Hey y’all,
I know I only recently posted about my current other work that I am still awaiting part deliveries for, but I’m too excited about this new design not to share and ask for some feedback. Would love to hear your opinion!
This is a triple hard keyer based on the LZX Cadet VIII Hard Key Generator. It has three key and three CV inputs which normalize down, respectively. It has three respective positive key outputs and negative key outputs and attenuversion and offset for the CV.
Since I chose to expose the negative key outputs (compared to LZX Keychain which only has single output for each key), I expanded the panel to 10hp. With the added panel space, I added some switches to swap each negative output for a particular logic combination (thanks Lars for the idea!).
The first negative output will be able to be switched to be the sum of all 3 positive outs. The second negative output will be able to be switched to be the AND result of the 1st and 2nd positive outputs. The third negative output will be able to be switched to be the AND result of the 1st and 3rd positive outputs.
I hope that explains it without confusing people too much. Anyways, here are the current schematics - broken out into control board, main board, and power board (shoutout Syntonie for the design strategy). Would really appreciate any and all feedback. Thanks!
Control Board
Main Board
Power Board
11 Likes
This looks great! I kind of went down to a similar path with Seuils: comparators + switchable logic operations between the comparators 
First thing I notice is that you have the output of the AND gates connected directly to jack output, however logic chips usually cannot drive outputs directly so you’d need to move your output buffers (U5A, U6A, U11A) after the switches and before the 499R output series resistor, this way both the comparator and AND gate outputs will be properly buffered.
Then by doing so, since both the comparator output and AND gate output will share the same voltage divider, and comparators usually doesn’t go all the way up to the voltage rail, whereas the AND gate will (producing a 0V-5V signal), the output amplitude will be a bit higher when the AND is selected. From LM361 datasheet, it seems like the logic high (1) voltage is ranging from 2.4V to 3.3V, and the logic low (0) can have an offset up to 0.4V, hence why the voltage divider divides around 3.3, and there is a tiny bit of offset added at the negative terminal of the op-amp.
Then the extra amplitude isn’t too much of an issue since we’re dealing with logic signals, so even if it goes up to 1.5V when the AND is selected, the extra voltage will be ignored at the encoder stage, and even when processing it or using it as CV prior to the encoder, it shouldn’t cause any issues.
Then I guess a simple way to correct this would be to add another series resistor between the AND gate output and the switch, so it will be in series with the 2.32k, and with a value of 1.5k, it would give about 1V at the output buffer when the AND gate is selected.
I had to deal with a similar issue with Seuils, though ended up using a logic buffer at the comparator output to bring it from 3V-ish to 5V, so I could then use the same 5V->1V divider. The first revision would output a signal less than 1V when the comparator output was selected, resulting in a slightly darker output, whereas the output of the AND/OR/XOR were going up to 1V.
Keep up the good work!
4 Likes
Thanks so much Bastien! I am thinking of redoing the whole logic portion of the circuit to perhaps add some other operations besides AND, so I’ll definitely be incorporating your advice on using logic buffer so I can have consistent voltage to resistor divide. Really appreciate your thorough advice as always 
1 Like
I decided to take Bastien’s advice of moving the op-amp buffers to after the switches and adding 1.5k resistors in series with the logic outputs. I also scrapped the sum and AND outputs and replaced them with 1A XOR 2A, 1A XOR 3A, and 1A XNOR 2A. Went ahead and did an initial layout/assembly. Next steps, routing and front panel design.
Updated Control Board
Updated Main Board
3 Likes
Very cool, have you thought about low impedance 75 ohm outputs, for even harder / sharper keys in HD? Technically I think that’s supposed to be combined with 0.5 W resisters on the inputs.
But I’m not an electrical engineer and am not competent to give advice.
EDIT: I was off on a bunch of things. For starters, I have larger 2W 75Rs, not 1/2W. This is good, though, because that’s the recommended size to use in this case. Since the LM6172s can drive up to 12V, whereas the ADA485x only 5V, they need a higher power rating than the 1/2W ones used in the Gen3 modules. I’ll be able to jam the 2W ones in this go around, but I can expand the footprint for the next rev.
I honestly didn’t know they had been confirmed suitable for LM6172 output buffers, but I now know they are. I have some larger 1/2W 75Rs I can jam in there to test this time around. Fortunately, it looks like there is a 1/2W available in the same small packaging as the 1/4W counterpart - so should just be an easy swap!
Because I used LM6172s, I didn’t include any input clamping circuitry - so it seems with the 75R output resistors there is a risk of blowing them in very niche cases. Lars has suggested something in the 1-200 ohm range as a compromise.
1 Like
Available for purchase now on my site
5 Likes
