EEVblog

[@movax20h]

This is component of HP 3350 series, which is a lab automation system. HP itself also was a big player in analytical chemistry, especially chromatography and mass spectroscopy. But they were selling instrumentation, data logging, control, printing, storage, crts, etc, entire system as modular system, for other companies and to laboratories to build custom scientific equipment, both in chemistry, medical diagnostic, but probably also in other serious fields (physics, astronomy, etc). Here some pictures : pubs.acs.org/doi/pdf/10.1021/ac60330a715 , including this exact A/D converter model. pubs.acs.org/doi/pdf/10.1021/ac60332a722 for more details.

[@bobweiss8682]

This was a small part of a Gas Chromatograph setup. Used for digitizing the output from the detector.

[@Daveyk021]

That RED "gunk" is high voltage varnish. We used to call it "glip" to lock in the pot adjustment so that it would not vibrate to a different value and you could tell if someone adjusted it.

[@Chriva]

Looks like glass wool. I'd not scratch my balls after handling that if I were you :D

[@RuneTheFirst]

Capacitors: "Don't be vague, ask for Sprague!" (Their old slogan.)

[@volvo09]

Someone should revive the company! I miss seeing sprague caps.

[@RuneTheFirst]

volvo09 They sort of have 2 owners now - Vishay for tantalums and Cornell-Dubilier now makes the "Orange Drops". Vishay uses the Sprague name now and then but it is obviously not the same. I miss them too. It was nice to have a dependable domestic brand. Guess we have to rely on Japanese caps now (Nichicon, Matsushita, etc.) I've had too many Chinese caps fail (new ones).

[@spwicks1980]

Ray Rooney Wish they'd make their Sprague Atom electrolytics. Damn they were good for filter caps in valves amps.... they are highly sort after these days.

[@JGnLAU8OAWF6]

Do a review of that screwdriver please.

[@EEVblog]

Next mailbag.

[@DrakkarCalethiel]

AvE style? :D

[@moshegalimidi2302]

bet dave dosnt have a bosch nanoblade

[@lroy730]

Yeah that's bitchin !

[@vovan8547]

Доброго времени, а подскажи пожалуйста что это за аппарат.

[@dvlachy]

Dave, I've seen one of these at a Garrett manufacturing plant in UK. It was used to sample analog output from a speed sensor of a turbocharger under test. I believe there was kind of primitive digital closed loop system with adjustable PID controller. The purpose was perhaps to control the speed under various engine loads and rpm. There was more HPs in the rack, but I remember this one only due to the fact it was 40years old A/D.

[@jagardina]

Before the reveal I would guess some kind of precision data collection device. HP did a lot of automated test equipment stuff. We used a lot back at Bell for digital transmission testing. This seems to predate GPIB/HPIB.

[@cuteswan]

I immediately thought the gunk on the pots was just fingernail polish: My father and his colleagues used to put a dab of red fingernail polish on the edges of trimmer pots - just like that - after calibrating instruments. Supposedly it kept them from drifting due to vibrations, especially in heavy industrial settings. The blobs held well enough for however many months, then broke away easily when it was time to re-adjust. I don't know if the vibration issue was a real thing, though maybe it was also a bit of a deterrent for customers to futz around with the settings too. Thanks again for another fun & lively bit of education.

[@PedroDaGr8]

This device was part of a GC and/or LC Lab Automation System from HP. It would normally be used in a much more harsh environment than most electronics labs. Vibration from things like pumps, rough handling, thermal changes, etc. might certainly be an issue and a valid worry. The nail polish would make sense there, as would the conformal coating on the important and sensitive part of the board.

[@EEVblog]

Loctite (and others) make various grade sealers for this type of application

[@johnfrancisdoe1563]

EEVblog But generally, any red seal material symbolize that the authorized calibration is intact. Easily faked, but good enough in a non-adversarial environment.

[@tomgeorge3726]

You have to use "Coles" brand "Starlet" nail polish, "Cherry" colour...

[@cambridgemart2075]

The stuff we used to use was called Glyptol if I remember correctly; it was designed to stop inadvertent adjustment and to indicate if someone had tampered with the adjustments.

