TI-45: Rejuvenation of a faded VFD-Display

[Pjwum]

Nice Sunday wishes,

I thought I'd share some calculator rework on a rainy Sunday.





The sample on the left suffers from a considerable dim display. As far as I know, the TI-45 is the only TI pocket calc featuring a vacuum fluorescent display or VFD.

It basically works like a valve tube. A filament wire, the cathode, is heated by a low current and starts to emit electrons. The electrons accelerate to the anode with a potential difference of about 30 Volts, they hit the phosphorized anode coating and the kinetic energy of 30 eV is mostly transformed into photons. The segments of each number are separate anodes and can be switched on and off individually. A grid between cathode and anodes for each digit can be either positive (on state) or negative (off state). This way the driver electronics multiplex digits.







The emissive layer on a cathode filament consists of metal oxides which can degrade over time by disuse. But it is possible to bring it back to life by injecting high current pulses. Usually the electronics drive the filament at about 1.5 V AC. Here 6 V DC are directly connected to the filament pins of the display package and the filament wire starts to glow. This treatment is applied about 20 seconds and two times with 8 V. The result is clearly visible.





One advantage of VFDs over LEDs is very low current consumption. Only the filament draws some milliamps. Anodes and grid have extremely high impedances and do not contribute to the current consumption. Hence, total current draw does not depend on the number of lit digits.
The pictures below show current draw in off-state (28 uA), on-state one digit (15 mA) and on-state eight digits (16 mA).

[Steve Simpkin]

Very interesting! Somewhat similar to the principle of a TV picture tube rejuvenator.

There were a handful of TI handheld calculator models that used a VFD. I know at least the TI-1025, TI-1265 and TI-1650 four-function models used a VFD. I was surprised to learn that the power consumption was less than an LED model. I would have thought that the high voltage power supply needed to drive the anodes alone would consume more power than an LED model. Thank you providing the data to confirm the lower power used by this model. For many years VFDs were also cheaper than LEDs as well.

The above TI models appear to use a capacitor voltage multiplier instead of a transformer to achieve the required anode voltage.

http://www.datamath.org/BASIC/TI-1000/TI-1025.htm

[Steve Simpkin]

Here is a photo of my TI-1025. It came with its box and looks like it was hardy ever used. I can't tell if the VFD is dimmer than when it was new but it is very readable.

[attachment=9547]

[toml_12953]

(06-06-2021 05:56 PM)Steve Simpkin Wrote:  Very interesting! Somewhat similar to the principle of a TV picture tube rejuvenator.

There were a handful of TI handheld calculator models that used a VFD. I know at least the TI-1025, TI-1265 and TI-1650 four-function models used a VFD. I was surprised to learn that the power consumption was less than an LED model. I would have thought that the high voltage power supply needed to drive the anodes alone would consume more power than an LED model. Thank you providing the data to confirm the lower power used by this model. For many years VFDs were also cheaper than LEDs as well.

The above TI models appear to use a capacitor voltage multiplier instead of a transformer to achieve the required anode voltage.

There was even a TI with a Panaplex display (TI-150)!

http://www.datamath.org/BASIC/DATAMATH/J...tm#display

[Massimo Gnerucci]

(06-06-2021 07:45 PM)toml_12953 Wrote:  

(06-06-2021 05:56 PM)Steve Simpkin Wrote:  Very interesting! Somewhat similar to the principle of a TV picture tube rejuvenator.

There were a handful of TI handheld calculator models that used a VFD. I know at least the TI-1025, TI-1265 and TI-1650 four-function models used a VFD. I was surprised to learn that the power consumption was less than an LED model. I would have thought that the high voltage power supply needed to drive the anodes alone would consume more power than an LED model. Thank you providing the data to confirm the lower power used by this model. For many years VFDs were also cheaper than LEDs as well.

The above TI models appear to use a capacitor voltage multiplier instead of a transformer to achieve the required anode voltage.

There was even a TI with a Panaplex display (TI-150)!

http://www.datamath.org/BASIC/DATAMATH/J...tm#display

Not to mention my preferred one, the SR-22!

[Steve Simpkin]

Or the TI Minimath early LCD model from 1972 that never made it past the prototype stage.
http://www.datamath.org/BASIC/DATAMATH/Minimath.htm

[Pjwum]

:-)))) I have to admit the calculator era for me starts with the TI-30.

(06-06-2021 05:56 PM)Steve Simpkin Wrote:  The above TI models appear to use a capacitor voltage multiplier instead of a transformer to achieve the required anode voltage.

The TI-45 uses the exact same 4 ceramic capacitors plus 4 diodes to boost the 30 V anode voltage. No other components beside the TMC chip.

[dhe]

Any chance of getting a picture of the back with the power supply connected?

[Pjwum]

(06-07-2021 07:06 PM)dhe Wrote:  Any chance of getting a picture of the back with the power supply connected?

Here you go. The filament usually is connected to the outermost two terminals of a VFD and can easily be verified by measuring the resistance (should be < 100 Ohm). Since it is only a wire polarity is irrelevant.

Patrick

   
   

[voltaage]

Thank you Pjwum for the post!

I tried to do this on one Elektronika MK 61 of mine which had a very, very dim, unevenly lighted display. In my case, the voltage through the filament had to be between 12V-15V to make a difference. Initially I went for 8V but the filament was not glowing at all, even with a couple of cycles I couldn't get the digits any brighter so I opted to raise the voltage.

This calculator was such a pleasure to work on since each end of the filament is directed through a daugther-board that does all the voltage magic. I had to cut two of these wires that the small board used to hook up to the main pcb to make sure the voltage was applyed solely to the VFD, so no other component could blow up. When monitoring the improvements I just re-soldered the wires each time to power up the calculator. After I was satisfied with the results, those tinned wires had to come out, nothing that some donor resitor legs can't help with.

The number of times I've seen a worn VFD is very reduced. My local supermarket has one IBM-branded POS information pole (I'm not sure how you'd call these things) that dates back from the early 2000's. That's thing has been on literally, 24/7 for the last 22 years and no fading, no unevenly lighted portions are present.

My MK 61 has a very unique display. Most of the models I find online only have the -88888888°88 layout, whilst mine is arranged as 888888888888. Any dedicated digit for the sign display is just a simple middle segment for your regular "8". Probably the earliest ИЛЦ2-12.8П tubes were like this before realizing that is was better to leave just the bare basics. While true, under normal circumstances using the -88888888°88 layout is fine, I found some tricks to trigger the use of those segments that shouldn't.

Therefore, I had no choice but to work on the old part, which gave great results.