HP Part Numbers

[Radio Wrangler]

I was watching an episode yesterday of EEVBlog, where Dave was trying to figure out what a number of ICs were in his recently acquired bench DVM. The parts all had HP's proprietary part numbers marked on them.

The xxxx-xxxx format, usually called four-dash-four number was used for parts HP purchased from elsewhere where x is any digit

The xxxxx-xxxxx was used for custom made parts specific to one instrument, though they may later get used in other instruments. Usually the prefix five digits would be the instrument model number with leading zeroes padding it out. If so, the first digit after the dash sometimes gave a bit more information.
^ for PCB assemblies and 2 for unloaded PCBs. If so the following digits related to which PCB assembly in the instrument

EG 03725-60001 decodes as the loaded PCB called the A1 assembly in the 3725A (If I remember, it's the frequency counter board in the upper box of the Baseband Analyser) If later boards were suffuciently far redesigned to get a new part number rather than just a revision code letter, then the last 5 digits might be jumped by +100

4 letters-dash-5digits was a format used by HP's semiconductor arm for devices they sold outside. For maximum confusion, they also would have 4digit-dash-4digit numbers as stores numbers where they were also used insode HP.

EG 1SB1-0021 was a true RMS thermal converter used in the HP3336A Level Generator and several later instruments, though not sold outside that I know of.

EG HSMS-2800 was a very common 70v Schottky diode (the surface mount option in this case) It was sold outside. These part numbers went to Agilent then Avago when the semiconductor operation operation was spun-out and cast adrift on the stock market.

So that's the oddball stuff out of the way.

There is more detail in the 5-dash-5 digit parts. The last 5 digits were just issued chronologically assigned by corporate (Galactic HQ on Page Mill Road, Palo Alto)
The first 5 digits, the 'prefix' tell you what sort of part it is.

0698- resistors
0757- also resistors
0160- non polarised capacitors
0180- electrolytic capacitors
1850- germanium transistors
1853- silicon PNP
1854- sinicon NPN
1855- FETs - all types, J, MOS, either polarity etc.
1820- Digital ICs, but some early linear ones got included!
1826- Linear ICs
1858- transistor arrays
1901- general diodes small signal and power
1902- zener (and avalanche, if being pedantic) diodes
1906- diode groups like bridge rectifiers.
0121- variable capacitors of the mechanical kind
0122- varactor diodes

THis is incomplete and off the top of my head so there may be an odd error.

The HP stock numbers caused difficulties for parts supplier, so a number of ICs with HP stock numbers on the top (as per the purchasing contract) had their manufacturer's type number printed on the underside. Not all, just some, but if you get desperate, it's somewhere worth looking. I used to build dead-bug breadboards with ICs upside down with superglue on raw PCB material, so this was handy for me!

Also the HP stock numbers sometimes got shortened to fit on the smaller parts.

1854-0247 for example might have 4-247 printed on it. It's a 2N3866 with better quality control. When a division was closed down >30,000 were given away for free to the amateur radio world. They've never been seen since. They got spread around the globe and must have been grabbed by handfuls and salted away in junk box hoards.

Complements weren't coded together or anything fancy like that. The last 4 digits were simple chronological

1853-0036 (3-36 to its friends) was 2N3906
1854-0215 (4-215) was 2N3904

in the 4-4 parts numbers there are prefixes for nuts, screws, washers springs, even office and cleaning supplies, paint, abrasives, glues, hickory cocktail sticks! tools, power tools and solder. While doing some training, I got to set a new part number up as an exercise. I don't know if it ever got corporate approval. Can't remember the number but the description was "Washer, non-ferrous, mint-flavor" with the correct American (mis-)spelling of 'flavour'. I've been a friend of horses for a long time!

A bit of a brain dump, but I hope it helps people trying to fix some nice classic test gear.

David

[Andrewausfa]

There's a spreadsheet on one of the vintage electronic forums with a lot of HP part numbers and equivalents. Can't remember where though at this time.

Andrew

[Humber888]

There are a few cross references about and most have been mentioned before on this forum. I have three different ones saved. I attach one here, but the main one is 22MB in size and too big to post here. There is also another one, put together by WME_Bill, that I will leave to him to post if he wishes. The problem I find is that the particular device I need is never listed. Would it be worth making this thread into a sticky if we could bring together the various versions in one place?

[TonyDuell]

The thing to search for is 'Bench Briefs'. Some issues contained cross-reference lists of ICs and transistors. Note that they only listed parts used in istruments that were being supported at the time, meaning you want to get the older lists as well.

xxxx-xxxx part numbers can certainly be HP custom parts. HP custom ICs used in their calculators, for example, have numbers of that format. xxxxx-xxxxx parts tend to be subassemblies rather than components.

[Radio Wrangler]

Yes, the scheme got varied and evolved a bit over the years. I worked there through much of it. Still have my original HP45 calculator.

So, yes, there have been various exceptions, but the overall structure is as described above

David

[factory]

Some of the pdf lists are on the HP archive site here; http://hparchive.com/hp_parts

It includes the 1973/74 parts price list too.

David

[Terry_VK5TM]

Originally Posted by Radio Wrangler

It's a 2N3866 with better quality control. When a division was closed down >30,000 were given away for free to the amateur radio world. They've never been seen since.

David

The GQRP club has a stash of them and I have a few left after I gave some to a new ham.

[Radio Wrangler]

They got distributed via the people who came to the Rochdale doo one October. Thousands went to the US, Russia, Germany and so on. Glad some made it to VK. In view of the quantity, I'd expected Sprat to be full of things using them for the next year or so, but they never got mentioned.

They were all distributed pro gratis as Agilent had requested.

They'd have gone to landfill otherwise so any getting used would be a net win. They'd been a life-buy for the calibrator oscillator of RF power meters. The new products were going SMT and the whole product line was being moved out east, where they didn't want the burden of carrying historic stock.

They seem to have found a niche hereabouts in the amplifiers of active loop antennae.

David

[WME_bill]

HP part numbers.
Picking up the hint from Humber888, I attach my list of commercial equivalents. I collected this some years ago, mainly for transistors and diodes when I was working on HP oscilloscopes. It wants to be read with the later Bench Briefs lists.
It raises the problem of parts specially selected from the commercial versions. As with David Radio Wrangler's comment about 2N3866 seleted = 1854-0247. It seems better to use the nearest appropriate part and accept a working machine which does not quite meet the proper performance is always better than one which does not work at all.
I admire these perfectionists who want the exact equivalent 50 years afterwards.
wme_bill

[Radio Wrangler]

Plenty of the 'selected' ones were not necessarily for better performance parameters. They were sometimes for things like more extensive testing, or corrosion testing or statistical checking of life expectancy with accelerated tests at elevated temperature.

I'd have to go look at HP's specification control document for each part to see what was special. and I don't have access to that database any longer... it was all microfiche in 80-column card stuff, and I guess it's buried in landfill, somewhere.

David