[@marklange8894]

Hi Dave, From what I have been able to find those 1852A’s were part of an HP 3352B Laboratory Data System, used widely in gas and liquid chromatography equipment. It had an HP “computer” if you could really call it that, with a whopping 24K of memory, and output to a plotter for accurate plots of chromatograph peaks. Used in everything from Medical research, drug testing, forensics, to oil and gas industry, and chemical manufacturing. High precision for the day, and HP apparently didn’t want anyone working on defective equipment (other than HP themselves) because you are correct about absolutely no manuals or schematics “in the wild”. That was pretty much standard practice for HP in the day. I hope this satisfies a bit of your curiosity. P.S. I actually saw one of those in the early 1980’s in a Mud Logging Trailer at a 22,000 Feet deep gas well drilling site in South Texas. It was connected to a gas chromatograph monitoring samples of gas coming out of the drilling mud that was pumped through the drill string then back up the bore hole into the mud pit. It was running in parallel to a Honeywell analog chart recorder, I guess they didn’t want to take a chance of the “new fangled” digital machine missing something.

[@JimGriffOne]

17:13 - I believe the "Start/Stop" button controls external equipment. It acts as a simple trigger to start other equipment connected to it.

[@borkowsm]

I think that HP mount this crystal upside down for AU market 😀

[@JimGriffOne]

Apparently it could be used as part of a gas chromatography setup - using the HP 5890 SERIES II Gas Chromatograph. There is mention of a cable in the document I found: Part number: 05890-60790 | Edge Connector: HP 18652A A/D Converter, 1V input

[@JimGriffOne]

Search this document and there's a single reference of "18652A". www.agilent.com/cs/library/support/documents/a15282.pdf

[@JimGriffOne]

There are four references to "18652A" in this document: www.aimanalytical.com/Manuals/7673B/HP%207673B%20-%20S.pdf

[@mrmobodies4879]

I found that document as well and there seems to be a clue in the filename: HP 1090 LC as an Auto Sampler by Agilent. It looks like that unit is part of a set of systems.

[@CentiZen]

this document has a wiring diagram that shows how to connect a 5890A to a 18652A using another piece of equipment in between www.willstein.com/Resources/HP%205890%20Cable%20Configurations.pdf

[@antigen4]

yeah i guess data acquisition would make sense

[@RuneTheFirst]

The Navy still had some of this stuff kicking around at their engineering station test facility in the early 90s. Long gone now.

[@sarreqteryx]

as mentioned already, it was used with a couple models of gas chromatographs, to provide a digital signal to computers. it was also used in industrial process control units, for much the same. I haven't been able to find much more than that. apparently, it was also co-produced by Agilent.

[@laptop006]

That old case is a giveaway if you have some of their other rackmount kit from that era.

[@KeanM]

Yep, the paint and case style was typical of 1970's HP

[@leisergeist]

More than just rack mount stuff afaik

[@BenHelweg]

And the typeset

[@Goonygoon84]

Heat resistant... insulation... from the 70's.. in the US.. That ranks up there on things I don't needlessly play with.

[@davidf2281]

Yeah. That woollen stuff terrifies me. I'm sure just a very short exposure would be okay but jeez, I wouldn't touch it.

[@leisergeist]

It's the Best-os!

[@kyaputenharokku9574]

That is absolutely not asbestos.

[@davidf2281]

Kyaputen Harokku How do you know? Not trolling, just interested.

[@JGnLAU8OAWF6]

Even if it is it's not THAT bad.

[@redtails]

our old GC setup had this ADC I believe.. the bit-depth wasn't so important, but the dynamic range was

[@DreitTheDarkDragon]

IDEA: What if input and output connectors are interconnected so you could simply daisy-chain your devices? Maybe real (digital) output is on that edge connector and pulse on "output" connector just tells all other devices to start or stop their operation? It could be connected via transformer to be insulated (I heard it was common before optocouplers were a thing), so it might explain negative pulse. Try to check that edge connector with oscilloscope, I'm pretty sure one of pins might be clock signal and is triggered automatically by timer you set inside.

[@unmanaged]

www.aimanalytical.com/Manuals/7673B/HP%207673B%20-%20S.pdf

[@DreitTheDarkDragon]

Pages 107 and 108 makes me feel that I might be right. *checks manual for my mid 70s multimeter* Maybe one of pins (pin 4?) is trigger to start single measurement? And then samples per second "switch" only changes accuracy (and length) of measurement.

[@5Komma5]

Cool thanks for setting up the referral links.

[@SeanBZA]

Were common as part of a HP ATE suite, where it was the ADC for some of the channels, switched by a relay board that used mercury relays to select the input.

[@StreuB1]

Great video Dave, thank you!

[@uzaiyaro]

Like that screwdriver. Highly recommend the ES120/121. Might even send you one. They’re a fantastic piece of kit. Same guys that make the TS100/80.

