A recent addition to the desktop calculator collection is the Canon Canola L100A desktop 10 digit gas discharge display calculator. This unit was purchased as a working calculator, and was in good shape except the heavy yellowing of the beige upper case..

The unit arrived with it’s power cord wrapped separately, and was fairly well packed. Upon opening the package, it was obvious that the calculator came from a cigarette smoking environment. So the first order of business was to disassemble all the major components and thoroughly clean all the casework, keys and power cord with my favorite all-purpose citrus cleaner. The circuit boards were cleaned with 91% Isopropyl alcohol. There was quite a bit of flux residue on the boards from initial hand soldering of wires, and quite a few areas where the main board had been reworked to remove solder bridging and to touch up connections from the original wave soldering.

The L100A is comprised of three main modules consisting of the main board and display, keypad assembly and connector, and the AC power connector, switch, and transformer assembly.

My L100A was factory set for 115 volt 60 Hz AC input. The three wire input receptacle routes the AC line leg to the power switch and then to circuit board connection J1, which then connects to the fuse with the other end of the fuse connected to transformer wires K6 and K7. The neutral leg of the connector is tied directly to circuit board connector J2 which then connects to K2 and K3 transformer wires. Transformer wires K1 and K8 are soldered to the circuit board, but are on isolated pads. Pads K4 and K5 are linked together but have no wires attached.

The transformer primary windings in this configuration measure 126 Ohms. The two yellow leads on the transformer secondary are the high voltage supply for the display and have a resistance reading of 319.4 Ohms. The two grey transformer secondary leads are the low voltage supply and measure 11.46 Ohms. Both secondary’s attach to the main board thru a 4 pin connector.
Measured current drawn at 120 VAC input is 0.062 Amps with all digits displaying 8’s. My measurements show a maximum of 5.8 watts and an idle wattage of 4.9 watts after the display was cleared.

The keypad is fairly simple with two teal colored clear keys, a 0–9 with decimal point section, 4 basic math function keys, and an equals key. The keypad is attached to the main board with a 20 contact edge card connector with 13 wires. I ended up installing new cable lacing, as the old stuff was very dirty and starting to unravel. The keypad had some foam strips under the keys that had deteriorated, so it was removed and replaced with some adhesive backed felt strips

The keys use a magnet and reed switch that closes when the key is depressed. The circuit board uses a 2 X 11 matrix for connecting the switches to the main board. There are only 18 switches, so that allows for several spares. Transistor TR2 is controlled by pin 25 of the TMS1824 IC and alternates selection of the two columns in the keypad matrix (b1 & b2).

The AC power assembly connects to the main board via a connector with 4 populated pins. The high voltage secondary is connected to two diodes D2 and then is filtered by the C2 vertically mounted axial capacitor pair rated at 4.7 uF and 200 volts each. The low voltage secondary connects to 4 D1 diodes in a full wave bridge configuration, and then is filtered by C1 a 470 uF at 25 volt capacitor. The HV filtered output is 297 volts and the LV filtered output is 20.3 volts DC.
You may have noticed the different labeling of the components by type and not by individual component. So all 4.7 uF @ 200 V axial capacitors are labeled C2, both HV diodes are labeled D2 because they are the same, there are 11 each C8 capacitors. Only different values are given different designations. This threw me off a bit when I was putting together my list of replacement capacitors, with two capacitors labeled C2 and three of them labeled C3.
Both 200 volt C2 capacitors were showing signs of leaking, so I decided to replace all of the 50+ year old electrolytic capacitors. Most of the capacitors were Shoei branded, with one of the C6 capacitors being an original Nichicon unit.

Original capacitor and replacement list:
- C1 x 1 — 470uF @ 25V radial 7.50mm lead spacing replaced with
- 470uF @25V radial Nichicon UPX1E471MHD1TN
- C2 x 2 — 4.7uF @ 200V axial replaced with
- 4.7uF @250V axial Vishay 500D475M250DC2
- C3 x 1 — 33uF @ 6.3V axial replaced with
- 33uF @ 10V axial Nichicon TVX1A330MAD
- C5 x 2 — 1uF @ 100V axial replaced with
- 1uF @ 150V axial Nichicon TE1500-E3
- C6 x 3 — 4.7uF @ 25V axial replaced with
- 5uF @ 50V axial Nichicon TE1303-E3

The smaller axial capacitors are located under the display assembly, but are easy to replace with their polarity clearly labeled on the board.

