Kosmo format multiples

That sounds like an issue with the converter and/or KiCad (which I don’t use, so cannot help you with that the details).

I did a quick test in EasyEDA; the first row is copper and mask on laminate, the next row is silkscreen and mask on copper, the last one copper and mask on laminate (so same color as mask on copper, but surrounded by bare laminate). EDIT: Added the examples to the original post instead.

1 Like

Note, by the way, that both white and black solder mask are pretty opaque, a copper graphic with white or black mask will probably be near invisible. Other colors, particularly yellow, have more transparency.


I was gonna say, how of the top left would you really see? Would be like a shadow even on yellow

1 Like

I mean like this:

It’s a subtle effect, but I like it quite a lot!



Switch mult—
8X 3mm LEDs, 2 colors of your choice, 8 of each

1X 2x5 IDC Euro power connector

8X Kosmo style (5pin switched stereo) jacks

2X 1x6 female 2.54 pinheader male

2X 1x6 female 2.54 pinheader female

8X switches DPDT on/on or on/off/on
(On/on forces one channel or the other, on/off/on you can mute/disconnect a channel without unplugging)

8X resistors for LEDs. Mileage may vary depending on specific LED and preference. I chose 6.8k for a decent brightness (still pretty bright) but maybe check one switch setup before continuing with the rest.

Build notes—one of the resistors is not silkscreened with value. Use the same as the other resistors.


Buffered Mult BOM:

16x Kosmo-style (5pin stereo) jacks

2x12mm m3 stand-off’s (optional spacer support)

6x trim pots, 2k preferred, 10k works just a little sensitive when you adjust each output

2x 1N5817 schottkey @power supply or just use jumper cable or 10R resistor if you don’t have.

2x male/female 2.54 pinheaders, each 1x10

1x eurorack style power 2x5IDC socket

2x TL074

2x 22uF polarized caps

6x 10pf ceramic caps

2x 6pf ceramic caps

4x 0.1uF ceramic cap


6x 12k
6x 300R
12x 51k
4x 100k

Build notes:

Each of the buffered outs has its own trim pot to match to each other. Use a known voltage source and plug it into the first input. Adjust the trimmers until the 3 outputs match the input voltage. Repeat for input 2.

Build as expected. all components silkscreens are correct. However, the trim pot silkscreens are mislabeled. Pretty easy to figure out once it’s built and you are adjusting the channel voltages.

1 Like

Noob question, if anyone’s got the time - what’s the need for some many parts in these buffered mult circuits? In particular, the resistor/cap combo in the feedback look from the output to the negative pin of the op amp? I’ve seen mult schematics/stripboard layouts that have no components on the input and the feedback loop, and only have a resistor on the output, which I understand is for impedance

is a voltage follower. V_out = V_in. Unity gain by design.

is an inverting amplifier. V_out = -(R2/R1)*V_in. Unity gain if R2 == R1. Well, negative unity. Greater parts count but if one of the resistors is or is in series with a trimmer, the output can be adjusted.

You want unity gain and in principle the voltage follower delivers it. But if you don’t want “in principle”, you want controls that guarantee it, the inverting amplifier gives you that. Or if you need slightly non unity gain, which I do for a purpose not worth getting into here. Also the inverting amp keeps its inputs close to zero volts (the inverting input is what’s known as a virtual ground), well away from the op amp’s rails, which is a virtue. It does invert the signal so you need to use two in series.

Something I just realized (I’m still learning this stuff myself) is the inverting amp has input impedance R1 while the voltage follower’s input impedance is that of the op amp which is up above 1M somewhere. That can have implications if you’re trying to use it with a 1V/oct CV.

Edit to add: The AJH Synth V-Scale is a Eurorack buffer that not only has trimmers, it has them on the front panel! So it’s easy to use it to compensate for 1V/oct tracking problems.


Legend! Thanks for that, helps a lot. I’ve seen reference to inverting amplifiers/voltage followers before, so this clears a lot up. Sounds like the voltage follower set up is more what I’m after at the mo’ so will get on that!

Here’s schematic and build document for a voltage follower buffer

but basically it’s just jack, voltage follower, several parallel output jacks with series resistors. Seems to me it would be better to have a voltage follower on each output.


Nice. Yeah I agree that surely you’d need a voltage follower for each output? That’s what I was aiming to do at least.

Here’s a design with stripboard layout etc.


aha!!! I just saw this! sorry I have been rather off grid this past week! funnily enough I just put up a buffered mult too! but these look awesome!!


Nice, looks like voltage followers with 100R output resistors and 10M to ground on the input?

(new LMNC pcb? brb ordering one just as an excuse for ordering more 90’s toy pins.)


yeah thats what I ended up doing, the 10m to ground stopped it from hanging, is that a common thing?? however the voltage followers do exhibit some very interesting other uses! especially left floating, is there any thing out there documentation wise about using voltage followers and just floating pads? you can get a good chunk of max out of it from just touching it, figured a good means with a bit of slew to make some simple touch pads.


really cool that you merged passive and buffered on the same panel! genius


aha haha yeah cant call it the f words :smiley:


The input impedance of the non inverting opamp can be Giga Ohms. The datasheet is your friend.

1 Like

Not beyond “don’t do that” :smiley: – See any article about how to wire up unused opamps, e.g. here:

Improperly terminating unused op amp pins has the potential to lead to degraded the op amp or system performance. Leaving either the input or output pins floating will create unexpected voltage shifts, which leads to unpredictable behavior.

The JFET gate has extremely high impedance, so there’s not really anywhere for any electron that makes it in there to go next, so it doesn’t take much to mess with the JFET.

Hanging, as in locking up or something else?

1 Like