HP-75 RAM expansion?

[Martin Hepperle]

Dear HP-75 experts,

some weeks ago I read somewhere among my CD/DVD files an old publication about a home made RAM upgrade for the HP-75 using two RAM chips. These were to be soldered piggyback on existing RAM chips and the corresponding Chip Select lines routed to the PCB.
The article also explained that one has to mill out part of the case down to the metal cover plate to provide enough space.

Now I cannot remember where I read this. I believe that this was in some conference proceedings, but I cannot find them anymore.

Does anyone remember?

Thank you,
Martin

[Paul Berger (Canada)]

The only memory upgrade that I can think of that would be that simple would be adding the equivalent to the 8K module, but even that would not be that straight forward. The MAR chip in the 75 that is the interface to the RAM is set up for using 2K chips, so it provides 4 chip select lines to the connector for the 8K module. If the upgrade was done using 2 chips that would suggest they are 4Kx8 and addition logic would be required to encode the 4 select lines into 2 4K select lines and the extra address bit on the 4K chip.

[Michael Fehlhammer]

Hmm, I don't understand all these hardware intricacies, but a general question:
I thought 24 kB is the maximum RAM an HP 75 can address due to limitations of the address space.
Could more RAM be managed by the operating system (without patching it)? Wouldn't a complex bank switching mechanism have to be implemented?

[Sylvain Cote]

(10-12-2016 08:59 AM)Martin Hepperle Wrote:  some weeks ago I read somewhere among my CD/DVD files an old publication about a home made RAM upgrade for the HP-75 using two RAM chips.
These were to be soldered piggyback on existing RAM chips and the corresponding Chip Select lines routed to the PCB.
The article also explained that one has to mill out part of the case down to the metal cover plate to provide enough space.
Now I cannot remember where I read this. I believe that this was in some conference proceedings, but I cannot find them anymore.

I have looked at all the 1980's proceeding and found nothing on this, but I will have another look just in case I have missed it.

(10-13-2016 10:38 AM)Michael Fehlhammer Wrote:  I thought 24 kB is the maximum RAM an HP 75 can address due to limitations of the address space.

You are correct.

(10-13-2016 10:38 AM)Michael Fehlhammer Wrote:  Could more RAM be managed by the operating system (without patching it)?

Not to my knowledge.

(10-13-2016 10:38 AM)Michael Fehlhammer Wrote:  Wouldn't a complex bank switching mechanism have to be implemented?

Yep! and you also have to manage the file system and the runtime space before doing the BS.
IMHO this should be part of the OS, it may be possible to do it in a LEX file but the person doing it must really understand how the OS works.

I am following this thread to see where it goes ...

Sylvain

[Martin Hepperle]

(10-13-2016 10:38 AM)Michael Fehlhammer Wrote:  Hmm, I don't understand all these hardware intricacies, but a general question:
I thought 24 kB is the maximum RAM an HP 75 can address due to limitations of the address space.
Could more RAM be managed by the operating system (without patching it)? Wouldn't a complex bank switching mechanism have to be implemented?

Michael,

yes, the directly accessible RAM memory of the HP-75 is indeed limited to 24K.

Only with some "tricks" additional RAM could be switched into the ROM window as ROM replacement. The Pod and PDS seem to be examples of such applications.

The background of my inquiry:
I have a rather battered up sample of an HP-75 without the RAM module and I wanted to see whether it would be possible to add the missing RAM without locating another original RAM module for $$$.

In fact, after reading said article (or was it a nightmare while I slept?) I bought six (4+2 just-in-case) TOSHIBA TC5516AP 2048x8 BIT CMOS Static RAMs and now I cannot locate the article anymore...
And yes, it were FOUR chips to add, not TWO... my mistake.

On the other hand, if someone would design a PRM-75 similar to the PRM-85 for the HP Series 80 machines or even a FRAM-75 that would be something even more interesting...

Martin

[J-F Garnier]

(10-13-2016 12:13 PM)Martin Hepperle Wrote:  In fact, after reading said article (or was it a nightmare while I slept?) I bought six (4+2 just-in-case) TOSHIBA TC5516AP 2048x8 BIT CMOS Static RAMs and now I cannot locate the article anymore...
And yes, it were FOUR chips to add, not TWO... my mistake.

I remember an article on this subject in the Journal of the French PPC-Toulouse Club, and it was not too difficult to recover it...
Attached are the diagrams, note that it uses a single 8K chip. You may find some information useful.

I once considered to build it, but later I found a cheap 8k RAM module so I didn't do it.

J-F

[Paul Berger (Canada)]

(10-13-2016 03:57 PM)J-F Garnier Wrote:  I remember an article on this subject in the Journal of the French PPC-Toulouse Club, and it was not too difficult to recover it...
Attached are the diagrams, note that it uses a single 8K chip. You may find some information useful.

I once considered to build it, but later I found a cheap 8k RAM module so I didn't do it.

J-F

That looks like what I was thinking of doing to repair a broken 8K module, but then I found a source for 2K chips in SOIC package so I just replaced the defective chips instead.

[Martin Hepperle]

(10-13-2016 03:57 PM)J-F Garnier Wrote:  I remember an article on this subject in the Journal of the French PPC-Toulouse Club, and it was not too difficult to recover it...
Attached are the diagrams, note that it uses a single 8K chip. You may find some information useful.

Thank you - I had gone through les Journals Français (PPC-T and PPC-Paris) available on the internet, but had not seen this one. Another option.

Martin

[J-F Garnier]

(10-14-2016 07:25 AM)Martin Hepperle Wrote:  ... I had gone through les Journals Français (PPC-T and PPC-Paris) available on the internet, but had not seen this one. Another option.

This issue (Micro-Revue N4) is not yet scanned. I scanned these two pages especially for you :-)
I will briefly speak about the French PPC Journals at the Allschwil meeting.

J-F

[Martin Hepperle]

(10-14-2016 09:52 AM)J-F Garnier Wrote:  This issue (Micro-Revue N4) is not yet scanned. I scanned these two pages especially for you :-)
I will briefly speak about the French PPC Journals at the Allschwil meeting.

J-F

Ah, merci beaucoup!

That probably means that I have to travel to Basle to get the full information...

Martin

[Martin Hepperle]

(10-13-2016 12:13 PM)Martin Hepperle Wrote:  ...
In fact, after reading said article (or was it a nightmare while I slept?) I bought six (4+2 just-in-case) TOSHIBA TC5516AP 2048x8 BIT CMOS Static RAMs and now I cannot locate the article anymore...
..

Finally! After searching day and night since October 2016 for this article I finally found it again. Now I can sleep better.
The article appeared in the Jan/Feb 1984 Computer Journal of PPC on page 21. It does not show up in the table of contents because it is disguised as a letter from a reader.
It describes how to implant four 2K 5516 RAM chips into the HP 75 to spare the cost of a 8K RAM module. While the soldering in piggyback fashion would be easy, one would have to mill out part of the case to make the chips fit (or one could machine the DIL cases thinner, but that's more for the explorer type).

Martin

[rprosperi]

(04-21-2017 01:24 PM)Martin Hepperle Wrote:  It does not show up in the table of contents because it is disguised a letter from a reader.

Yes, this is quite frustrating when searching PPC/CHHU, and to a lesser extent Datafile, issues for old articles that one partially remembers. Some very significant 'articles' appeared only in the Letters or NOP (bugs/errors in past issues) columns, and generally were not included in the contents. Some examples include the HP-71 Disassembler, HP patches for the FORTH/Assembler and FORTH/41 Translator ROMs, and some key LEX files.

Though frustratingly tedious, a page-by-page review is needed to find many such items, though the reward is one often finds other jewels tucked neatly away in the same places.

[Paul Berger (Canada)]

(04-21-2017 01:24 PM)Martin Hepperle Wrote:  

(10-13-2016 12:13 PM)Martin Hepperle Wrote:  ...
In fact, after reading said article (or was it a nightmare while I slept?) I bought six (4+2 just-in-case) TOSHIBA TC5516AP 2048x8 BIT CMOS Static RAMs and now I cannot locate the article anymore...
..

Finally! After searching day and night since October 2016 for this article I finally found it again. Now I can sleep better.
The article appeared in the Jan/Feb 1984 Computer Journal of PPC on page 21. It does not show up in the table of contents because it is disguised a letter from a reader.
It describes how to implant four 2K 5516 RAM chips into the HP 75 to spare the cost of a 8K RAM module. While the soldering in piggyback fashion would be easy, one would have to mill out part of the case to make the chips fit (or one could machine the DIL cases thinner, but that's more for the explorer type).

Martin

If you got the chips in a SOIC package you could probably easily work something up that plugs into the connector for the 8K expansion, unlike other edge connectors in the 75 it is a standard 0.1 pitch edge connector. You can use almost any 2K SRAM chip the 5516 was the part number Toshiba used but 6116 is a more common designation. I had some trouble tracking down 2Kx8 SRAMs in SOIC package to repair a couple 8K modules but ended up getting about 25 when I finally tracked down a source. You could also use a single 8K SRAM but you would need a little extra logic to re-encode the 4 chip selects into the top two address lines and chip select.

Paul.