Its no Fluke...

[saturation]

Thanks alm for that historical perspective. Each paragraph in reply to bolded items.

Was doing some research after reading your post.  I see the 8508A is a metrological DMM, a truly super meter, with amazing specs, and a price to match, ~ $10,000 used  and sadly the 'net says its discontinued.  Amazing, it was built back in 2002.  By comparison, a new 3458a as a basic DMM is $8000, new, $9800 with the high ppm stability board.  Going rate at eBay is ~50% off.



I can see why.  I'm not sure what the 8508a ran for when new, but it seems for calibration use, there's nothing like an 8508a, its accuracy is even mentioned in confidence intervals, which is really the most precise way to mention accuracy, and its AC accuracy is peerless.

In the HP 345x pedigree, there is a leap in accuracy between 3455 and 3456, not from 6 to 7 despite the higher scale reading, and again a big leap from the 7 to 8.  As a simple reference, the 10V scale [ 30V for the 3457a] at 90 days all converted to ppm, including the LSB:

345x ppm

5   51 6   25 7   54 8   4.15

Yes, I guess because of the current range so the 6 is not really a 'complete' DMM?  But the amp accuracy on the 7 is low, a good handheld is better.  Its LSB count error is absurdly high.

Yes, the 3457a go for more, given the added capabilities, I presume beyond accuracy.  I'm comparing the final purchase price, whether it be buy it now or bid, whichever is lower from eBay.  Also comparing unit equitably whether its for parts, working but not calibrated, and calibrated. 

 

The Fluke HP 3458A had better specs than the HP one (though I believe the difference was just one resistor). The Fluke 8508A is currently the most accurate general purpose lab multimeter, as far as I know.

It's kind of funny. For 'normal' folks like us, the HP 3458A is like the holy grail of lab multimeters. I was reading the PMEL (army/air force/navy calibration lab) forum, and they complained that the 3458A was useless for anything but DC, and preferred the 8508A.

The HP 3458A is fairly old (1989 or so), from memory, I think the 3457A came out in 1986/1987, and the 3456A in '81, not sure about the 3455A (late seventies?). I would have expected a 3459A somewhere in the past 20+ years.

Does the 3457A really fetch $200 more than the 3456A in similar condition? They both have some pros and cons, no clear cut winner as far as I'm concerned. I had a choice between a 3456A (tested, good history) and 3457A (lightly tested, gave calibration error about current range) for $50 extra. I chose the 3456A because of the better display and better stability, and because the current range on the 3457A was probably bum anyway. So in my opinion, the prices should be pretty close. The 3456A was more expensive than the 3457A when both were current products, and the 3455A was even more expensive.

[saturation]

At least for the Fluke 80s series, you can buy replacement LCD at eBay for about $10-20.  Depending on the type of model, some other LCDs too can be easily replaced.

Thanks for the various confirmations of Fluke's good name in DMM's
I used Flukes at work for many years  until the LCD displays went dark or leaky

Wanted one for myself but a new one financially out of reach and ePays prices were astronomical for 2nd hand ones.  Eventually got an ex Forces TEK DMM916
Checked it against my in CAL Agilent 34401A  and it agreed right across all ranges.
Ime lucky to own reasonable DC/AC Cal sources (Fluke and eSi) including an eSi Dekatran DK72A
and the 916 sits smack in the centre of its published specs on all ranges!
Not bad for a 2nd hand DMM.  Tek must have learned well from Fluke (and its blue rubber holster is
easier to keep clean than Fluke yellow!)
This DMM  is also not well known either. If you cant afford a Fluke its a good alternative

John

[alm]

Not that it matters, but the revised specs for the 3456A (from the 1987 catalog) are actually better than that, 15ppm of value + 2ppm of range. I wonder why the offset specs of the 3457A are so bad for the 30V range, since you need that for 10V standards (don't think anyone uses saturated Weston cells these days). It doesn't have an 'infinite' (10Gohm) input impedance at 30V either.

Prices from the 1987 catalog are:
- 3455A $5600
- 3456A $4100
- 3457A $2800

Not sure if the 8508A is discontinued, Fluke still lists it on its website. But I do admit that I haven't bought any for a while, so I haven't actually checked .

At least for the Fluke 80s series, you can buy replacement LCD at eBay for about $10-20.  Depending on the type of model, some other LCDs too can be easily replaced.

There were also lots of these problem with the older 8000 series, not that bad if you consider the age. Our good friend from Greece has some experience with this.

