*** SOLD*** HP3456a 6 digit DMM

[saturation]

Yes, what I meant was its had its share of life's experience  who knows what it went through, always good to know its pedigree to 'guestimate' its remaining life.

http://www.vias.org/transistor_basics/transistor_basics_07_05_01.html

Thanks for the tip about the ROMs, I'll keep a look out.

Given its a c1982 design, how much longer it will continue to function is unknown.

The actual life is always unknown until in dies . There's no such thing as a programmed max. life span. It's designed so a certain percentage will still be working after the designed life span. These target percentage and life span back then by HP for a top-of-the-line piece of equipment are probably much higher than for current consumer equipment from any consumer equipment manufacturer (which ironically includes HP). The probability of failure might raise slightly after this time, but many electronics components are likely to live on for a very long time if they don't die as an infant. They have been burned in for almost thirty years (if it was actually produced in 1982) . Look at the Mars Rovers, designed to be likely to work for at least 90 days (with some safety margins I'm sure, since RMA is even more expensive than on earth), actually survived for years. It didn't blew up on day 91 as you might expect from the design lifespan. Plenty of equipment built in the sixties that still works (probably not with the original tubes, but these were designed to be replaced periodically). It depends more on the storage and amount of use it saw than on actual age. Electrolytics can be a problem, but then again, some electrolytics from 50+ years old that still work, and plenty five years old electrolytics that are already dead.

There's one known issue with the 3456A, and that is that the ROM's are known to fail. The contents and instructions for replacement are available online, so not a big deal. Not a lot of reports of 3456A's (or other instruments from similar vintage) dieing, though. This instrument was used for a long time in a lot of places, especially since the 3457A didn't offer much advantages (lousy display, but faster, higher resolution via GPIB only, stability slightly worse, more features like current measurement).

[tecman]

Nice to see that I was somehow a bit responsible for some good discussion.

I actually have, and have had, quite a few pieces of 20 year old HP and TEK equipment.  I currently have an 8457a DMM and quite a bit of TEK TM500 stuff.  I have made some repairs, but in general the reliability is far greater than a lot of modern equipment that I have at work.  I frequently end up bringing one of my instruments to work as it does a better job than some of the equipment we have to use.

In general, the most frequent failure I have seen in the equipment is tantalum caps.  Never an electrolytic and rarely a semi.  My most used scope is still a TEK SC504, 80 MHz analog.  It has suffered 2 dead tantalum caps (PS bypass) in the years I have had it.  Likewise my TEK AFG5101 has also suffered 2 tantalum cap fatalities, but otherwise no issues at all.  I also have a 30+ year old TEK 547.  Needs an overall cal every year or so, but it keeps going and going.  It is more a show piece for me, but still very usable (for those unfamiliar, a mostly tube based with some semis, 5" 50 MHz with plug-in vertical modules.  Weighs about 80 lbs).

As for ROM failures, that seems rare.  I assume that they use EEPROMS, which do have some finite limits on retention life.  As a safeguard, one could unplug and copy them and keep a HEX file or a fresh set of PROMS.

Nice to have lab grade equipment on a beer budget.  While I could afford newer stuff, why spend the money.

Paul

[saturation]

Hello tecman,

Again my best thanks for the DVM.  

These pearls of experience are invaluable and thanks for sharing.  My experience with quality is similar with biological and medical equipment, somehow the 1970-1990s produced a range of electronic devices that were indestructible and retired because they were  slow and obsolete rather than ceasing to function.  

Your experience with tantalum cap failures is interesting, since its considered one of the more reliable types today.  However, 40 years ago as it was still new, it was reported to suffer failures akin to electrolytics because of corrosives its in construction.  I guess if we have devices still existing with such caps, we'd expect to see them fail.

http://powerelectronics.com/passive_components_packaging_interconnects/wet-tantalum-capacitors-dscc-93026-0101/

Were those older equipment of yours bought new or bought used?

Nice to see that I was somehow a bit responsible for some good discussion.

I actually have, and have had, quite a few pieces of 20 year old HP and TEK equipment.  I currently have an 8457a DMM and quite a bit of TEK TM500 stuff.  I have made some repairs, but in general the reliability is far greater than a lot of modern equipment that I have at work.  I frequently end up bringing one of my instruments to work as it does a better job than some of the equipment we have to use.

