Construction of the Funky −
a key using pressure sensors
Some surplus electronics companies are offering relatively inexpensive pressure sensors.
The sensors I used are buyable from Conrad Electronic (FSR-151, order no. 182546).
Actually I wanted to see what I could do with one many times.
Amazingly the sensors "played" well, so I started immediately to look for a project.
I had long wanted a durable and light key.
Therefore I decided to make a key using the sensors.
The pressure sensors need an interface for the connection to a keyer. First I had a solution with an operation amplifier (CA3098) in my mind. I discarded this idea, however, as there was a simpler way. Sensors of this type have a resistance of over 1 MΩ in the non-actuated state according to the data sheet. With my sensors it was measured even higher, as I could not measure anything within the highest range (10 MΩ) of my Ohmmeter. If one presses carefully on it, the resistance changes rapidly down to a value of approximately 100 kΩ. Therefore the interface had to detect only this threshold. Thus I searched again in my junk box and found a CMOS trigger (CD4093). This circuit family uses almost no current and can be used with a voltage range of 3 to 15 V.
The interface circuit is shown in the Figure (Key interface circuit; the outputs connect to the inputs of a keyer). The sensing of the trigger level is done via two trigger inputs of the CD4093. Two 10 MΩ resistors with hold the voltages at the inputs with non-actuated sensors to a fixed value below the trigger level. Thus the circuit is stable and it does not break into unwanted oscillations. If one operates a sensor, then its resistance is greatly reduced and the potential at the trigger input shifts over the trigger level toward the plus potential. The trigger output thereby switches abruptly.
According to the data sheet, the resistance of the sensor drops very rapidly with applied pressure. A force of 20 g drops it to 2 MΩ (that is, for instance, the weight of two normal writing papers). The more strongly one presses on the sensor, the further the resistance drops. If you want to make the key somewhat more insensitive, then you need only reduce the resistance at the input. If this is reduced to, say, a value of 100 kΩ, then the sensor must be strongly pressed before the trigger is switched.
Additional transistors on the output to drive a keyer are not necessary, since the internal output transistors of the circuit can sink up to 20 mA.
"Wobble" or "Squeeze" Key
With a mechanical wobble key you can produce either dots or dashes, but not both simultaneously. This function should also be in this key. Therefore I inserted a "barrier" which prevents the simultaneous output of both characters. That is the purpose of the RS flip flop at the input of the circuit. For fans of the squeeze (iambic) technique, the flip-flops can be disabled by a DPDT switch or jumpers as shown.
Oops, I almost forgot a very important point! The CMOS circuit needs very little current, but it can't work without power. My first the idea was to build the key into a transceiver so that its power supply could be used. However, then the key and transceiver are a unit and I could not use the key with a different transceiver. The second idea was to use a small 9 V battery to power the interface circuit. Even though the circuit uses very little current one could forget to disconnect (or turn off) the battery. Murphy would undoubtedly show up and kill the battery at a most inopportune time!
Thus I looked for and found another option. The key its powered (the current is in the microampere range) over the normal key cable, making use of the voltage normally present on the keyer input lines. Normally no keyer modifications are necessary. How does it work? Quite simply. It is possible to attach a resistance at the keyer input instead of the key without a keystroke being detected. With the K2 transceiver, for example, the "critical" resistance value is approximately 22 kΩ. This results in a current up to 220 µA at the "open circuit" voltage of 5 V on the key line. The current actually necessary is much smaller and is not measurable with my meter. A 1 µF capacitor stores sufficient energy during the output of a character to power the CD4093. The diodes prevent an inadvertent discharging of the capacitor due to the "short-circuit" on the appropriate line to the keyer. Since in most cases the key lines alternate, voltage is supplied even when one line is low. Thus the capacitor stays continuously charged. I have tested the "holding time" of the capacitor when both lines are held low and it is greater than two minutes − more than enough for even the slowest QRS station!
The sensors can be attached to any smooth and flat surface. However, no aggressive adhesive should be used since the plastic of the sensor could be destroyed. I used double-sided tape to attach the two sensors to an aluminum paddle. If you don't like the flat surface of the sensors when operating with the key, you can apply a small self-adhesive device foots (idea of Dieter, DL2LE) as buttons on the sensors. The buttons should not be larger than the active surface of the sensor (the pattern in the center); otherwise the key will not function correctly.
I built the circuit on a universal PCB without copper foil using bare wire. The building of chassis is probably the most unpleasant part of homebrew construction. Therefore I would like to make a suggestion for a simple chassis for you. I made a mini chassis from the sawed-off part of an old chassis with guide rails. I do not indicate the exact mass, since that is strongly dependent on the material used. Since the key is quite light, it should to be fastened to a heavy piece of sheet metal. Naturally you can also use the key on a leg strap or like the Palm Paddle with a magnet holding it to a desk or to the transceiver. You can also hold the key with one hand with one hand and operate it with the other.
PS. Many thanks to Larry East, W1HUE, who had edited my in the QRP Quarterly published article to "Americanize" my bad English.
PPS. In the meantime Chuck Olson (WB9KZY) has undertake some futher experiments with this key and ascertain, that the key works better without the pressure sensors. His variation is named "Touch Paddle" and he sells it via Jackson Harbor Press (Photo by WB9KZY).