Tek must have learned well from Fluke (and its blue rubber holster is
easier to keep clean than Fluke yellow!)
This DMM  is also not well known either. If you cant afford a Fluke its a good alternative

Not sure if they learned from Fluke, Fluke actually bought their DMM line (after a lawsuit about patents or similar looks or some crap like that) and used them as basis for their high-end (18x, now succeeded by 28x) series, so you might say they learned from Tek .

[saturation]

Thanks alm, great stuff. That's  ~ 10 ppm reduction.  The data I quote is from the Agilent supplied user and service manual which spans 1980-2.  Missed it, even if I do have a catalog entry from 1992, with a price near $6k.

http://home.g-net.net/%7Ejmarchetti/Photos/HP3456A/HP-Catalog-1992_3456A.pdf

How would you know the unit you own is subject to the revised specs?

The service manual has several revisions, the 2015Axxxxx series went through a lot of changes, including ROM related issues, and the 2201Axxxxx series is golden, for those interested those serial numbers mean 1980 15th week, and 1982 1st week respectively.

There are 2 zener reference boards for either line, one with the LM399H and another with a zener part I cannot find anywhere.  I presume the LM399H variants are the more accurate models but I have to open the DVM to find out!

Yes, the 3457 is awful from a precision point of view.  IMHO its not in league with what the 6 and 8 represent, metrological quality.  

Yes, the Fluke site describes the 8508a but does not mention its discontinued, but normally all order-able products will have a link to the buy column of the Fluke store.  Tequipment lists it as discontinued and no vendor I scanned has a new one for sale.  Maybe its made to order now?  Not really looking to buy one at this price, as it way above my league, but if one showed up on eBay reasonably priced, I'd certain give it a detailed look, particularly the service manual to see what I can do to keep it healthy.  

Which BTW, is another down point for the 7 and 8, no schematics or detailed how-to as the 6, making it very limited in terms of user serviceability.

@ hpxref, also Tek belongs to Denaher, the parent of Fluke, so now they are 'one' in the same.

Not that it matters, but the revised specs for the 3456A (from the 1987 catalog) are actually better than that, 15ppm of value + 2ppm of range. I wonder why the offset specs of the 3457A are so bad for the 30V range, since you need that for 10V standards (don't think anyone uses saturated Weston cells these days). It doesn't have an 'infinite' (10Gohm) input impedance at 30V either.

Prices from the 1987 catalog are:
- 3455A $5600
- 3456A $4100
- 3457A $2800

Not sure if the 8508A is discontinued, Fluke still lists it on its website. But I do admit that I haven't bought any for a while, so I haven't actually checked .

At least for the Fluke 80s series, you can buy replacement LCD at eBay for about $10-20.  Depending on the type of model, some other LCDs too can be easily replaced.

There were also lots of these problem with the older 8000 series, not that bad if you consider the age. Our good friend from Greece has some experience with this.

Tek must have learned well from Fluke (and its blue rubber holster is
easier to keep clean than Fluke yellow!)
This DMM  is also not well known either. If you cant afford a Fluke its a good alternative

Not sure if they learned from Fluke, Fluke actually bought their DMM line (after a lawsuit about patents or similar looks or some crap like that) and used them as basis for their high-end (18x, now succeeded by 28x) series, so you might say they learned from Tek .

[alm]

Missed it, even if I do have a catalog entry from 1992, with a price near $6k.

The updated catalog is available either from the Agilent website (HP archive?) or the HP Museum. Any catalog after 1987 that still includes the 3456A should be fine.

How would you know the unit you own is subject to the revised specs?

The updated specs were the result from new long-term stability data from the voltage reference, gathered for the 3457A that used the same reference, not from hardware revisions. This is stated in the February, 1986 edition of the HP Journal (can be downloaded from those same sites, I posted a link some time in the past, not sure if that was a response to you?). Not sure if this is valid for both the A24 and A25 reference boards. The later A25 board used the LM299H, and I think this change was made well before 1986 (my 1982 copy of the service manual lists it). I can't find it in the backdating section, so that suggest that there might be some overlap. My guess is that it applies to the A25 reference board. The 3457A service manual lists the reference assembly as 03456-66525, which is the 3456A A25 reference assembly, so I think it's a safe guess that that's the reference they were talking about in the HPJ article. Not sure about the A24 reference, can't find anything about that zener. Agilent's find-a-part claims that it was replaced by the 11177-69501 (3455A reference assembly), which was replaced by the 03455-66511, which suggests it's also used in the 3455A. I can't find a schematic or parts list for that in the 3455A service manual to check.

