Resonant LPG - My first proper KiCAD Project

KOSMO MODULES
1

Hi all, after having a chat with CTorp the other day about his aweome k25 utility modules, I decided to have another crack at using KiCAD to build myself a proper module of my own specification (I’d say design, but so many other people have built this module over the years I really can’t make that claim) - A Bucla style low pass gate with resonance and some attenuverted CV control.

As this is my first attempt at doing anything properly in KiCAD I figured before I move forwards to the stage where I am assigning footprints and laying out the PCB, I would first see if any of you lovely folks could give me any feedback on my schematic itself.

EDIT: This updated schematic has now been tested and shown working by CTorp in a post below!

Known Issues: Low level of resonance at audio out in VCF mode due to conversion from 15v to 12v - Anyone got any ideas how I can fix this?

LPG Schematic
LPG Schematic2204×1559 221 KB

I have chosen to use dual VTL5C3’s as per the original Buchla 292 schematic rather than the VTL5C3/2 of the later versions because they were easier for me to source. Similarly I have listed one of the OpAmps as a TL084 just because the original schematics say to, and I happen to have some lying around. For the CV summing stage and the attenuverter I have used a TL074 instead though because I have loads of them. I also have a floating OpAmp left over from the TL072 that I have just closed off the connections on, but if I can get this module to work I would like to make a Quad version with some interesting normalisation between the four channels and a summing mixer stage at the end where I will be able to put them to work. There is also a 47k resistor next the ‘The Deep’ switch (R17) which I have seen it suggested to replace with a multi turn trim pot so you can tweak it behind the panel and get it to sound better, but I am not sure how much use it’ll be. I intend to try it both ways.

I will of course be breadboarding this before I design/send out for any PCB’s, but I need to wait for my vactrols to arrive from Thonk before I can do this, so I figured getting some feedback may help in not blowing anything up when they get here.

Have I made any glaring mistakes? When I run the electronic rules test I get errors on my three power lines (shown in the bottom left box marked shockingly ‘Power’) because I am not driving them. Since the power is coming from the euro header, can I ignore this?

Also, has anyone else put one of these together before - I know Eddy Bergman has a stripboard layout for a similar design (he basically just handles his CV In’s differently) - and know of any componant swaps to make it better on 12v, as whilst it is confirmed to work at +/-12V it is designed to be a +/-15V module.

Thanks for reading, any input you might have would be most appreciated!

4 Likes
2

The pots look like they’re backwards to me. Terminal 2 shorts to terminal 1 when fully counterclockwise, so for instance turning RV1 “up” will make the offset go from the full CV value to none, and I think that’s the opposite of what you want. I get confused about this myself a lot, so think each one through carefully.

U1D should not be left disconnected. Connect pin 12 to ground and pin 14 to pin 13.

TL074s and TL084s are pretty much indistinguishable nowadays.

I believe knowledgeable people when they say ferrite beads on the power rails are generally useless in audio circuits. I put 1N5817 Schottkys there instead to protect against power reversal.

Best practice is to add 100 nF bypass capacitors close to each IC, one from each power pin to ground.

The KiCad complaint about undriven power pins can be addressed, if you feel like doing so, by adding power flags (from the same menu that has ±12 V and Ground symbols) to the rails. Or you can just ignore those complaints.

J3 is connected incorrectly. Should be like this.

image

(Note the use of a symbol where the numbering is odd numbers one side, even numbers the other; that’s the standard way. Pins 1 and 2 to -12 V; pins 9 and 10 to +12 V; all remaining pins to ground.)

5 Likes
3

By the way, maybe this is a Buchla-vs-Moog terminology thing, but is “Gate In” actually a CV (such as an envelope generator or slew output)? To me a gate is a binary thing, either 0 V for off or (say) 5 V for on, and one wouldn’t ordinarily add an offset to a gate.

1 Like
4

Thanks for the help! So like this…

SCHEMATIC REMOVED - SEE FIRST POST FOR LATEST REVISION

I’ll try replacing the ferrites with some diodes though, since reverse power protection sounds like a good idea :slight_smile: I just can’t find the right symbol in the component library for Schottkys (assuming they are different from regular diodes)

The original 15v schematics have a 150k resistor at R11, by the Offset pot, is this something I should change with it being powered from a lower voltage supply? It strikes me as something with a 1:10 ratio to the supply voltage, but I could very easilly be wrong.

