Interesting find: HP 9875A dual cassette drive



#2

I took a chance and bought this HP 9875 Cartridge Tape Unit from a recycled equipment company recently.

It's an HPIB device designed to be used with the 9820, 9825, 9830, etc. with the appropriate interface, so the back panel will look familiar to those of us who've used HPIB devices before.

The 9875A is an "intelligent" peripheral with its own processor and memory. This explains the giant circuit board inside. In fact according to HP it could be connected directly to HPIB measuring equipment and accumulate data on its own without a host computer. Flipping a "Self test" switch on the back will cause the unit to perform write and read tests on cassettes in both drives and blink a drive light if any errors occur.

Inside are two 9825A-era cassette drives that use DC100 tapes. Fortunately I have one good DC100 tape! Both tape wheels were goopy, but the good 'ol "heat shrink tubing" fix, inelegant though it may be, fixed that.

Being expensive, 70s-era HP equipment, it's built like a tank, with a thick aluminum frame supporting all components.

A clever feature: angled mirrors and clear ports above the tape drives allow you to see the label on the tape, if any, or the spinning tape reels if not.

Although the device predates the HP-85, it's a lot easier to play with on an 85 (with the optional HPIB interface) or HP-87 than on earlier machines, because these let you do BASIC things from the command line, like dimension string variables to hold status results, which can only be accomplished inside a program on earlier machines.

Although "intelligent", the 9875A doesn't have anything in the way of file system smarts, nor is it supported by "mass storage" ROMs. When you insert a tape, the first thing you need to do is send a "rewind" command, because otherwise the unit doesn't know where the tape is. You create files, specifying the number of records and bytes per record, and can write to and read from files or records programmatically. Various operations let you select optional read-after-write verification, select drive 0 or drive 1, and retrieve error codes, which are typically delivered as a single long string consisting of comma-separated numbers you must then parse yourself.

You can't (easily) save or load programs from the unit; and using data files is quite tedious: for example, you must programmatically position the tape to the start of a file with the "find file" command before reading or writing to the file; and files have no names and there's no concept of a "directory", so you'd better make notes.

Still, it's an interesting little bit of HP history, and the only one I've ever seen. $150 well spent.


Edited: 2 Nov 2011, 5:58 p.m.


#3

Nice find!

What is the part number of the white ceramic chip? It is probably a HP proprietary microprocessor, and it is in the "SLAM" (Single Layer Alumina Metallized) package. This package was mostly used by HP, but Motorola used it for a short while for their DSP56000/DSP56001 DSP processors.

In the mid-1970s when this product was designed, there weren't any commonly available IC packages with more than 42 contacts. The SLAM package was developed to provide a larger number of contacts, as well as to package chips with higher power dissipation than could be accommodated by the more common plastic packages. By using a single metallization layer etched on the alumina substrate, it was a fairly inexpensive package to produce, compared to other high-contact-count ceramic packages of that time.

The industry generally went a different direction, with the JEDEC CLCC (Ceramic Leadless Chip Carrier) packages and ceramic pin grid arrays. Today ceramic packages are rarely seen in commercial products, with the notable exception of microprocessors for general purpose computing, where they are used due to power dissipation requirements of 50 to 145 watts.

The Motorola 56K DSP parts used a 100-contact SLAM package, but this one might have fewer contacts.

Edited: 2 Nov 2011, 8:33 p.m.


#4

Here ya go (in passing, the iPhone 4s camera is really quite good...)


#5

I was wrong, that isn't a microprocessor. I've seen that one before in other HP gear. It is a "CHI" HP-IB (IEEE-488) interface chip. Possibly a 48-contact SLAM package. It is probably NMOS. Given that it doesn't even include the bus buffers, it seems unlikely that it would draw more than 150mA at 5V, so the reason for use of the SLAM package must be the contact count rather than power dissipation.


#6

You're sure that's a CHI chip? Given the vintage, I would have thought it was a PHI (Processor-to-HPIB Interface) chip made with HP's proprietary silicon-on-sapphire process technology for high speed and low power.