Yes, the 3457 is awful from a precision point of view.  IMHO its not in league with what the 6 and 8 represent, metrological quality. 

My guess that they were aiming for the system DMM market (automated test setups), not really the metrology market. You'd have to closely read the marketing material to be sure. It had features like high speed and optional scanner cards, so it might be more like a 34401A (which isn't metrology class either). The only other meter in that class (apart from the 345xA) at the time was the 3478A, which was pretty slow and had lousy accuracy.

Yes, the Fluke site describes the 8508a but does not mention its discontinued, but normally all order-able products will have a link to the buy column of the Fluke store.  Tequipment lists it as discontinued and no vendor I scanned has a new one for sale.  Maybe its made to order now?  Not really looking to buy one at this price, as it way above my league, but if one showed up on eBay reasonably priced, I'd certain give it a detailed look, particularly the service manual to see what I can do to keep it healthy. 

You may be right, but it's also possible that they just don't keep them in stock, and only sell directly.

Which BTW, is another down point for the 7 and 8, no schematics or detailed how-to as the 6, making it very limited in terms of user serviceability.

The 03457-90011 service manual from the Agilent website appears to include full schematics. The 03457-90012 copy I've got appears to only include sections I-V. There's a CLIP (component level information package) available for the 3458A from Agilent for $$$ ($150 or so?), which includes full schematics. There's a copy of this floating around on the net, but without Agilent's approval (they'd rather sell you a copy), as opposed to the manuals from obsolete instruments, which can be legally distributed.

[saturation]

Thanks alm, that's very helpful and vital information!  I'll be scanning the HP journal site, I just downloaded the 3457 service manual but didn't go through all the pages.

The volt-nuts forum has referenced several Chinese language forums that are very revealing.  I've attached some of the most intriguing ones here.

Here's a comparison of most top metrological quality DMM in production, its just one user's blog, but he has them 'all':

http://translate.google.com/translate?u=http%3A%2F%2Fwww.eefocus.com%2Farticle%2F09-04%2F70848s.html&sl=zh-CN&tl=en&hl=&ie=UTF-8


Also, in comparing the high end HP and Fluke, this author implies, even Fluke prefers the HP.  Here he labels and identifies key devices in what he claims is a Fluke official photo of their metcal lab [from a brochure?]:

http://translate.googleusercontent.com/translate_c?hl=en&sl=zh-CN&tl=en&u=http://www.hellocq.net/forum/showpost.php%3Fp%3D1892062%26postcount%3D284&rurl=translate.google.com&usg=ALkJrhgWNBGZplO4K1Lz4c1xlE0sg7REZQ



Probably the most demonstrative photo, this just how enthusiastic they are about electronics, the first I've seen, making creative photos from recycled LTZ1000:

http://translate.googleusercontent.com/translate_c?hl=en&sl=zh-CN&tl=en&u=http://www.hellocq.net/forum/showpost.php%3Fp%3D2087552%26postcount%3D509&rurl=translate.google.com&usg=ALkJrhjmRRwg8F5L8Li7Pdp103YOYIZmHw

[saturation]

Just reviewed the full service manual for the HP 3457a, yes, schematics in detail are available.  I've retracted my comments.

[alm]

Nice overview. I was aware of the Keithley (though I've heard bad comments about the stability of either the 2001 or 2002), Datron (Fluke's 8508A is basically a Datron meter), Solartron/Schlumberger (I've seen them on Ebay, the non-standard cable always turned me off) and Prema (construction seems cheaper and less professional than the others, though this doesn't influence the stability and accuracy, they often had tons of features like math and graphs, sometimes even with a built-in PC with hard drive), but not of Advantest. It appears that the 8508A is still the most accurate DMM. It's possible that Fluke prefers the 3458A (a Fluke HP 3458A?), it's only DC measurement as far as I can see. The PMEL technicians mainly disliked it for its AC performance, though they preferred the 8505A for any work. It's also possible that they didn't get around to replacing the 3458A, there's no date on the pictures.

I found the comment about the 8508A lacking frequency and period measurement strange, seems to me that if you want to measure frequency in a metrology lab, you'd use some high-end counter with precision reference, not a DMM that probably does a mediocre job.

Datron used a few parallel precision zeners with temperature compensation, at least in the older 1080/1081. I found it odd that they didn't used a heated zener (LMx99, LTZ1000) like most of the other manufacturers, though they appeared to get good stability and tempco with this. I wonder if the same technique is used in the later 1281/Fluke 8505A.