In general, the most frequent failure I have seen in the equipment is tantalum caps.  Never an electrolytic and rarely a semi.  My most used scope is still a TEK SC504, 80 MHz analog.  It has suffered 2 dead tantalum caps (PS bypass) in the years I have had it.  Likewise my TEK AFG5101 has also suffered 2 tantalum cap fatalities, but otherwise no issues at all.  I also have a 30+ year old TEK 547.  Needs an overall cal every year or so, but it keeps going and going.  It is more a show piece for me, but still very usable (for those unfamiliar, a mostly tube based with some semis, 5" 50 MHz with plug-in vertical modules.  Weighs about 80 lbs).

As for ROM failures, that seems rare.  I assume that they use EEPROMS, which do have some finite limits on retention life.  As a safeguard, one could unplug and copy them and keep a HEX file or a fresh set of PROMS.

Nice to have lab grade equipment on a beer budget.  While I could afford newer stuff, why spend the money.

Paul

[alm]

The ROM's aren't EEPROM's, I think they are Mostek mask ROM's. They are known to die after twenty years or so. The contents are available, and there are instructions about alternative EPROM that work (rom_info_rev.txt is the most recent), so replacing them is not a big deal (you just need an EPROM programmer or someone else to program them). I think I even found some usenet posts from Winfield Hill (of Art of Electronics fame) about them .

I'm not claiming that transistors and IC's don't have failure modes, but the question is if these are relevant on this timescale. These semiconductors usually don't handle high current and have a huge feature size. Most failures (fortunately) tend to be in simple things like power supplies and connectors, like the already mentioned tantalums. There are some semiconductors with (relatively) high failure rates like the mentioned ROM's and some Tektronix/Maxim IC's in their 2400 series of analog scopes, probably because there was something wrong with that semiconductor process. But then again, people rarely post about their equipment working fine, so you always get a distorted picture (popular equipment seems to break often).

Tantalums from that timespan seem to fail often, I believe those cheap tants were originally intended for consumer audio. It also doesn't help that some companies like Tektronix liked to run things close to the specs, like using 6V tants on a 5V rail.

Modern equipment doesn't always have a great long-term reliability either, think about the crappy electrolytic capacitors leaking after a few years, or the Nvidia bumpgate issue, or tin whiskers due to lead-free processes. So I wouldn't bet on modern day equipment outliving the old stuff (but I wouldn't bet against it either), think of the old stuff as having had a really long burn-in . At least the older stuff has schematics available and less custom parts.

[saturation]

Thanks Alm.  I have those discussion links, but I thank you for the real links you provide.  For reference to all.  It also has the brief discussion with Hill on his 3456A back in 2005, and Stephan who not only takes the lead, but compiles most of the info.

http://www.electronicspoint.com/hp-3456a-multimeter-roms-gone-bad-t27876.html

The ROM's aren't EEPROM's, I think they are Mostek mask ROM's. They are known to die after twenty years or so. The contents are available, and there are instructions about alternative EPROM that work (rom_info_rev.txt is the most recent), so replacing them is not a big deal (you just need an EPROM programmer or someone else to program them). I think I even found some usenet posts from Winfield Hill (of Art of Electronics fame) about them .

I'm not claiming that transistors and IC's don't have failure modes, but the question is if these are relevant on this timescale. These semiconductors usually don't handle high current and have a huge feature size. Most failures (fortunately) tend to be in simple things like power supplies and connectors, like the already mentioned tantalums. There are some semiconductors with (relatively) high failure rates like the mentioned ROM's and some Tektronix/Maxim IC's in their 2400 series of analog scopes, probably because there was something wrong with that semiconductor process. But then again, people rarely post about their equipment working fine, so you always get a distorted picture (popular equipment seems to break often).

Tantalums from that timespan seem to fail often, I believe those cheap tants were originally intended for consumer audio. It also doesn't help that some companies like Tektronix liked to run things close to the specs, like using 6V tants on a 5V rail.

Modern equipment doesn't always have a great long-term reliability either, think about the crappy electrolytic capacitors leaking after a few years, or the Nvidia bumpgate issue, or tin whiskers due to lead-free processes. So I wouldn't bet on modern day equipment outliving the old stuff (but I wouldn't bet against it either), think of the old stuff as having had a really long burn-in . At least the older stuff has schematics available and less custom parts.