (vs. 34401A), I'd say you
might typically see approximately these amounts of short-term "noise" (for
10Vdc range, 100 Power Line Cycles):

For 3458A: +/- 0.2uV variation from reading-to-reading
For 34401A: +/- 10uV variation from reading-to-reading

(At 10Vdc this would be +/- 0.02ppm "racking" for the 3458A vs. +/-1ppm
"racking" for the 34401A.)

I think you might find the 3458A's linearity to be typically < 0.06ppm on
its 10Vdc range. (Fluke's 720A Kelvin Varley Divider isn't linear enough to
verify this, so you'd need a JJ-Array to "prove" it.)

As for long-term DC stability of the 3458A: I'd say you might typically see
< 2ppm drift / 6 months. (I'm currently seeing < 2ppm drift / 12 months on a
couple old 3458As here. ...New 3458As would likely drift more than this
until they age for several years.)

For added confidence, you could use a trended Fluke 732B 10Vdc standard (or
a bank of them) to correct for 3458A long-term drift. (But IMO this would
require 732Bs with known history and as trended over several years minimum.)

Disclaimer: The above numbers are what I'm actually used to seeing here.
Published specs are of course greater than these numbers. So, for
traceability purposes you'd have to use published specs (k = 2) for your DVM
& Fluke 732B (unless you have several years historical data / trending).

Cheers,
Greg

----- Original Message -----
From: "The-al-bundy@home.nl" the-al-bundy@home.nl
To: "Discussion of precise voltage measurement" volt-nuts@febo.com
Sent: Saturday, November 14, 2009 1:56 PM
Subject: Re: [volt-nuts] Which meter?

Hi Greg,

Well I've made a design with the LTZ1000A and must test it's stability. The
best DVM we have is the 34401, but with the 6.5dig you don't get far with a
LTZ1000 as DUT. So we need something better. As far as I've seen the only
options are the 8508A and the 3458A.

The 8508A is relativly new (2002 I thought), so one would expect better
specifications than the 3485A and more options like other connectivity than
GPIB.
However the 3458A is already around there for >20years and it seems to be
one of the best DVM. The 100k sample rate is a feature which could be handy
in the future.

Money is not so much an issue, just we have to make a good investment for
the next 5+ years or so.
Of course we expect to use the DVM more often for ultra stabile
circuits/sources (in currents as wel as volts).

Thanks,
Bram

Bill - The Fluke 845A/B, 750A, 332 series and 335 series are obsolete.
Spec'd at only 10ppm accuracy, how would the 750A reference divider fit
Bram's project? (For comparison the 3458A's linearity is > 100x better than
the 750A.)

Also how would the 332 or 335 series be good enough sources for a LTZ1000A
UUT project? (For comparison the 332 & 335 short-term noise might be
typically 50x greater than the 3458A, according to my measurements.)

:) Greg

----- Original Message -----
From: "WB6BNQ" wb6bnq@cox.net
To: "Discussion of precise voltage measurement" volt-nuts@febo.com
Sent: Saturday, November 14, 2009 2:13 PM
Subject: Re: [volt-nuts] Which meter?

Bram,

You need to consider a different method !  You would be better off renting a
Fluke 732 voltage transfer standard with current calibration and a Fluke
845A/B
(desk version - not the rack mount version) null meter.  The desk version
has a
high isolation resistance then the rack mount unit.

The 845 series null meter has a DC output on the rear that is a full scale
representation of whatever scale you have selected on the front panel.  This
way
you can use just about any piece of crap meter with computer connections to
log
the results.

I am assuming you are setting the LTZ1000 to a standard 10 volts.  If not
you may
need to use a good divider; either a Fluke 750A or the Fluke 720A Kelvin
divider
if you have a weird voltage level below 10 volts.  If you have a voltage
that is
higher then you may need a good high quality voltage source like the Fluke
332 or
335 or one of the newer ones from Fluke.

Bill....WB6BNQ

WB6BNQ wrote:
...Once you have a referenced null condition it
is a matter of recording the variation from that point.
-----clip-----

Bill,
Bram stated that he wants performance better than the 34401A for his LTZ1000
UUT project and I agree with his premise. For starters I think he wants more
resolution? ...With that requirement, I'd think it prudent to have low
enough noise and linearity in order for the added resolution to be
meaningful. The linearity wouldn't matter if doing 1:1 voltage transfers,
but would matter if doing non-unity voltage ratio or absolute measurements.
For example, if the DVM is calibrated at 10Vdc, its accuracy at 7Vdc will,
in part, be dependent on its linearity. ...And in the case of the 3458A, its
accuracy on ranges other than its 10Vdc range will also depend on its
linearity (based on the way it's calibrated at 10Vdc and subsequently
auto-calibrates its other ranges as dependent on its linearity).
Greg