Calculator Counter

One of my current projects involves cycle counting for durability purpose. My original plan was to build a microcontorller circuit, using the ATMEGA328P, to do the counting and display. I have been building the circuit in my head and was putting together a parts list over the weekend when I realized that I may be putting a lot of time into designing this circuit without first testing my mechanical system. I needed a quick and easy circuit which got me thinking about using a cheap calculator. I’m sure I have seen this method used before, it is probably where I got the idea, but I though I would post something about it anyways. Let me explain how this works.

Most calculators (citation needed) have this feature where you can add the last number input to the current answer. For example, if I put 1 + 1 = in my phone calculator I get 2. Pressing = again gives me 3 then 4 then 5 ect. It is adding 1 to the answer every time I press =. Therefore if I can press the = button on a calculator at the end of a cycle it will then add that cycle to the count and act as a cycle counter. On to the cheap calculator.

I had a Sharp ELSI MATE EL-233S in my desk drawer that was unused. I had to purchase it for a economics class since scientific calculators where not allowed during exams, this was the cheapest and most basic calculator I could find at the time. Before opening it up I made sure it still functioned and would have the feature I was looking for. I found that the keystrokes required could be reduced by just pressing 1 + =. This is great for when I will need to initialize the calculator.

Opening up the calculator reveals a bare minimum circuit. There is the ASIC hidden behind its black goop, 3 capacitors and the LCD. I didn’t expect much. My plan was to solder wires to the key pad contact so that I could then control the key presses using a transistor as a switch. Initially Looking at the contacts I assumed it was copper tracks covered in a conductive black substance, graphite? This is not true. The entire keypad circuit is built using this black conductive material which can not be soldered to and there is no copper underneath, you can see in the pictures my attempts to scrape if off.

With that plan in shambles I looked at another method to hook up my transistors. Luckily on the reverse side there are copper tracks and the manufacturer was nice enough to leave large exposed test pads. Figuring that a combination of these 11 test pads creates the keystroke inputs, I started probing around. I found the following keystrokes.

This method of probing around was taking too long and had unexpected results. I then traced out the keypad contacts from the front to the test pad they were connected to.

Also in order to be able to use NPN transistors I will need to determine which pin is more positive so that it can be connected to the collector. This is easy to accomplish with a multimeter. Using DC voltage mode measure the voltage between the two pins. The following table displays the polarity between the keystroke connections.

The wired up circuit is below, please excuse the horrible pictures but this should give anyone a good start.

To initialize the calculator into “counting” mode an Arduino was programmed and wired up. A frequency generator is used to find the maximum counting frequency that the calculator can count. The below oscilloscope screen capture shows the maximum counting frequency of this setup. Because of the low counting frequency I abandoned using this method as a counter.