[@genixia]

Given the age of that gear, I wonder if Bob Widlar had any hand in those National Semiconductor Op Amps. I note that the fixed resistors are all 5 band 1% variants. Note just how bad electrolytic capacitors are: +/- 20% when new. More like +10/-30% as aged. The metal film cap has to be the sampling cap. I'm not seeing any DACs or R2R networks, so we can rule out any ADC type that needs a DAC. The thermal insulation ensures that both power rails are affected equally by ptc/ntc component effects, thus preventing the derived reference voltage from drifting. The signal path will have been given the same design consideration.

[@NICK-uy3nl]

A huge discrete A/D converter, before the advent of dedicated A/D ICs...it is amazing to see that entire thing on a single 10mmx10mm IC today, consuming 1/1000 of the power, with 1000 times higher sampling rate.

[@GeorgeGeorge-xj2bc]

Maybe something like ultra low noise-small signal,wide band and constant temperature amplifier full of custom components.

[@a1guitarmaker]

Man, I had HP as the company. It has the look. Thought it might be an audio product with digital processing. Wondered whether it was higher frequency stuff... Good game!

[@renanpils]

Hi Dave! I have seen something like this before. I recognized a part of the circuit made by HP in an old analog-digital conversion board I found on junk. The board I found seemed to be a prototype the electrical transmission company. No datasheet or manual was found with it. I didn't see anything strange as the woolen stuff you found though.

[@SkyCharger001]

The thickness of the can made me think of the can that houses an old-school FDR's tape. Which brings me to a possible reason you couldn't find it as an HP device. A few years ago I heard about people having trouble finding the purpose of a rusted through component they found at a beach as the design and stuff for the component was with the client Boeing instead of the (still identifiable) original manufacturer. Something similar could have happened here: HP build the device, but the rights are held by the client.

[@xl0xl0xl0]

Hey, Dave. Did you know that you could use fine grit sandpaper on the screen to get rid of the glare?

[@UberAlphaSirus]

polishing compound would probably give a better result

[@stofbag]

Yeah, the card-edge connector next to the insulated section is identical to the one on the old HP "integrator" graph plotter for the HP GC's. it's contemporary with the HP5710 gas chromatograph which I believe is the first with the form-factor agilent uses today

[@FaustFotographyLLC]

ONLY thing I found on this online is there is a ton for sale and on auctions. Nothing about it.

[@Jedda73]

That screwdriver caught my attention for a second till I saw the $450 Aus price tag! That's more than my Panasonic cordless drill cost with two batteries!

[@EEVblog]

Yup. in my mailbag upcoming video I mention the Panasonic.

[@ThinkingBetter]

Wonder if you could find a first generation audio ADC/master recorder from the beginning of digital music?

[@victorvarsanyi6702]

Did a little research and all i could find was that an HP 1050 Series Variable Wavelength Detector had an output for that specific ADC.

[@therm0tt0]

That Twin Peaks reference was fantaaaaastic.

[@revivedash2679]

Just a question I was wanting to ask. Why is it then whenever you wire up a breadboard to power you put the positive in the blue rail and the negative in the red rail? On other breadboards I’ve seen and have it’s the other way around.

[@Knight8365]

Really enjoyed this "Guess what it is" type of video. I originally thought it was a form of data capture, then switched to frequency reference. Close but no cigar :D

[@DrewskisBrews]

Don't be vague, ask for Sprague

[@TheDefpom]

@ 11:30 I'm pretty sure they are just trimmers.

[@Peter_S_]

That was another classic WTF facepalm event. Those are 5% trimmers. They even say "CIRCUITRIM" on top. Dave talks out his ass a fair bit.

[@pa4tim]

They are, I have seen them often in measurement gear in non-precision circuits.

[@mrmobodies4879]

The filename on the PDF that Jim Griffiths suggested points to an autosampler Hewlett-Packard - HP 1090 LC

[@NivagSwerdna]

OK so I guessed wrong initially. It's an ADC from a Gas Chromatagraph where the output is typically between 0 and 10mV. I'm surprised there isn't more analog goodness.

[@NivagSwerdna]

Ah... the card edge connector does go into the box... that's where the goodness for the front-end is.

[@TheHermitHacker]

Mid 70's... The smell is bang on. Could you send me this so I can smell what the 70s we're like? Thank you.

[@00Skyfox]

That card edge connector looks like the same connector as the Commodore C2N Datasette uses on all old Commodore computers.

[@dhpbear2]

10:55 - Those packages were common in The States back in the late 60's

[@sandeepkerala1982]

This is the ADC used in old HP(now agilent) gas chromatograph. This is a ADC for calculating the peak area(flame ionization detector femto to pico ampere current converted to plus or minus 1 volt using a transimpedance amplifier and the same will be converted to digtal using this ADC box. Sampling can be done up to 32 hertz(more number sampling more narrow peaks can be detected). We inject a standard, calculate are under the peak then inject sample and compare the area under the peak to the standard and will get the quantity. The peaks are eluted as an order and its called retention time. With respect to the retention tie , qualification can be don for different materials. Now many manufactures have gas chromatograph including us :). Data apex clarity is an excellent product now a days...

