[volt-nuts] What made a HP3458A so expensive

Bill Gold wpgold3637 at att.net
Sun Jan 19 16:27:20 EST 2014


    I would guess that a local JJ would have to be compared with a
"standard" JJ.  Since NIST says "Ours is correct and yours could be a little
more or less than ours, you need to know the difference".  But this all get
into a really "gray" area in the 0.01 ppm region.  So what are you going to
do?  Move your JJ to NIST?  Not practical!  My suggestion about a JJ in a
9.5 digit was just to get a more stable voltage reference.  Of course at the
present time given the fact that the JJ has to be cooled to below 4.2 K
makes this idea more than a little impractical.  Of course you could setup
to reference the voltmeter to a stand alone JJ in your workshop.  See this
website for a turnkey unit you can get.  www.supracon.com    They have two
models, one which uses liquid He and the other has a Cryocooler.  I am sure
that you can talk the little lady into this as a birthday present?
    I am not sure what Fluke does to be "traceable" to NIST but I would
imagine that Fluke regularly sends carefully characterized 734A units to
NIST in a round robin so that they can accomplish this.  There is an
application note somewhere that talks about this.
    Same with the Hz, the ohm and so on, you have to somehow reference it to
NIST, to claim you are "traceable to NIST".


----- Original Message ----- 
From: "Joe Hobart" <nova at npgcable.com>
To: "Tom Miller" <tmiller11147 at verizon.net>; "Discussion of precise voltage
measurement" <volt-nuts at febo.com>
Sent: Saturday, January 18, 2014 4:44 PM
Subject: Re: [volt-nuts] What made a HP3458A so expensive


Would a Josephson Junction standard need to be calibrated?

Adjusted and maybe compared, yes, but you should not need to calibrate a
standard?  Years ago we had HP Cesium Frequency Standards at work.  There
primary standards and good to 4E-12 with no additional calibration.


On 1/18/2014 3:37 PM, Tom Miller wrote:
> Can you imagine what it would cost to get that calibrated?
> Tom
> ----- Original Message ----- From: "Bill Gold" <wpgold3637 at att.net>
> To: "Discussion of precise voltage measurement" <volt-nuts at febo.com>
> Sent: Saturday, January 18, 2014 5:17 PM
> Subject: Re: [volt-nuts] What made a HP3458A so expensive
> In fact if you look at a 1982 HP catalog you will see that the 3456A was
> selling for around $3,700 and given inflation the 3458A is still a pretty
> good bargain when it was introduced in 1989.  In 1989 the 3456A was
> for $4,600 while the 3458A was selling for $5,900.  I wonder if there will
> ever be a "3459A" 9.5 digit meter?  With a super miniture Josehpson
> for a reference?
> Bill
> ----- Original Message ----- From: "Tom Knox" <actast at hotmail.com>
> To: "Discussion of precise voltage measurement" <volt-nuts at febo.com>
> Sent: Saturday, January 18, 2014 1:29 PM
> Subject: Re: [volt-nuts] What made a HP3458A so expensive
> Actually in spite of it's high price I feel the 3458A represents a
> When it was introduced in 1989 it was $5900. A price it held for decades.
> Although designed in
> then 80's it's accuracy is still unsurpassed.  The 8846A is much less
> expensive because it has only about a quarter the parts. With few that
> are hand selected and/or aged. And all though both meter have well
> executed designs the result is the 3458A has about a magnitude greater
> accuracy. And anyone in Metrology can attest to the fact that the cost
> of accuracy is exponential. But the thought I wanted to contribute to the
> dialog is the real value of the 3458A is the body
> of knowledge built around the thousands of 358A's some running
> continuously for nearly three decades. The characteristics of the 3458A
> are perhaps the best documented of any electronic instrument ever made.
> That is priceless. When comparing that body of knowledge to individual
> units I have found every 3458A is a little different and seem to each
> have their own personality. In a side note, recently I owned what appeared
> to be the original 3458A. I cannot remember the serial number, but it had
> the numbered stickers identifying the boards that can be seen in the
> original service manual. And it was still  working flawlessly.
> Thomas Knox
>> Date: Sat, 18 Jan 2014 17:41:26 +0100
>> From: frank.stellmach at freenet.de
>> To: volt-nuts at febo.com
>> Subject: [volt-nuts] What made a HP3458A so expensive
>> Well,
>> the 3458A now is in production for about 25 years , and only around 50k
>> units have been sold.
>> (That's an estimation by Joe Gellers series number collection)
>> As the 3458A is a niche product with homeopathic series volume, the
>> development, verification/validation, special reliability engineering,
>> selection and burn-in measures, QA costs had to be paid per unit in the
>> beginning.
>> To my opinion, especially the 3458A was designed mainly for military
>> requirements (Tamb 55°C). The military was willing to pay a premium (HP
>> = High Price) at that time. So HP was able to realize that price.
>> All that development budget is long paid, and after end of the Cold War,
>> the military does not order so many devices anymore, what caused the
>> problems of the T&M business (finally => Keysight, urgh!).
>> But as that market is tight, competitors are few, so the price is not
>> going down, instead it's increasing over the years, from $5900 to around
>> $8500.
>> It would be interesting to calcualte the BOM of the 3458A.
>> As they use many custom specific components, it should be relatively
>> expensive.
>> There is no parameter in the HP3458A specifications, that the device has
>> to be powered constantly to meet the specs.
>> Those very high 8ppm/yr. drift might apply only during continuous
>> operation.
>> During power down, the LTZ should not drift at all, as the ageing
>> mechanism is driven by temperature.
>> But there might be (there are indeed) considerable hysteresis effects.
>> I have set the temperature of my HP3458A to ~ 60°C, I shut it down after
>> usage, and the periodical comparison to 3 other references shows a drift
>> of less than 1ppm/year.
>> Frank
volt-nuts mailing list -- volt-nuts at febo.com
To unsubscribe, go to
and follow the instructions there.

More information about the volt-nuts mailing list