That package was first used in the HP 9835A/B for the language-ROM drawer and to expand the capacity of the front option-ROM drawers. In those days, the option ROMs were high-margin products so finding a way of putting four different option ROMs in each ROM drawer was akin to printing money.

I used the same package, but with a pc-board substrate for the laser-trimmed precision-resistor networks in the never-produced HP 98037A Analog I/O card (http://www.hp9825.com/html/analog_i_o.html) for the HP 98x5A/B desktop computers. It was a good way to get a high-density, pluggable component in those days.


#7

The service manual for another HP product (I forget which one), referred to that part number as a CHI. It didn't give the expansion of the abbreviation. Maybe "Computer to HPIB Interface"?

Certainly functionally it is a processor to HPIB interface.

I thought HP started using SOS in the late 1970s. Were they really using it around 1975?

Regarding the ROM package, I imagine that you must be talking about the rectangular package with holes for the contacts? I've never seen ROMs packaged in the square SLAM package like the HPIB interface chip.


#8

The first HP Journal mentioning silicon on sapphire at HP is the July 1978 issue. (http://www.hpl.hp.com/hpjournal/pdfs/IssuePDFs/1978-07.pdf) That article references another HP article about silicon on sapphire dating back to April, 1977. There's also an HP ad you can Google up from November, 1977 that talks about the PHI chip and silicon on sapphire. That article says that the HP silicon-on-sapphire program dates back five years from the end of November 1977 or some time in 1972.

The HP 9875A tape drive doesn't date back to 1975. It came later, as evidenced by the use of the Tape Controller chip (TACO) that was designed for the HP 9845A. (http://www.hp9845.net/9845/hardware/processors/ and scroll down) I'm pretty darn sure that The HP 9825A (introduced in 1976) used discrete electronics and TTL chips for its tape controller. The NMOS TACO came later. According to the HP Computer Museum, the intro date for the HP 9875A was 1978, so that is the same time period as the HP Journal article. Also, the HP 98034A HPIB interface card for the HP 9825A was based on HP's NMOS Nanoprocessor (as was the 98035A Real Time Clock card and the 9885A floppy drive) because there was no HPIB chip available during development of that I/O card.

All of which doesn't prove that the chip in the HP 9875A is a PHI. It does prove that HP's silicon-on-sapphire chips existed in the proper time period to go into the HP 9875A. My hazy memory recalls hearing a lot about the PHI chip and nothing about the CHI chip at the Desktop Computer Division. I didn't use any of the HPIB chips for my design work at HP. Didn't need to.

As for the ROM in the HP 9875A, indeed I was referring to the ceramic package with the holes. The other ceramic package in the clips which you call SLAM (never heard that term, but sounds good) is clearly the HPIB chip. That's how the PHI was packaged. I don't know about the CHI chip.

Edited: 6 Nov 2011, 9:15 p.m.


#9

Is there any published information on the Nanoprocessor?


#10

As far as I know Eric, details of the Nanoprocessor were never published. There was no external development system available for it, so there was little reason to publish that info.

#11

Eric, I don't know if he actually did but I was told while I worked at HP's Santa Rosa Instrument Division with Larry Bower and Dyke Shaffer that Larry was responsible for the Nanoprocessor. He's long since left HP and now has Larry Bower Engineering in Santa Rosa. Details are via Google. He may have more info on it. Dyke may be a good source as well, having done much of the work on the 9825 and its peripherals (as well as the serialized HP-IB known as RIB: Redwood Interface Bus, but that's another story). Dyke can be found on LinkedIn. Both are tremendous engineers!

A tiny version of the nanoprocessor was used in the Logic Pulser probe back then, too...


#12

The HP9825.COM web site has some pages that talk about using the nanoprocessor in the interface cards for the HP98x5 desktops.


