TI-74 RAM Cartridges

[mrvan]

I recently obtained a TI-74 and found another being parted-out that had an 8 KB cartridge. To my surprise both the parted out TI-74 and the RAM cartridge work. I successfully expanded (call addmem) the main memory to the cartridge to have 16 KB. 

 

A few questions:

1. The battery on the RAM cartridge undoubtedly is dead. If the cartridge remains in the TI-74, will its memory be retained by the TI-74's batteries? I'm only interested in the RAM cartridge to expand the size of RAM, not for persistent storage.

2. Can the cartridge be opened reasonably easily so it goes back together without damage, to replace the battery? I read it is somehow soldered on but I can deal with that easy enough.

3. Are newly built TI-74 or TI-95 RAM cartridges being built or kitted, etc. by anyone in our community? 

4. Can the RAM chip on the TI-74 RAM cartridge simply be replaced with a 32 KB version?

 

Thanks in advance for any answers. I get the impression the further I stray from the TI-99, fewer are the folks I'll find working on these machines and able to provide support. 

[brain]

#3, not that I am aware, but if you can get pics and dimensions, I can make a new PCB.

#4 depends on seeing the schematic (which is dependent on #3)

[apersson850]

As long as you have working batteries in the calculator, it will keep the memory alive. You may be able to change them without losing memory, even in the module, if you do it quickly.

I've not opened my RAM module, so I don't know for sure what happens. But there's always a sharp knife and crazy glue.

The battery is probably welded, not soldered. Then the welded connectors are soldered to the PCB.

 

I just realized I've never replaced the battery in my RAM module. But since I keep it in the calculator all the time (it's the only module I have for it), I would probably not notice if it dies.

[mrvan]

9 hours ago, apersson850 said:

As long as you have working batteries in the calculator, it will keep the memory alive. You may be able to change them without losing memory, even in the module, if you do it quickly.

I've not opened my RAM module, so I don't know for sure what happens. But there's always a sharp knife and crazy glue.

The battery is probably welded, not soldered. Then the welded connectors are soldered to the PCB.

 

I just realized I've never replaced the battery in my RAM module. But since I keep it in the calculator all the time (it's the only module I have for it), I would probably not notice if it dies.

Thanks for this response. I only want to use for RAM extension so should be fine. I imagine that any possible corrosion, etc. the battery could have done is long over so probably no need of a rush to address.

[mrvan]

16 hours ago, brain said:

#3, not that I am aware, but if you can get pics and dimensions, I can make a new PCB.

#4 depends on seeing the schematic (which is dependent on #3)

Oh wow, cool.

 

This is the best I've found for a schematic and parts list. https://www.hpmuseum.org/forum/thread-7146.html. Not sure if the board connects all address pins. Supposedly there's a TI-95 cartridge with 32 KB that does fit but all these are pretty difficult to find. Making a new board is probably easiest.

 

I'm willing to open my cartridge and get pics / dimensions. I'll try to be careful, and I have two of the slug cartridges that TI shipped with the TI74s to protect the cartridge connector and make the unit look nice. Pretty much everyone with a TI-74 has these, at least one, so a new board that fits within one of them would be perfect. I could even send you one of mine if you continue to agree to do this.

 

The TI-74 and CC-40 are close brothers that just look different. Very similar BASIC, both have HEXBUS (or DockBus).

 

BTW, are you the one selling the memory cartridges or PCBs for the CC-40? 

[mrvan]

I've first opened a slug cartridge and it opened fairly easily but not without a small amount of damage at the seams. An xacto knife cutting of the debris and perhaps a small amount of paint or even ink touchup should make it more than acceptable to make the unit functional.

 

I then opened the RAM cartridge. It's clear to me that TI had no intention that these be taken apart. They should have had a door at least for swapping batteries. In any case some damage to the snaps, but should go back together ok. Thinking that alteration should be considered so the case can be reopened, even if all snaps removed and just wrap with a nicely made sticker.

 

Jim, I can provide measurements for the board, but first want to know if the link I sent you is adequate for assessing the schematic. The battery is attached to the top of the PCB on top of the memory so will have to assess how to remove without damage. 

[brain]

1 hour ago, mrvan said:

I've first opened a slug cartridge and it opened fairly easily but not without a small amount of damage at the seams. An xacto knife cutting of the debris and perhaps a small amount of paint or even ink touchup should make it more than acceptable to make the unit functional.

 

I then opened the RAM cartridge. It's clear to me that TI had no intention that these be taken apart. They should have had a door at least for swapping batteries. In any case some damage to the snaps, but should go back together ok. Thinking that alteration should be considered so the case can be reopened, even if all snaps removed and just wrap with a nicely made sticker.

 

Jim, I can provide measurements for the board, but first want to know if the link I sent you is adequate for assessing the schematic. The battery is attached to the top of the PCB on top of the memory so will have to assess how to remove without damage. 

