jack wrote:
Thaler has the best voltage references -- there was a design in one of the
rags (Nuts n Volts or QEX) a few years ago in which 3 Thaler references
were placed in parallel for a super reference. Email me privately if you
would like a copy.

Also, get a copy of Linear Technology Application Note 86 "A Standards Lab
Grade 20 bit DAC with 0.1ppm/degC Drift" -- Jim Williams is the author --
and make sure to read the appendices. Williams also did an ApNote with a
comprehensive history of voltage measurement techniques, but I forget the
application number.



Thanks for the suggestion. I must admit to never having heard of of
'Thaler', but a quick search has thrown up some really interesting
stuff. No online ordering, but an 800 number to find the distributors.
I'll follow up on that.

I'll get the LT AppNote. Thanks for the info.

73 Alf NU8I

Thanks for the reference to the references, Jack. I emailed them, and
got back a nice reply that said, in summary:
-- Yes, they sell singles, no problem.
-- Order direct from them.
-- Current prices for four items I asked about are
** VRE100C $43.50 4 week lead time
** VRE102C $48.73 in stock
** VRE302AS/AD $11.04 in stock
** VRE302CS/CD $6.52 in stock

Those first ones are extremely low tempco parts (typ. 0.3ppm/C) with
very good initial accuracy (0.01% max error), +10V and +/-10V. The
second two are 2V parts with still very good tempco and good initial
accuracy. See their website for details.

Cheers,
Tom

Tom --

If you didn't see the subsequent post -- the VRE305 used by Tipton has a
Temperature Out pin -- 630mV with a 2.1 mV/C coefficient. This could be
used with a transistor array to maintain a constant temperature for the
device.

Jack

On Fri, 21 Apr 2006 00:40:19 -0400, "jack"
wrote:

Tom --

If you didn't see the subsequent post -- the VRE305 used by Tipton has a
Temperature Out pin -- 630mV with a 2.1 mV/C coefficient. This could be
used with a transistor array to maintain a constant temperature for the
device.

Jack


The LT1019A Linear Technology chip that I played with has the same pin
to give a changing output as the junction temperature changes. However
LT says that it is pretty much useless for actual temperature
relationships. I took measurements with that pin and also found it to
not correlate with temperature close enough to do anything with other
than give an indication of change. I believe it is buried to far into
the junction.

73
Gary K4FMX

Hi Alf,

Look at this item on ebay. This guy makes little reference boards for
a good price. I have had conversations with him and got some good info
on standards.

http://cgi.ebay.com/10V-L-STANDARD-V...QQcmdZViewItem

If the link doesn't work search ebay on "voltage reference".

I built a little standard using a different chip than he used. I used
an LT1019A and trimed it in.

I have a small spread sheet comparing his board with mine and the 731B
transfer standard if you would like me to send it to you.

He had sent me one of his boards to play with. I found that they move
quite a bit with temperature. Temperature is the biggest problem in
maintaining a stable reference with those chips. I have a fluke 731B
transfer standard to reference to, using an hp 3456 voltmeter to
measure things with.

The board he has is not a bad deal as he sends it out calibrated. If
you can duplicate his temperature you will do pretty good but probably
not quite as good as he advertises.

A good pre regulated supply is important to feed the reference chip
too.

73
Gary K4FMX


On Fri, 14 Apr 2006 19:47:03 -0700, Alfred Green wrote:

I have recently aquired a nice 5.1/2 digit DVM, and would like to assess
the absolute accuracy. Many years ago I had a 'Standard Cell', but that
got lost in one of my several relocations.
What is a good alternative these days? I would think that there is a
semiconductor device that has a known threshold. I am familiar with the
Wheatstone bridge method to measure voltage without affecting the reference.
Once I have a good DC reference, getting AC & RMS calibration is fairly
easy.
Having a GPS locked time and frequency reference, I am used to being at
the 1e-10 accuracy level for those, but if I can get to 1e-5 for analog
voltage I would be very happy.
Any suggestions gratefully received.

73 Alf NU8I
Scottsdale AZ DM43an
160m - 10 Gigs

One good thing about the Thaler references is that they provide a
Temperature Output -- so if you are using a Fluke, Keithley, HP-Agilent DVM
with GPIB, HP-IL or RSwhatever you can compensate for temperature
differentials.

Of course, you can also use the temperature output to drive a differential
amplifier and "cook" the unit to a constant temperature much as we used to
with crystal oscillators (boy, am I dating myself now

"Gary Schafer" wrote in message
...
Hi Alf,

Look at this item on ebay. This guy makes little reference boards for
a good price. I have had conversations with him and got some good info
on standards.


http://cgi.ebay.com/10V-L-STANDARD-V...QQcmdZViewItem

If the link doesn't work search ebay on "voltage reference".

I built a little standard using a different chip than he used. I used
an LT1019A and trimed it in.

I have a small spread sheet comparing his board with mine and the 731B
transfer standard if you would like me to send it to you.

He had sent me one of his boards to play with. I found that they move
quite a bit with temperature. Temperature is the biggest problem in
maintaining a stable reference with those chips. I have a fluke 731B
transfer standard to reference to, using an hp 3456 voltmeter to
measure things with.

The board he has is not a bad deal as he sends it out calibrated. If
you can duplicate his temperature you will do pretty good but probably
not quite as good as he advertises.

A good pre regulated supply is important to feed the reference chip
too.

73
Gary K4FMX


On Fri, 14 Apr 2006 19:47:03 -0700, Alfred Green wrote:

I have recently aquired a nice 5.1/2 digit DVM, and would like to assess
the absolute accuracy. Many years ago I had a 'Standard Cell', but that
got lost in one of my several relocations.
What is a good alternative these days? I would think that there is a
semiconductor device that has a known threshold. I am familiar with the
Wheatstone bridge method to measure voltage without affecting the
reference.
Once I have a good DC reference, getting AC & RMS calibration is fairly
easy.
Having a GPS locked time and frequency reference, I am used to being at
the 1e-10 accuracy level for those, but if I can get to 1e-5 for analog
voltage I would be very happy.
Any suggestions gratefully received.

73 Alf NU8I
Scottsdale AZ DM43an
160m - 10 Gigs