[@tomasxfranco]

Are the input leads shorted by an aluminum plate or does it just look like it from that angle?

[@zaprodk]

The Kemet caps in the shielded box are Wet-slug tantalum.

[@mrmobodies4879]

It is an interface for a Hewlett Packard 1090 Series II UV-Visible HPLC System. It is listed as a part here: se-source.com/hp-1090/

[@mikesradiorepair]

The US military used them. NSN (national stock number) was 6625-01-100-2596. It was part of a digital acquisition system made up of: Components includes, HP2113B processor with 32K memory, HP7906A disc drive with interface and controller, HP2645 crt with dual tape cartridges, HP966B thermal line printer, hp 18651A digital transmission loop controller, HP18652A analog module, HP29402B cabinet, 56 in 60 hz 120/208 V, HP18661D laboratory automation system software

[@PlasmaHH]

You find that thing mentioend in various HP/agilent gas chromatograph systems. There are alternative parts mentioned, so it must be some defacto standard for these things how to communicate. All I could find was that it seems to work on 1V signals. I also found it mentioned in a variable wavelength detector. Even a 2000s electrochemical detector mentions a connection to that thing. My guess is that its a pretty high performance part to increase the performance of an existing system.

[@afp.connex9289]

The New axial Cockroach egg capacitor from digilock at 4:10

[@wb5mct]

Good eye! Damn bugs are learning how to disguise their egg pods! LOL

[@amiralozse1781]

I guessed right! Its an ADC! Have to admit I'm no real electronics guy, just fascinated. ... the 'RDGS/SEC' selector was my final clue

[@oldblokeh]

It’s a 18652A a/d for the HP 335x series of lab data systems. I was an HP systems engineer for these systems for several years. They were manufactured in Avondale, Pennsylvania. A sister module was the 18653B event controller module (ECM) which has 16 bits of binary input and 7 triac outputs, mostly used to control an autosampler/autoinjector. These modules were daisy chained in a loop with a single shieded twisted pair. You could have up to 15 modules in a loop. The loop controller, 18651A, sat in an HP 1000 minicomputer. This started with the early 2100 series core based machines up to the later 21MX models. With chromatography, it’s area you are interested in, so this was an integrating a/d. The sampling rate was not high, but it didn’t need to be. You could have up to 100 meters of cable between modules as far as I can remember. With the 3357 system supporting 4 loops, you could have up to 60 channels of chromatographic data acquisition, integration and reporting going on simultaneously. There was a limit on the total sampling rate per loop. This module was also supported on the later ‘A-series’ systems, although it was superseded by an updated module. These modules were extremely reliable. I don’t recall ever having one fail at any of my customers.

[@RandyLott]

Dave's been hitting the weighted screwdriver at the gym. Just look how fast he's going at 3:50!

[@redsquirrelftw]

I figured it was either some kind of high precision clock, or a turbo encabulator.

[@electronic7979]

Useful video 👍

[@Psychlist1972]

When you showed the woolen stuff, I was imagining a bunch of exploded caps.

[@MikeB_UK]

Readings/sec seemed to originally be at 2 at 5:20 in video before you moved it down to 4. In case you care... :-)

[@JW-ce4lx]

It is actually an A to D interface controller

[@PyroRob69]

I seem to remember seeing something like this attached to the gas chromatograph when I was in college. It was a lot of years back.

[@cubeistgames7985]

The outside was obviously HP, about the same era as my HP 2114B in my collection.

[@diatomsaus]

What's that Wiha screwdriver you're using?

[@rich1051414]

In the US, we call that 'insulation' or 'fiberglass insulation'. As a kid, we would grab handfuls of it(with gloves on), rip it into pieces and drop it down peoples shirts as 'itching powder'. Wash your hands after handing that stuff xD

[@uwezimmermann5427]

4:51 "do we have a date code yet" - they obviously missed to camouflage the TI chip - a SN74161 from 1974 w27

[@fabimre]

The color and style of the case already gave away the brand! The "wool" look to me like "Rock-Wool".

[@8bitMicroFan]

The diode at 13:50 in the row of vertical components looks like a germanium point contact diode. I wonder why they used one...