I think so.  In the schematic on that page:

 

They left off the address and data lines.  That's probably OK, I just need to verify I have a pinout of the TI74 connector so I can make sure I connect all those up.  If, though, you can trace one of the lines (maybe pin 13 on the IC), that would help me  ensure I get things ordered right.  TI's  D7 being D0 for everyone else messes with me sometimes.  It's a shame TI74 carts have the connector on the cart, as that means it'll be harder to reproduce.  Good measurements of the pitch of the connector (if that info is not already known) and the pad size would be helpful, to see if a connector can be sourced.  I've never done a TI74 cart, so maybe the above information is already known in the community.  If so, a link is fine.

[mrvan]

31 minutes ago, brain said:

I think so.  In the schematic on that page:

 

They left off the address and data lines.  That's probably OK, I just need to verify I have a pinout of the TI74 connector so I can make sure I connect all those up.  If, though, you can trace one of the lines (maybe pin 13 on the IC), that would help me  ensure I get things ordered right.  TI's  D7 being D0 for everyone else messes with me sometimes.  It's a shame TI74 carts have the connector on the cart, as that means it'll be harder to reproduce.  Good measurements of the pitch of the connector (if that info is not already known) and the pad size would be helpful, to see if a connector can be sourced.  I've never done a TI74 cart, so maybe the above information is already known in the community.  If so, a link is fine.

 

Below should be some useful images of the RAM PCB, battery, case and EMI wrapper.

 

The battery is soldered in in the two larger holes and mounted on top of the memory chip. The link I sent you has the EMI wrapper image. There's also a piece of plastic tape that covers the two diodes on the top of the PCB to prevent touching the battery.

 

The PCB length is 1 1/16 inches long x 1 inch wide.  Total length with connector is 1 17/32 inches. There's a small amount of margin inside the case but that's used up by the EMI wrapper. There's no jiggle.

 

The pic of the connector with part # should help on the pitch.

    

        

[mrvan]

Jim,

 

I'm definitely willing to work this through with you. I do like the CC-40 but the form factor of the TI-74 is much nicer. I've read the TI-95's 32 KB cartridge works in the TI-74 so this cartridge in 8 or 32 KB should also open the doors for the TI-95. This memory upgrade paired with the HEXTIr, and your future external display and these computers will really become something worth putting effort into.

[+Ksarul]

You might want to take a look at this document I found on WHT. . .lots of TI-74 technical data in here.

TI74 technical data manual.pdf

[mrvan]

21 hours ago, Ksarul said:

You might want to take a look at this document I found on WHT. . .lots of TI-74 technical data in here.

TI74 technical data manual.pdf 2.07 MB · 2 downloads

There's some excellent info in there. It's nice to see the pinout.

[mrvan]

@brain, I'm back from two weeks of vacation and now a week of work, so have had a bit of time to get back into this.

 

Doing this with a new PCB obviously would be a lot nicer--have you been able to make any progress? I know you have a lot of irons in the fire.

 

But in the interim...

 

I reviewed the technical manual you linked and noted all the proper signals are present at the edge connector so that's a good start. There's also some notes about generation of "wait states" by essentially slowing down the CPU when accessing addresses that are slower than the CPU can handle (it doesn't have a concept of wait states built-in). I also wonder how "CALL ADMEM" works, whether it attempts to write/read in the expansion RAM range until it hits a bad return. Since I've read that 32KB cartridges have been made for the TI-95 and claims are they work on the TI-74 I'll ignore these.

 

Comparing the existing 8KB RAM chip (HM6264LFP-15) and the candidate 32KB (HM62256LFP-15) yields minimal differences:

The 8 KB memory has two fewer address pins, as expected, but also has two CEs (-CE and CE).

The 32 KB memory has the two additional address pins and eliminates CE. A14 is assigned to pin 1 which is NC on the 8 KB chip. A13 is assigned pin 26 which is CE on the 8 KB chip. 

 

The simple hand-drawn schematic seems simple enough. To alter the board, it looks like memory pin position 26 should be eliminated from the existing circuit and connected to A13 on the edge connector. A14 needs to be connected from the edge connector. It's a shame the memory chip is surface mounted. That'll probably require drilling a hole to run a wire from edge connector pin 29 A14 to the chip's A14--there's very little space on the sides of the board, but maybe there's a little room at the back of the board to run a wire from bottom to top. The board is packed very tightly in the module.

 

I should note that I've obtained a second 8 KB RAM cartridge so have some additional freedom for experimenting. Also have three TI-74s.

 

I'll look into purchasing a few of the 32 KB RAM chips. 

[brain]

I have, but I got stuck at the schematic of the cart port connector in the technical doc you linked (which I actually already had in my possession, but my copy if just as blurry.  I have it laid out, but can't fully connect the pins until I can decipher the A0-A15 and D0-D7 pins on the schematic.  I looked on the Internet for a cleaner copy, no luck so far.  Any help would be appreciated.

[mrvan]

Are you looking for an edge pin connector pin to chip pins mapping? I can provide that I think. 

