That components below R5 are not easy to solder , but they look aligned . The input jack looks nice , good solder joints. Great Photos

Since my last post, I've known that the trace from R10 to the asic was in bad shape. When measured, the line wasn't open but there was a significant amount of resistance. I'm assuming I messed it up accidentally when soldering on the resistors. This really only left me with 1 option, which is to pass a jumper from R10 to the 5th pin from the top on the left of the ASIC. Easier said than done
Of course, the pins on the ASIC are extremely small and close to one another and I don't have any fancy micro-soldering equipment...
So anyways, that's what I did today and I somehow managed not to short a bunch of pins in the process

Its a bit hard to see, but I used 36 AWG insulated jumper wire.



Here is a mechanical pencil lead next to the joint for scale



I measured to make sure everything was good and closed it back up... Unfortunately, this didn't fix my issue. Just like before, the input lines still work and at least one of the output lines doesn't. On the bright side, I can rule that out as a potential issue.
There's almost certainly a couple cold welds though since the soldering job leaves a lot to be desired (done with basic soldering tools and no flux). I'll probaly try to reflow the ones that look the worst and if that doesn't work, I'd probably just try again from scratch with proper tools and some flux (I have a bunch of half-dead calculators anyway)

Its been a few years, I have gotten much better at micro soldering.
I decided to revisit this project by attempting it again from scratch with two new calculators (I do not trust what I did back then and trying to fix it would be harder than just trying again)

This time, it worked!

Here is the board before and after, and a close-up view of the comm circuit that was added:


I did exactly what I described in this post except moving J14, stealing the circuit from a rev Q TI-82 instead of rev U (because its what I had on hand, it is the same circuit)

Of course, I had to cut a hole out of the bottom of the calc to accommodate the link port:


As some of you may know, the TI-80 only supports one operation through the link port, taking screenshots, so that's exactly what I did.
I managed to obtain this screenshot:


Finally, here is a video of the screenshot being taken:

He he, nice work

That's really cool! Glad to see it work this time!

After seeing this thread and having gone through the TI-80 schematic, it occurred to me that one could add a temporary graph link to a TI-80 on a breadboard. The unfitted components comprise of diode for protection on the inputs, transistors to control the output, pullups, and L/C filters on the outputs. This uses 4 pins (in+out tip, in+out ring) similar to the design in the gray serial graph link which means enough circuitry needs to be added to resolve these 4 pins on the calculator to the 2 signal + GND pins of the graph link.



Since this is temporary, I omitted the L/C filters and L on GND. The pullups are already on the board, so just need to worry about 1 diode, 1 NPN transistor, and two resistors for each tip/ring. mr womp clued me in that TiLP supported TI-80 screenshots and this was a success. I had previously sacrificed the graph link jack in my black serial graph link, so I had to solder the red/white/black wires to it in order to interface with the breadboard which can be seen behind the breadboard. The other two wires on the TI-80 PCB provide 5V power. Success!

I was curious if new parts could be ordered to add a graph link to a TI-80 rather than salvaging them from an existing ViewScreen or Graph Link cable. Based on Zeroko’s and mrwhompwhomp’s levely photos of TI-80 ViewScreen calculators from earlier in this thread, the graph link circuits have the following parts. I was wondering if anyone had looked these up previously to confirm?

• R5 (pull-up already populated): `103` = 10kΩ
  
• R6: `471` = 470Ω, 0805
  
  • Replacement: YAGEO 470Ω RC0805JR-07470RL
• R7: `333` = 33kΩ, 0805
  
  • Replacement: YAGEO 33kΩ RC0805JR-0733KL
• CR4: `M1M` or `L4` - same diode as used in gray and black serial graph link cables for powering from serial pins. Likely Schottky Diode
 BAT54.
  
  • Replacement: (Many manufacturers) BAT54
• Q2: `LG`NPN transistor. Based on marking font, looks like Toshiba 2SC4116-GR
  
  • Replacement: Toshiba 2SC2712-GR,LF
• JACK: 2.5mm 3-conductor, no switch audio jack. Same one as used in black serial graph link and TI keyboard
  • Replacement: Kycon STX-2500-3N
• L1: `2R7` = 2.7µH 1210 inductor, same as used in gray serial graph link and silver USB graph link
  • Replacement: TDK NLV32T-2R7J-EF
• C15: no marking (as usual), likely same as ones in silver graph link

As the TI-80 calculator was manufactured in the mid-1990s, its components have been extremely difficult to find in catalogs. The resistors are the easiest since those have been standardized and well labeled, but the small surface mount packages for transistors and diodes have few symbols and no indicators on manufacturer. The inductor and capacitor just form a filter on the inputs and so are not necessary for the link port to work, however, may provide some protection when first connecting to limit current. The transistor I'm fairly confident based on the font of the package marking, however, the diode I'm still not certain of. The parts total about 2.50 USD (+tax+shipping), but before I order, I'm going to try to confirm the electrical characteristics of the capacitors and diodes.

I came across an early TI-80 with date code `I-0195` which didn't have the `A` suffix. I saw this thread mentioned "Revision A was produced between January 1995 and April 1995" but hadn't observed any pictures of the January builds so not sure if all the January ones are like this. Similar to the Rev A, this uses ROM 3.0, PCB "APLMB-30F 1-3" and solder on the pads for the components for the graph link circuit. No change in jumpers or resistors.

What is unique to this one is the ROM chip is not the usual LH5359, but an Atmel AT27LV256R. This is apparently UV erasable EPROM, but in a package without a window so it is effectively OTP. Full image album https://imgur.com/a/7xMt3gJ

Hm, interesting, we knew about pre-A with PCB suffix 1-1 and 2-1, but not 1-3 yet.
I just added it to the spreadsheet, thanks

Thanks, indeed I see a few other pre-A in that spreadsheet. mrwompwomp also confirmed on discord the pre-A with PCB 2-1. The Atmel EPROM was a bit surprising, but this is likely why A15 has the jumper configuration and is tied high. The EPROM, this is VPP and per the datasheet must be set to VCC during typical read, output disable, and standby operation.

Back on the topic of adding a graph link port to a non-ViewScreen TI-80, I ordered the following modern parts and they recently arrived. I put them in a I-0495A model I have and the graph link port works well as tested with TiLP and Black Link cable! The 2.5mm audio jack was my biggest concern, but both its electrical and mechanical footprints fit the pads and holes. The capacitors I ordered for the DBus filter were based on measurements for the TI silver link, but they had too much capacitance, so I omitted them for now. If any of you have a TI-80 ViewScreen with the C15 and C16 capacitors, I'd be interested in getting a rough measurement, out of circuit if possible.

Code:

That's awesome! Looks like a great soldering job and the hole in the case is very clean!


I had scribbled down the values on that post-it note, but unfortunately this is the only picture I have showing the note.