[@ollieb9875]

This was fun. Thanks man 😀

[@elonmuskmtmt886]

Given its heavy use of logic chips, and apparent analog amp components, my guess is that it is some type of early digital radio comunications expiriment.

[@TheError404]

11:34 maybe they are in-circuit adjustable precision resistors (9.95k-10.05k)?

[@andthen0170]

You put the channel selector back in the wrong slots. Just saying. 5:23

[@GeoffreyFeldmanMA]

It's Spray-gue capacitors. I went to school near their Headquarters in North Adams, MA, USA so trust me.

[@MarkTillotson]

Spray goo?

[@sydmichel]

"Have you figured out what it is yet", reminds me of Rolf.

[@xDevscom_EE]

HP thingy :) Thanks for teardown.

[@Screamingtut]

IC 74161 - Synch 4-bit Counter, IC 7404 Hex Inverter

[@dhpbear2]

Maybe this is a bit of HP's in-house test gear?

[@DreitTheDarkDragon]

Might that be some kind of industrial circuit tester? *Dave reveals what it is just few seconds later* In that case, readings per second could change integrator slope?

[@lroy730]

I new it was HP by the case . But an AD converter ? I thought it was some kind of counter.

[@dhpbear2]

00:18 - It's looks like Hewlett-Packard. With an unusually sparse front panel!

[@petroldevo9934]

It's a flux capacitor controller....

[@DShadowWolf]

Part of a lab data system - found a reference to a "3352B Laboratory Data System" that has the 18652 A/D converter as a component...

[@KnightsWithoutATable]

Guess at the 00:05 mark - it is a single channel transmission amplifier for a microwave telecommunications system.

[@TheDefpom]

I think it needs new caps

[@Aardvarkzombies]

Does the video go choppy for someone else??

[@keshabapatel8322]

What is the input impedance back in the days?

[@micahjoel7507]

Interesting. Could a 2018 Arduino do everything this piece of equipment does?

[@johnfrancisdoe1563]

Micah Joel Maybe the digital part, but double-ramp A/D at 8388608Hz / 4sps = 2Mi clocks per sample = 6 digit / 20 bit precision if a double ramp converter. That temperature controlled isolated box with the heating resistor next to the precision capacitor is the difficult part. Although you could wire up a 20bit SAR ADC chip from Analog Devices to the SPI bus (I think there's an Arduino SPI library already) with a nice voltage reference, but you'll still want to control the system temperature by driving a power transistor from a PWM output and using the μC built-in temp sensor to keep a steady temperature, and the μC EEPROM to store adjustment values from calibration testing of each device plus a digital signature from the calibration lab. Also sync the onchip CPU clock PLL to an external PPS or 10MHz signal from your lab time reference, to ensure consistent clocking of the ADC chip (syncing to the USB 1kHz signal is not precise enough for a 1ppm system).

[@dhpbear2]

11:25 - Yes, those are trimpots.

[@papaalphaoscar5537]

Glass wool. Avoid rubbing your eyes! :-D

[@OM7AHU]

Good video.

[@teuton8363]

where is the big thumb? I only see the regular one!

[@JamesDoebbler]

Looks to be part of a system used in lab liquid/gas chromatography. No specs, but very brief mention: www.willstein.com/Resources/HP 5890 Cable Configurations.pdf

[@SaberusTerras]

That board is totally baked. I wonder if it was by design or just the former operator didn't take good care of it.

[@mfx1]

Unmistakably HP style case.

[@frenchcreekvalley]

The "red goop" us Glyptal.

[@RoyHess666]

3:09 what does "bula" mean? Why do you say that so often twice? Is this some sort of new trend?

[@sonnyjimbod]

It's from 'Ferris Bueller's day off'

[@r3bers]

Guessed ADC before black tape gone.

[@zvpunry1971]

Your whistling screwdriver was better. :)

[@SebastianScholle]

@Eevblog electric screwdriver, life changed!

[@PedroDaGr8]

It's an ADC for an HP gas chromatography system. I found the 18652A referenced in the HP 5960 Series II GC manual. Based on the description, it digitizes a 0-1V differential analog signal (common in old analog chemical lab equipment). HP was (and still is, as Agilent) one of the market leaders in Chromatography (Liquid, Gas, HPLC, UPLC, etc.) Chromatography systems are normally VERY modular, you add what you need and nothing more. Want to change detectors from light absorption to electrical conductivity, switch out the detector. Want to add an auto-sampler for unattended operation, easy. As such, it was likely sold as part of one of these systems as an optional add-on, which is why the documentation is harder to find. Honestly, you can see the influence of this modular format even in today's Agilent HPLCs.

[@thenaimis]

don't need to see a label to know that's HP :-)