#13

Uh, the HP9825.com site is mine. There's a discussion of the use of the Nanoprocessor in HP 9825A I/O cards on the site but not the internals of the chip. It's more of a sequencer than an actual microprocessor (hence the name). The page I referenced on HP9845.net (http://www.hp9845.net/9845/hardware/processors/) has a pinout for the processor about four-fifths of the way down the page, derived I think from the complete product-level schematic in the back of the HP 98034A HPIB I/O card manual. Oh for the simple days when we could actually include schematics in the back of a manual and they'd be useful.


#14

Steve, thanks for your great site about the 9825, which I found after a nostalgic search a while back. I have very fond memories of using a 9825T at my college work study job in a chemistry lab to manage experiments, perform data collection and post analysis from various analytical instruments, as well as other non-work related tasks. We produced crude hidden line 3D plots of geometric shapes for fun, and even on topic for this forum, plotting 41C barcode on a 9872C. I also used it as a dial up terminal for CS class assignments with a 9871 printer. Those memories prompted me to acquire two each of the 9825T and 9872C to play with in the future. Now all I need is to find a 9895 ROM so I can have some real storage!

Edited: 8 Nov 2011, 6:40 p.m.

#15

I just found the service manual on hpmuseum.net. It lists the HP-IB interface chip as a 1AA7-6001, so apparently some revision was made to the chip during the production life of the product.

Could you also post a closeup of the 40-pin chip with the heat sink on and the funny rectangular chip near it? They are near the edge connector. The service manual lists the processor as an 1818-2500 "Binary Processor Chip" in a ceramic DIP, but doesn't show a heat sink on it. The BPC is one of the chips used in the processor hybrid of the 9825/35/45, and does draw a lot of current, so I'm not surprised that they decided to heat-sink it.

The funny rectangular part is the ROM, which the service manual lists as P/N 1818-2814. It's a 64K ROM, probably 4K*16 of the same type used in the 98x5.

There's another HP custom chip with a heat sink on the "tape control and power" board, the 5061-3012 "Tape Control Chip". I expect this is probably the same chip used for tape control in the 98x5.

Thanks!
Eric


#16

Here you go:

I gotta admit, I've never seen anything quite like that white chip!


#17

It's a ROM chip, HP also used those for the firmware chips of the HP 9835A/B.

-Rik


#18

In the 98x5 computers, the ROMs were packaged that way both because the 28-pin DIP wasn't yet an established standard, and for compactness. Essentially the same reason as for the SLAM package for the CHI part.

In the 9875A, there would have been enough room for a 28-pin DIP if it had existed, but I'm sure it was easier to stick with packaging all of their ROMs in the same way.

If I remember correctly, though, the 9872 series plotters did not use ROMs in that same package, even though it used the same processor.


Possibly Related Threads...
Thread Author Replies Views Last Post
  HP-80 History, Design and Interesting Facts BShoring 1 198 11-30-2013, 08:50 AM
Last Post: Xavier A. (Brazil)
  Interesting Base Conversions - Porting a 1975 HP 25 Program to the HP 35S Eddie W. Shore 1 184 10-13-2013, 07:49 PM
Last Post: BruceH
  Prime cant find Memory available Dougggg 5 292 10-07-2013, 07:24 PM
Last Post: Han
  HP Prime can't find how to use partfrac function ! dg1969 3 261 10-04-2013, 09:25 PM
Last Post: Joe Horn
  OT - A lucky find - Busicom Handy-LE LE-120A Cristian Arezzini 2 221 09-26-2013, 04:43 AM
Last Post: Cristian Arezzini
  Need service information for HP 82161A Cassette Drive Peter N1 2 189 07-14-2013, 04:11 AM
Last Post: Tony Duell
  HP 9885M disk drive David Ramsey 0 136 05-12-2013, 09:38 PM
Last Post: David Ramsey
  9114A Drive aj04062 4 485 11-17-2012, 07:48 PM
Last Post: Raymond Del Tondo
  Craigslist find and fix - HP 19C Matt Fegenbush 12 506 08-13-2012, 02:57 PM
Last Post: Matt Fegenbush
  Where to find the HP 39GII in London Miguel Toro 10 382 08-01-2012, 01:23 AM
Last Post: Thomas Radtke

Forum Jump: