Questions and Answers: Keyall Discrete kit 1/27/2019, Chuck Olson, WB9KZY q: Why so many LEDs ? a: The MOSFETs used generally are specified to turn on fully with a gate to source voltage of 10 volts. With 6 red LEDs (illuminated by 6 other red LEDs) the Keyall Discrete will deliver roughly 9.5 volts at an LED current of 20 mA. While it would be nice to exceed 10 volts, 9.5 is probably enough since another LED pair would make it harder to drive with 13.8 volts. q: Why use red LEDs ? a: White, blue and green LEDs were tried but none worked as well as the red LEDs. White LEDs use a fluorecent coating which apparently prevents them from generating any appreciable voltage when illuminated. Green and blue LEDs do generate voltages which are actually greater than red LEDs but they don't have the "oomph" (a new technical term ;) of the red LEDs. One figure of merit for photodiodes is short circuit current, I think of it as reflecting the internal resistance of the LEDs when they are operated as photodiodes (solar cells). Blue/green LEDs had a short circuit current of a few uA. The red LEDs used in the Keyall Discrete have a short circuit current of 40 uA or more. q: Isn't a lot of light escaping through the ends of the LEDs and through the soda straw ? a: Yes, I did try using aluminum foil wrapped around the LEDs and it actually did raise the short circuit current a little but I though it would compromise the isolation between the input and output by providing a conductive path. q: what about LEDs with flat ends ? a: There aren't as many LEDs available with flat ends as there are with the rounded lens - some were tried and results were poor, they just aren't as bright as the ones with lenses used in the kit. The flat ends are much easier to work with though and result in a smaller layout. q: what about surface mount LEDs ? a: that will be the next step, will try to rig up a prototype with SMT LEDs and see how they perform. The commercial integrated PVI (Photo Voltaic Isolator) chips that are available as well as the SSR (Solid State Relays) chips use an array of something like 20 or more IR photodiodes which are illuminated by a single LED. That's the real advantage that the SMT LEDs might well share if grouped close together. q: the board is labeled with OUT1 and OUT2, does that mean two loads can be switched ? a: generally no, the Keyall Discrete is a Single Pole Single Throw Normally Off relay. With a mechanical relay it generally doesn't matter which end is connected to what part of the load. And that is true of the Keyall Discrete, OUT1 and OUT2 are both connected to the load (with a key this would generally be the transmitter ground and the tip of the key plug which could be a positive or negative voltage). However the Keyall Discrete can also switch AC just as a mechanical relay can. There is an exception mentioned in the manual where the common source connected (labeled output ground on the schematic) can be used with OUT1 and OUT2 to switch two POSITIVE voltage loads such as a VFO and a cathode keyed transmitter, but this is pretty rare. q: are there any disadvantages to the Solid State Relay design ? a: yes, since the two output transistors are used in series there is a doubling of the voltage drop across the Keyall Discrete. Also the high voltage transistors used (they are the cheap ones) have a relatively high on-resistance, so this drop can be considerable, depending on the current being switched. However with normal keying circuits on relatively low power rigs this isn't a problem. People do ask about using the Keyall design for switching higher frequency AC voltages, I've never tried it but I suspect the voltage drop mentioned above would make this a bad idea. That's one advantage of a mechanical relay. q: are there any advantages to the Solid State Relay design ? a: yes, the SSR generally will NOT bounce as mechanical contacts will. q: what about those relatively cheap solid state hockey puck type relays, why not use them instead of the Keyall ? a: because inexpensive SSRs usually use Triacs to perform the switching which means AC voltages only. An SCR or Triac will turn on a DC voltage but unless some involved circuitry is used to perform "commutation" the triac and thus the relay never turns off. There are commercial MOSFET solid state relays but they are going to be more expensive than the Triac designs of the same specification. q: what does the small MOSFET do ? a: the TO-92 MOSFET is an LND150, a depletion mode MOSFET. The reader can read more on the technical workings of depletion versus enhancement mode MOSFETs but the brief explanation is that when the gate to source voltage is zero a depletion mode MOSFET is ON the opposite of enhancement mode MOSFETs. If the gate to source voltage is made negative the depletion mode MOSFET will begin to turn OFF. In the Keyall Discrete a blue LED pair is used to generate a negative voltage of over 2 volts (larger than a red LED). So when the key goes down the LND150 will turn OFF allowing the full voltage from the 6 red LEDs to appear on the gates of the two output MOSFETs. When the key goes up the 10 megohm resistor will discharge the gate and turn ON the LND150. This will then discharge the gate capacitance of the two output MOSFETs. Without the LND150 the output MOSFETs will stay ON for seconds after the key goes up. This would certainly only allow Morse code to be sent at glacial speeds. One other thing the LND150 does is to prevent any incident light on the red LEDs from turning on the output MOSFETs while the key is up. q: but why not use a simpler discharge circuit for the output MOSFETs than the blue LED pair ? a: some PVI chips do use simpler circuits like a single discharge resistor or an JFET or MOSFET with a series resistor. But the problem with these circuits is that these resistances affect both the maximum output voltage of the LEDs and the short circuit current (figure of merit). The LND150 plus blue LED pair combo assure a high output voltage and won't decrease the short circuit current. q: will a mini 12V battery like the A23 work with the Keyall Discrete as a power source ? a: yes, it will work, the question is how long ? The A23 will have an open circuit voltage of 12.5 volts when fresh but with a 17 mA load, the voltage would decay to the 11 volt level in a guestimated 40 minutes of key down time. The battery is tiny though (used for garage door remotes and other portable applications). The A23 is available for less than one dollar at Digi-key with the battery holder also about that price. Another idea I'm pursuing is to use two 4LR44 batteries (6 volts each) along with a AA cell for a total of about 14 volts. The 4LR44 battery is supposed to be about the size of 1/2 of an AA cell so the idea is to use a two AA cell battery holder for all three q: what is the size of the board and the mounting hole co-ordinates ? a: the Keyall Discrete board is 1.95 x 1.95 inches or a little less than 50 x 50 mm. The mounting holes are .1" in from each edge, they are sized for 4-40 hardware (or M3). Don't try to use the TO-220 holes to mount the board, the tabs of the MOSFETs are "live" (connected to the drain). q: When the web page mentions 1000 Volts or 500 Volts, is that the rating of the Keyall Discrete ? a: No, it's the rating of the high voltage components, the two large (TO-220) enhancement mode MOSFETs and the .01 uF capacitor. Personally I'd take a cautious approach and derate them by at least half. The two output MOSFETs are connected in series so ideally they would equally share any voltage across them but that seems like an unproven assumption. Also, even if the applied steady state voltage is less than 1/2 of the component rating beware of any series inductance of the load - at turn off a large, reverse polarity spike can be generated by the collapsing magnetic field of the inductor - some kind of protective device (usually a reverse bias diode) should be connected to protect the Keyall Discrete.