#[1]Leaded Solder
[2]Leaded Solder
[ ]
[3]About[4]Past Events
TI-83 Plus One More
Tags: [5]computer [6]calculator [7]texas-instruments [8]ti83-plus [9]lcd
[10]repair
Although it may not be considered an “old” computer, a TI-83+ from the
futuristic year of 1999 has a lot of appeal. It’s got a Z80, it runs a
BASIC interpreter with machine-language program support, and lots of
fun homemade games were built for it over the years. There’s also the
small matter of me wanting a desk calculator with which to do
binary/hex conversions, so I picked up this broken one to attempt to
nurse it back to health. Will my efforts add up to a working
calculator?
The TI-83 Plus, showing its black screen.
I first attempted powering it on. As advertised, I got the black
screen. The brightness controls didn’t work, so I guess it’s not going
to be that simple. I soon found, however, that I could push 2nd + ON to
turn the calculator off, which means that the CPU and ROM are alive,
and the calculator is listening to its keyboard. This is likely to just
be a display fault.
The TI-83 Plus has a corrupted screen now at half-brightness.
Removing the CR1616 backup battery – which still reported a solid 3.0V
on the multimeter, good work Panasonic – and reinstalling it made
little difference, except now I could tweak the screen brightness and
enter characters. However, nothing was legible and the display just
kept getting progressively more screwed up.
Opening the case took a fairly long T6 Torx driver. Then I worked my
case spudger around the edges, which popped three clips loose. Once the
back is removed, it revealed this surprisingly wrinkly-looking
cardboard RF shield. Did this thing get wet?
The wrinkly, silvery RF shield.
Two small Philips screws come off to get the RF shield off.
The TI-83 Plus is opened.
As you can see, there’s not really much going on inside this big tall
case except for the keyboard. There’s SRAM and a flash ROM, and what
appears to be a big custom glue chip and gate array that probably
implements all of the actual system^[11]1. Everything else, including
the power supply, is scattered around the edges of the keyboard and
battery compartment.
The Z80 CPU on the calculator.
Here’s our friend, a Z80! This is of course a very small surface-mount
CMOS variety, a 44-pin Z84C0008 which is rated to run at 8MHz (it
appears to run at 6MHz.) There isn’t anything else on the board that’s
obviously damaged, although the surface-mount capacitors near the
opening for the backup battery are very sticky to the touch. Another
point in favour of “something got spilled in here.”
On the other side of the case, the keypad has one single, very big
membrane keyboard. It feels pretty good despite this, and the membrane
contacts are very clean.
The membrane pad.
The TI-83 keyboard PCB, with lots of vias
It’s a little weird how many vias are on this board. They don’t seem to
be integral to the keypad matrix, but they might be stitched into the
pours in order to provide additional cooling or noise reduction.
North of the Z80, we see a ribbon cable going to the LCD. There are
surprisingly few pins going to the LCD board, and I reflexively buzzed
them out to make sure they were continuous. Almost half of them (eight
out of seventeen) didn’t seem to work! Could it just be that the glue
holding the ribbon cable to the pads has deteriorated, or the cable
itself got damaged by whatever was spilled inside?
Cable For Mr. Instrument
A microscope close-up of the ribbon cable where it is adhered to a pad.
I suspected that you could repair the conductive glue on the ribbon
cable in the same way that some Game Boy LCD repairs are done. By
rubbing a soldering iron tip over it, you’ll heat-cycle the glue and
apply some extra pressure to make good contact again. It didn’t feel
like a solid solution to me, and since both sides of this connection
have fairly generously-sized solder points… why not just solder some
wires?
Steeling myself, I cut the ribbon with a pair of scissors and then
tried to peel it all off the board, but a thin layer of glue stayed
behind after the plastic was gone. I tried a few techniques to peel
this glue off, but ultimately ended up using a toothpick to scratch the
glue away from the pads. The wood of the toothpick is soft enough not
to damage the pad plating, but is big enough and coarse enough to be a
good scraper.
Scraping glue off of the pad with a toothpick.
Go Ahead And Jump (Jump!)
I started the laborious process of cutting off 17 jumper wires, and
stripping the ends. At first, I skipped tinning them, but that produced
some crappy joints and I ended up having to redo those anyway.
17 blue wires are sticking out of the motherboard.
After about 30 minutes of tennis-elbow-inducing wiring, I had the
motherboard side done. Then I used a screwdriver to bend the wires into
a gentle curve, so they wouldn’t get kinked.
Occasionally a wire would get crossed over and I’d get confused which
one I was working with, because they’re all the same colour. Were I to
do this again, I would suggest picking two colours of wire, and
alternating them, in order to reduce this risk, and also make it look
prettier.
17 blue wires are tacked onto the LCD panel now.
