It's been a busy week here at Cemetech, including breaking the news of the TI-84 Plus C Silver Edition calculator from TI, reaching Slashdot's front page with that story. We have accrued a lot of new members and new excitement, and I can begin to hint that more officially-linked things may be happening in the general vicinity of Cemetech in the future. We also announced Cemetech Contest #9, which has at least ten definite entrants covering all three categories.

To give you a bit of weekend entertainment, enjoy the following photos of my (successful) attempt to fix a TI-Nspire I got for a low price. On opening it up, I found that the flexible PCB connecting the logic board and the LCD board had torn where it attached to the logic PCB. I tried gently sanding off part of the flexible orange PCB to expose the conductor to solder it back on, to no avail. I resorted therefore to some of the finest-pitch soldering I have done to date, successfully soldering the many wires necessary to replace the flexible PCB. My previous work repairing TI-83+ LCDs was a cakewalk by comparison.

In other assorted news:
:: A Texas Instruments source who wishes to remain nameless has confirmed to Cemetech that the TI-84 Plus C Silver Edition will have 3.5MB user-accessible Flash and 21KB user-accessible RAM. Plan your programs accordingly!
:: The Cemetech staff are working to upgrade the Tools and Resources section with new pages helping beginner users and programmers get oriented with their calculators. Please help us collect internal links and topics for those pages.

Stay tuned for more exciting calculator developments coming soon!

That's some amazing soldering, Kerm. I didn't know it was humanly possible to solder that cable. Did you have to go thinner than 30AWG?

Nope, 30AWG did the trick, but it was a painstaking process to make sure everything was perfectly aligned and not touching any neighboring wires or pads. I forgot my Helping Hands with its magnifying glass elsewhere, which would have saved me a lot of eyestrain.

Like a boss, on all counts. Did you get any more WIP documentation of the soldering?

I'm afraid I did not, but it was mostly just trying and failing to give the flexible orange PCB, then ripping it off and soldering on the red wires one by one.

I'm surprised more people don't have things to say about the officially-confirmed TI-84+CSE memory specs.

21k RAM is kind of weak. :/ One would think they'd have been able to have made some adjustments and fixed that as a problem. :/

Aye, I think we have a lot of programmers who are disappointed by that figure. I guess it will push people to write Apps and/or use shells that automatically archive and unarchive programs (or give BASIC programs the tools to do that themselves). I know that I'm personally hoping for the latter to make Doors CS more popular.

That is good ... I want to help!

Great! We need two things from people to help with this project:
1) Document important calculator skills in clear, easy-to-read posts here on the Cemetech forum that we can link to
2) Help us sift through the forum and find existing topics that fulfill those criteria.

I can do that.... I think I can, must activate spell check first though... then I will!

Harvest Moon won't run on the 84-C.....sadness.

Nice ribbon repair! Is that thinner than the 84+ cable?

Many thanks! I think it's just slightly closer spacing than the TI-84 Plus, which I always assumed I could not solder. I guess I was wrong. Also, do you know the pinout of the keypad connector? Do you know how I can force it into Nspire mode without an Nspire keypad?

Kerm may I suggest getting those tweezers and that trace repair lead off DealExtreme would save you even more headache as its just a small soldering iron tap on each pad to add the wires


Very cool and nice job tho!

The one that the insulator melts off of from the heat of soldering? I very well may have to consider trying something like that. Can you point me to the item page again?

http://dx.com/p/titlelead-repair-0-1mm-soldering-enameled-wire-20m-132805

That would be it you better pick up a pair of fine pitch tweezers as well tho thin is a understatement on that stuff

see this picture for what I mean https://dl.dropbox.com/u/98196116/2012-11-18%2021.03.57.jpg

I can recommend these tweezers immensly if you dont have any they are really good so far. http://dx.com/p/19869

*All pin numbers are based upon the PCB's silkscreen inside the keypads.

Kerm, about the pins for keypad detection, I'm positive it's not 3-18, since they're the IO pins used to detect keys. I'm also fairly sure about 1 and 2 being just a vcc and gnd source.

If I had to place a bet, it would be on 26, since it's the only pin that wouldn't be used for an input key OR the link port wires OR the i2c touchpad that still goes HIGH on probing. My scope readings say it's held high, so I suspect it's an input with a pullup resistor like all the keys have. However, the rest (27-30) are also potential suspects, since it may be that the 84+ keypad chooses to pull one down and leave one floating, and the nspire swap them. Check to see if the nspire readings match up with the 84+ pad readings for those pins (in the link below) and we'll know for sure.

Unfortunately I no longer have my nspire (it has since become my sister's calculator) so I can't confirm any of it.


I find it interesting how they got keypad input. They had a matrix of inputs tied to outputs, such that they would test for all the inputs with each output separately, AFAICT. This is why in this post I reported square wave readings when buttons were pressed. When it was being polled it dropped to ground, but when it wasn't, pullup resistors brought it back up.

That's actually how nearly all keyboards work. They scan through the input rows or columns, checking the value of the output column or row for each input set individually high.

Edit: Thanks for the tip about pin 26; I'll do some experimentation and see if I can confirm your suspicion.

Cool. I've always wondered how matrix-based systems avoided cross-contamination.