[brain]

Yep.  I got the ones in the schematic I linked to (pwr, gnd, cs, r/w), but the addy and data line mappings are just a bunch of splotches.

 

[mrvan]

I believe this is what you're asking for. 

 

[brain]

https://github.com/go4retro/TI74_RAM

 

[brain]

So, one item to validate:

 

Schematic says 1mF and 100mF, but the text that the schematic was linked off of says 1nF and 100nF  Big difference

 

Jim

[mrvan]

On 4/29/2024 at 7:44 PM, brain said:

https://github.com/go4retro/TI74_RAM

 

@brain, X1 doesn't seem to comply with page 11 of the Technical Data Manual, in terms of pin definitions for addresses and data. For instance, A14 should be pin 29 but is pin 12 on X1.

 

The inputs to IC1 coming from X1 are not mapped 1:1, and A14 is connected to GND. The data pins are also not 1:1. I suppose in the end as long as every address and data pin is mapped uniquely, it doesn't really matter. A14 is a problem obviously.

 

As to verifying the capacitors, how can that be done? There's the pictures of the board with parts but I don't see values or identifying parts.

[brain]

To ease with routing, I took advantage of the fact that RAM doesn't care about address and data lines.  So, I remapped the data lines to clean up the routing, and the same for the address lines. 

 

A14 is connected to the ground plane, but it does not show in the PCB layout, as I left the ground pour off the pic.

 

You *might* be able to reach the original poster, as the post was in 2016, so perhaps he's still around.  Alternatively, you could try to measure with a DMM, if yours has a capacitance option.

 

Otherwise, we can just wing it.

[mrvan]

38 minutes ago, brain said:

To ease with routing, I took advantage of the fact that RAM doesn't care about address and data lines.  So, I remapped the data lines to clean up the routing, and the same for the address lines. 

 

A14 is connected to the ground plane, but it does not show in the PCB layout, as I left the ground pour off the pic.

 

You *might* be able to reach the original poster, as the post was in 2016, so perhaps he's still around.  Alternatively, you could try to measure with a DMM, if yours has a capacitance option.

 

Otherwise, we can just wing it.

Makes sense on the remapping day and address lines. 
 

I don’t understand the thinking on A14 being connected to the ground plane. Seems that would be separate from ground?

 

I’ve not used a dmm for a capacitor. I have a mid level Fluke not sure if it has the appropriate feature will check tonight. 
 

My RAM arrived yesterday so I can assess the capacitor soon before altering my RAM cartridge. 
 

I have multiple 74’s and 95’s now so will need several of these cartridges when we’re done. I hope others will join and this will grow the ‘40-95 user base, particularly if you get that external display going :-)

[brain]

20 minutes ago, mrvan said:

I don’t understand the thinking on A14 being connected to the ground plane. Seems that would be separate from ground?

My bad.  I thought the CS line would only trigger for 16kB, and I used a 32kB footprint (32kB is cheapest SRAM option at Digikey).  But, it appears the memory map allows a full 32kB RAM to be there, so I added it in (moved a few lines to make routing easy)

 

git updated

 

[mrvan]

I was thinking of the Fluke meter I have at home on the bench. Older but higher end unit. Never noticed a capacitance check but never needed one.

 

But I have a fluke 101 I keep in my truck for basis electrical issues with my toy hauler trailer. It has the capacitor check. 
 

I should be able to test on the weekend.

[mrvan]

I had a few hours available this morning to step into this a bit. I replaced the 8 KB RAM chip with a 32. It got fairly involved as multiple solder pads popped up and off, so had to run more cables. The board truly was quite fragile. I remember repairing some irrigation controllers built in the mid 1980s with PCBs that with the slightest anything gave up their traces. Maybe something of the era. Thankfully my TI-99s are not like this. To excise the CE connection I had to cut two traces and jump, as pad 23 for the memory was in line in the CE/-CE circuit rather than connected at an edge. 

 

Calling addmem only resulted in an additional 8 KB (total 16 KB). At least the board lives, but it is possible I goofed up A13 somehow so have some debugging to do. I read somewhere that at least with the TI-95 one had to set certain values for the 32 KB to be recognized. Maybe that's so for the TI-74. I should have saved links to all relevant but perhaps I missed one.

 

Hopefully I can get around to checking the capacitors later in the weekend and have pretty much ran out of time at least for today.

 

If this project only yields 8 KB RAM cartridges, I will be disappointed, but it will still be successful as there appears to be many TI-74 and TI-95 units out there without one. 8 KB vs 16 is quite a bit. My CC-40 with 18 KB seems to have plenty. At this point I still need 3-4 RAM cartridges.

 

Attached below is a pic of my now Frankenstein memory cartridge installed in one of my TI-74s. It'll never fit in the cartridge case again. It would have helped to have lower gauge wires but I also didn't expect to have to run so many.

[+OLD CS1]

I, for one, would be happy to have a RAM card for my 74 and 95.  It seems impossible to find an original one these days.