The display module was made by the (Japanese Radio Corporation) JRC and is a Septanix 237A1 . The display is similar to a Burroughs Panaplex display, but is made up of individual digit modules in what looks to be a ceramic case with a glass top cemented to it. There are 12 digits on the display, but only 11 are used by the calculator. The left most digit is used for the sign and error indication. The unused digit is covered to prevent unintentional information from being displayed.
The display bezel had what used to be an open cell foam gasket that has since degraded into a gooey mess. I was able to clean the display with 99.9% Isopropyl alcohol which easily removed the residue, but it was still quite a mess with black gunk everywhere. I have some very thin felt that was used as a spacer between the glass and the bezel to replace the foam.

The L100A uses two (Large Scale Integration) LSI (Integrated Circuits) IC’s manufactured by Texas Instruments specifically for Canon Inc. The TMC1824 and TMC1825 are both (Dual-Inline Plastic) DIP cases with 40 pins each, using a 10µ PMOS (P-channel Metal–Oxide–Semiconductor) process. This chip-set was Canon’s fourth series of calculator chips manufactured by Texas Instruments.
The calculator also uses integrated display drivers including two KH6248 IC’s for the display anodes one KH6249 IC for driving the display segment cathodes.
The KH5305 timing module is used for clock generation and its output is distributed to both 40 pin IC’s.
Component date codes and the quality control stamped label, set the date of manufacture at June of 1972 for this unit with serial number 217253. The Nichicon capacitor had the oldest date code from the second week of 1972 making it 51 years old as of the writing of this blog post.

The bottom case shell is fitted with a shield that has a Gold Iridite Chromate conversion coating applied to it. The shield is grounded to the AC power cord ground thru the AC module assembly which is mounted above it with four screws. The main circuit board is also grounded via the two raised shield tabs and the display mounting frame screws.

The AC module installs on top of the shield along with the insulator sheet used to protect the main board from accidently shorting to the shield. There is also a small vertically attached insulator that protects the back edge of the main board from contacting the AC module frame. The green wire is attached to a tab at the front of the shield and is later attached to the metal bezel shield.

This calculator uses a now obsolete plug found on Canon and Monroe (Canon OEM) calculators. It is a three pin rectangular connector with the center pin connected to ground and calculator chassis shield. This calculator also has a small pop-out handle so that you can easily carry it around.

It took a while after ordering the capacitor replacements to receive them due to some inclement weather affecting the northern USA, so I decided to do some further research on how the circuit board was wired together. One of my favorite ways to decode double-sided circuit boards is to backlite the board and photograph it and then use my image editing program to enhance the front and back traces. I use an edge detect filter that works good on certain board types, and a negative filter that almost looks like an x‑ray image was taken of the board. This helps me to follow the traces across most areas of the boards, but is limited by areas with large IC’s and modules that block the light. I then have to resort to continuity testing of the traces in those areas.


Pin | TMS1824 | Pin | TMS1825 |
---|---|---|---|
7 | VDD ‑8.54 volts | 9 | CP clock |
8 | VSS common | 23 | VDD ‑8.54 volts |
9 | CP clock | 24 | VSS common |
10 | SCT0 | 30 | KC inverted |
25 | TR2 Collector | 33 | VDD ‑8.54 volts |
29 | KC inverted | 35 | SCT0 |
38 | TB3 | 36 | VGG ‑15.35 volts |
40 | VGG ‑15.35 volts | 40 | VSS common |
The CP clock is a 100 kHz signal with a negative 14.2 volt pulse that is 2.62 μs wide.

SCT0 is a 970 Hz square wave at minus 8.4 volts.

TB3 is a 25.1 kHz negative pulse at minus 8.4 volts (no screen capture).
Now that all the electrolytic capacitors had been replaced and the calculator tested, it was time to try another “RetroBright” procedure on the upper case.
This time I used 12% hydrogen peroxide solution for 26 hours. The first afternoon was somewhat shady and cool, so I used an aquarium heater in an outer water bath to keep the solution warm. The containers were brought in overnight with the heater set to 86 degrees F. The following day was bright and sunny which kept the solution vigorously bubbling throughout the day without the heater.
Overall I give the results 95% with a few areas a light shade darker than the rest. All of the yellowing was gone, and I think the results are an artifact of the injection molding process for the plastic.

The L100A calculator is a well built unit by Canon, but marketed as an entry level calculator. It doesn’t have percentage, square root, or memory functions but what it does do, it does very well. For an algebraic 10 digit four function basic calculator it is a rock solid unit.