HP34703A Multimeter - Replacement ROM?

[rjmusto]

Hi all,
I have a soft spot for this range of HP test gear and the 34703 DMM is a real treat.

The one I have though is faulty - powers on, but shows no sign of functionality.

Having been through the logic fault finding sequence, it suggests the ROM (A3U11) is faulty.

Anybody know if there is a solution to this, or is it game over?

Thanks,
Ralph

[factory]

Since your last thread on this meter I've also bought a HP 34703A, I imported mine from the US and as I didn't have a manual I've bought the pdf from Artek.

Before condemning the ROM, have you checked all the supply rails are present & correct? As well as using +5V the ROMs have a couple extra rails of -2V and +10V, these are from a resistor divider, shown above the A1U19 ROM on the diagram.
Also check there isn't a bad connection due to the socket (if fitted), the connections between the boards and the other logic ICs on the input are working.



The ROM in question "A3U11" is HP 1818-2132, a 28 pin 4096 bit (256x16) NMOS part, unfortunately the microfiche* shows it was made by division HP09 (an internally made part). I've no idea if there might be a commercially available PROM that could be used instead.



But there is a NOS one available from a chap on eevblog, he actually has both ROMs used in this meter, I've no idea if you are on eevblog, but if your not I believe you have to have a certain number of posts before being able to PM a member (this is to reduce zero post scammers from making PMs about offered or wanted items);
https://www.eevblog.com/forum/buysellwanted/fs-new-test-equi...rts-hp-agilent-tektronix-fluke-etc/

*A few of the HP parts lists on microfiche were sent to CuroiusMarc, he scanned these and made them available on his webpage here;
https://www.curiousmarc.com/doc-archive

David

[TonyDuell]

Originally Posted by factory

The ROM in question "A3U11" is HP 1818-2132, a 28 pin 4096 bit (256x16) NMOS part, unfortunately the microfiche* shows it was made by division HP09 (an internally made part). I've no idea if there might be a commercially available PROM that could be used instead.

That sounds like the HP ROMs used in the HP9800 desktop calculators. Internally they are made as a pair of 256*8 mask-programmed ROMs, the top metalisation layer and bond out to the pins can connect them either as a 512*8 ROM (needing both A8 and A8/ to select one or other half) in a 24 pin package or as a 256*16 ROM in a 28 pin package.

At least one of the patents for the HP9810 or HP9830 gives the internal circuitry of the ROM chip. For interest here of course.

I've seen them in other HP instruments, the HP59401 HPIB analyser has a few of both configurations.

There is no PROM or EPROM that will directly replace them. The outputs are open-drain (similar to open collector on a bipolar device) not 3-state. And the enable inputs are 12V logic levels. People have used TTL open-collector buffers for the former issue and discrete transistors for the latter to interface normal EPROMs to the above calculators,

[rjmusto]

Thanks for the replies, much appreciated. And the tip on getting spares from the guy on eevblog is perfect. Thanks David, that really saves the day. (The two supplies to the ROM are good by the way.)

Ralph

[factory]

Originally Posted by TonyDuell

That sounds like the HP ROMs used in the HP9800 desktop calculators. Internally they are made as a pair of 256*8 mask-programmed ROMs, the top metalisation layer and bond out to the pins can connect them either as a 512*8 ROM (needing both A8 and A8/ to select one or other half) in a 24 pin package or as a 256*16 ROM in a 28 pin package.

Your right, HP designed their own ROM, due to lack of suitable commercial parts at the time.

I've found one of the patents you mention (for the 9820), it uses the 24 pin 512*8 ROMs with a similar resistor divider arrangement to get -2V. But no +10V as that is pin 28, on the 28 pin 256*16 version used in the multimeter module.


Creating a replacement for A3U11 would also need the code reading from a known good part, as they didn't add this info in the service manual, there is however a page with the info for the other ROM.

The 8 inputs of A3U11 are using TTL levels from four DM74L74, these are all National Semiconductor parts in mine, they seem to have quite a bad reputation for parts failing with age.
The outputs are mostly going to TTL ICs, one goes to a transistor and one goes off board (haven't traced where).

David

[rjmusto]

Haven’t managed to get a response from the guy on eevblog who had the stash of spares.

Any other options anybody know about?

I’d have a go at a daughter board with a newer part if I could get hold of the contents somehow. Might be turning into a labour of love here ?

[factory]

Did you send him a private message? Or just the reply to the thread, as he noted thread reply notifications weren’t getting sent to his email.

Next question would be, how to read the contents of the ROMs? There is some data for A1U19 in the manual to compare results.
I used to have access to a 386 based PC at work, with an IC programmer built in, even that would probably have needed some kind of adaptor board to read.
There is a standalone programmer of similar type here, but I don't have the PC ISA interface board that goes with it, or a functioning PC to use it with yet.

Wonder if anyone on the HPAK groups .io has attempted reading similar ROMs, then there is the vintage HP calculator/computer forums too.

David

[TonyDuell]

The main problems with reading out this ROM in a normal program would be :

It's 16 bits wide (many programmers assume 8 bit data)
You need +12V and -2V supply voltages along with the normal +5V
You need to swing the CS pin(s) to +12V to enable the ROM.

An adapter is possible, but it might be easier to just use an Arduino board to step through the addresses and grab the data.