Some of the joints were difficult to make because, even after multiple
rounds of cleaning with alcohol, they still had fragments of glue in
the adjoining “not-pad” region that got melted and pulled in. I’m not
enormously proud of this soldering job, but it is physically strong and
none of the joints were dull and cold in the end.
In retrospect, I probably should have used thicker-gauge wire for the
power and ground pins. Their pads are much larger, and I’m guessing the
LCD has a decent current draw. There’s a bit of “ringing” in the LCD
that I don’t remember being there when I was last using one of these,
but it’s possible that nostalgia has made my memory faulty.
30 more minutes of soldering onto the LCD panel, and it was finally
done. I slapped some Kapton tape onto the LCD-side joints for safety,
and then put the back case on and installed some batteries. Is it
working?
The TI-83 Plus screen reads: TI-83 Plus 1.12 / RAM Cleared
This screen is always shown on startup after a loss of battery-backup
power; the internal contents of the calculator’s RAM are scrambled and
must be reformatted on the first start. After this, it worked as
normal: I was able to write a BASIC program and do some simple math.
Annoyingly, unlike the TI89 Titanium I had been using previously but
put aside due to its voracious appetite for AAA batteries, I couldn’t
figure out how to easily do a base conversion with the built-in BIOS
ROM. I ended up writing a small BASIC program to do decimal to binary
and binary to decimal, but it’s annoying to switch between that and
calculator mode for things like bit flips and simple math, and I’ll
have to track down or make a serial cable to get a nicer one.
Let’s Talk About Ethics
Having opened the 83+, I now understood that it must have cost
virtually nothing to construct. There are only a couple inexpensive
ICs, and one of them is a proprietary TI gate array that can most
likely be fabricated cheaply. Over the course of several decades, they
must have been making a large integer multiple of the build cost in
pure margin while not updating the product.
This wouldn’t be so bad if not for TI’s pressure campaign to be the
mandated calculator for a lot of high school math classes and
standardized testing. Asking a bunch of rich suburbanite kids to spend
$150-180 Canadian on a TI-84 Plus is one thing, but I’m not sure I
could really sleep at night forcing a disadvantaged family to pay that
much for a machine – multiple days of work at even a $15 minimum wage –
that surely didn’t cost more than $20 to construct. TI’s effective
monopoly power hurts these kids as much as it has chilled the
development of better graphing calculators with superior feature sets
from other manufacturers.
While a lot of other devices featured on this blog are just toys or
luxury goods, this one is an essential tool for kids that is sold at
predatory markups.
Conclusion
Even though I think it entered into my life originally because of an
abuse of monopoly power, it’s a nice mini-nostalgia trip to have a
TI-83+ again. I think in the future it would be fun to write some more
Z80 programs for it, and maybe even do something ridiculous with it
like television output.
I wonder if anyone’s disassembled the ROM to add commands yet?
Repair Summary
Fault Remedy Caveats
Screen displays all-black, or garbage. Replace LCD ribbon cable
1. It [12]does. On top of the obvious decoding and refresh logic,
everything from keyboard matrix scanning to interrupt handling is
exposed to the Z80 through I/O ports. It seems like the LCD is
probably controlled by the glob top on the LCD board. [13]↩
[14]« Look, mom! X1's on TV!
Please enable JavaScript to view the [15]comments powered by Disqus.
Leaded Solder
* Leaded Solder
* [16]LeadedSolder
* Support with [17]Patreon
* [18]Subscribe with RSS
Old computer repair, history and programming.
References
1.
https://www.leadedsolder.com/feed.xml
2.
https://www.leadedsolder.com/
3.
https://www.leadedsolder.com/about/
4.
https://www.leadedsolder.com/challenges
5.
https://www.leadedsolder.com/tag/computer
6.
https://www.leadedsolder.com/tag/calculator
7.
https://www.leadedsolder.com/tag/texas-instruments
8.
https://www.leadedsolder.com/tag/ti83-plus
9.
https://www.leadedsolder.com/tag/lcd
10.
https://www.leadedsolder.com/tag/repair
11.
https://www.leadedsolder.com/2021/08/10/ti83-lcd-ribbon-repair.html#fn:gate-array-registers
12.
https://wikiti.brandonw.net/index.php?title=Category:83Plus:Ports:By_Address
13.
https://www.leadedsolder.com/2021/08/10/ti83-lcd-ribbon-repair.html#fnref:gate-array-registers
14.
https://www.leadedsolder.com/2021/07/13/sharp-x1-telopper-repair.html
15.
https://disqus.com/?ref_noscript
16.
https://www.twitter.com/LeadedSolder
17.
https://patreon.com/leadedsolder
18.
https://www.leadedsolder.com/feed.xml