[factory]

Some searching of the HPAK groups.io reveals a possible MCU based device to read the contents of these ROMs, which are also found in the 5328 counter (which has different ROMs depending on the options).
I've also started a discussion other there about the reading the ROMs from the HP 34703A DC volt/amp ohm module.

https://groups.io/g/HP-Agilent-Keysight-equipment/topic/reading_1818_roms_from_hp/114328082

I still need to figure out the ROM enable function, U19 gives pin 14 as 'enable line', the code listing for U19 gives this pin as the first decoder input, no data is listed in this column. For U11 pin 14 is one of the eight inputs, both are from 5V logic.
Could the +10V on pin be the enable input? I've not found a pinout diagram that explains the pinout of the 28 pin 256*16 ROM yet. The 5328 uses +12V on pin 28 for both of it's 256*16 ROMs

David

[TonyDuell]

Based on my experience with the HP9800 calculators which are full of the 512*8 version of this ROM, I would think that pin 28 is the enable input. It has to be pulled up to 12V (maybe 10V) to enable the ROM.

[factory]

In this module the +10V (12V enable) on pin 28 is permanently connected to the +10V supply rail, this is the same for both ROMs.

I'm glad you mentioned the HP59401, CuriousMarc has the service manual pdf, it has pinouts for both the 28pin 256*16 and 24pin 512*8 NMOS mask ROMs. I've not got round to finding out if the manual explains the pinout in any more detail (edit: it doesn't).
Pinout of the 256*16 matches that of the the faulty U11 ROM.



It seems HP also had problems with glitches/unwanted pulses on ROM outputs, several of the service notes (available at KO4BB) for the 34703 mention adding capacitors between output lines & ground to cure this, later revisions have these built in, mine is an early one (I've not checked if it got modified).

Forgot to mention, I've got no experience of using Arduino or other MCU boards.

David

[rjmusto]

Did you send him a private message? Or just the reply to the thread, as he noted thread reply notifications weren’t getting sent to his email.

I did send a PM. I guess I'll wait a bit longer, then try again. Don't want to hassle the guy.

[rjmusto]

If my understanding is right, there are no standard EPROMs available with 15 (ok more likely 16), parallel data output lines.

So to replace it, we'd need to use, say, two 27C256 chips stacked. A small daughter board could be made to do that I guess.

Similarly, a test jig allowing a TL866 programmer to read a good U11 in two sections should be doable.

Gets a bit involved certainly, but is the idea correct?

And we'd need a volunteer to read a good ROM

[TonyDuell]

There are a few 16-bit wide EPROMs, like the Microchip AT27C1024. But I think a pair of 8-bit EPROMs would be a lot easier. You'd need a transistor to convert the 12V chip-select signal to 5V levels but not much else.

[rjmusto]

I see the AT27C1024 is available from Mouser for under €4, so not expensive.

Am I right in thinking the U11 ROM is 256x16, so 256 addresses by 16bits wide?
But the AT27C1024 has 64k x 16, so 65,728 addresses by 16bits wide - huge by comparison. So will the 8 address lines we are using map across to the same 256 addresses when using this chip? Or how does all this fit together? Have tried to read up on this, but am still confused.

I'd like to have a go at making a jig to read the contents of the ROM I have on the TL866, but it would help if I knew what I was doing....

[factory]

Yes I've been trying to collect all the info in this thread.
In reply #2, the HP fiche refers to them as "256 words x 16 bits".

The pinout diagram in reply #11 will hopefully help getting the input/output pins matched up with a new part.
The modern 16 bits parts seem only available in "one time programmable" versions.

David

[rjmusto]

It occurred to me that Stephan's Retro Chip Tester is the obvious answer for reading one of these ROMs and at least we would then have a copy of the contents.

I checked on the RCT forum and, yes, it will read a 256x16 ROM. I just () need to create a custom chip definition and an adapter to generate the required voltages.

So will have a go at this and see what it reads from my, supposedly faulty, chip.

[factory]

Sounds like your almost sorted for reading them, a link to the RCT forum I found via web search; https://www.reddit.com/r/RetroChipTester/comments/1mdgkf6/testing_a_hp_special_256x16_rom/

HP just used resistive dividers for the -2V & +10V, zeners seems overkill.



Something like this would probably do for the RCT with -5V & +12V available.



The custom definition should probably reference it's only for the 256x16 HP custom NMOS ROM, as "1818" is a HP part number prefix for many memory & control ICs, many of which are regular commercial parts with a HP part number.

P.S. Might take me a while to sort my reader, I've got a HiLo All-03 somewhere, it's supposed to work with 16 bit devices.
I've ordered a blank PCB for the HiLo ISA board, problems to solve are; the PC with ISA slot (have an untested 386 motherboard) and a preprogrammed PAL IC for the ISA board, not sure what to do about that one yet. And of course an adaptor board for the HP ROM.

David

[rjmusto]

Have tried to read the ROM contents with the RCT, but not getting anything - or at least it reports that the ROM is empty.

I am wondering though if the addressing has to be inverted, that is, active low. But that does not seem to be an option on the RCT.

My belief is that the chip is faulty of course, so it could be that this just confirms that - kind of a false positive.

[factory]

Did you try adding some pull up resistors to the outputs, as it was mentioned they could be open collector.

Even if your reader can't invert the address lines, you could still add some inverting buffer ICs to test try that way, HP did use negative logic for some earlier parts.

David