Also, on the subject of the ‘Gate In’ jack, I realise that the gate is basically just an unattenuated CV for the Offset, but its supposed to be connected to a sequencer/keyboard gate out, fully opening and closing the Offset. Its labeled that way here because thats how I was going to label it on the panel. I think Make Noise described it as a ‘Strike’ input on their LPG’s and Buchla had them labelled as ‘Control In’.

1 Like
5

Oooh exciting!

I just put 10r and fully accept people to follow whatever religion they follow; jumper, 10r, ferrite bead, 5817, etc.

Definitely breadboard it! I’m sure you can find a use for the unused op amp (led driver?)

@analogoutput nailed it tho with all the suggestions. That 2x5 odd_even power footprint is a doozy and got me a couple times.

2 Likes
6

Schottky symbol is D_schottky

2 Likes
7

In my KiCad installation, 1N5817 comes up when you type that into the search bar in the Choose Symbol dialog. That’s the name of the symbol, under Diodes. It’s an alias of SB120, so you could use that if you don’t have 1N5817.

For the J3 symbol I advise using Conn_02x05_Odd_Even or something with the equivalent pin numbering. If you want a shrouded header the footprint for that is IDC-Header_2x05_P2.54mm_Vertical under Connector_IDC. It has the key slot located such that -12 V needs to connect to pins 1 and 2.

Regarding R11 etc.: I don’t have the original schematic, but I guess the idea is that the U2A circuit is acting as a constant current source for the vactrols. In the original I suppose the input voltages at R11 and R14 both could vary from 0 to 15 V whereas here they can go from 0 to 12 V, so the vactrol current range will be correspondingly reduced unless R22 is likewise reduced. Though the feedback’s complicated and I don’t fully understand it. [Edit: No wait, that last bit’s probably wrong. Maybe R22 needs to change but I’m not sure of the analysis.] Also I don’t know whether the VTL5C3 needs a different current range than the VTL5C3/2.

3 Likes
8

You know, I was just thinking it might need at least one blinkenlight… :smiley:

Once I get this single channel version running I want to put 4 of them behind one panel, with the outer (on the panel) channels output normalled to the inner channels input tip switch pole - taking the signal before the jacks at the output opamp - then all the output jacks switch poles going to a summing mixer. That way if you leave an inner channel input unconnected the outer channel would go to it be default, and by plugging a cable into an individual channel output would remove it from the sum. My original thought was to use the spare op amps for the mixer, but I guess it would do no harm to add another TL072 for that and have the spare opamp on each channel power an offset status LED.

1 Like
9

Awesome, found it, I must have mistyped the component name first time round. I want the arrow pointing away from the 10 pin connector for the positive rail and towards it for the negative rail, yes?

I have also updated the connector correctly as recommended.

10

image

1 Like
11

Hokay… latest version lol

SCHEMATIC REMOVED - SEE FIRST POST FOR LATEST REVISION

As for U2A, that is my understanding of its function, driving the two vactrols in a linear fashion until the voltage is high enough to trigger the Zener diode, at which point the current rises exponentially until you max out the output on the OpAmp. So its a controllable upwards ramp with a ‘kick’ at the end of its range (or so Muff Wiggler tells me).

R22 is pretty low though, at 470Ohms, as that part of the circuit is supposed to work just fine at 12v I’ll leave it be for now I think, maybe try a little up and a little down when on the breadboard and hope I don’t cook anything. Those vactrols are almost the most expensive part in the whole project and I don’t want to go through too many of them, I only ordered 4 for the first build (assuming I would cook it at least once, and hoping for the best)

12

What does the 3pdt switch do? That’s a more obscure part

1 Like
13

Ok, so from what I understand of the intricacies of the filtering element of this circuit -in the Upper position it grounds the filter circuit so it doesn’t feed back into the output and disconnects the resonance section leaving you with just an attenuator going to the output buffer for a simple VCA, in the middle position it leaves both the VCA style attenuation and the filter circuit connected, but disconnects the resonance section so you get a traditional Low Pass Gate, then when its in the lower position it disconnects the attenuation of the audio signal and adds in the resonance section to give you a Low Pass Filter with adjustable resonance.

Hence my desire to run multiple of these in series so you can do both low pass gating or vca’ing of a previously filtered (with resonance) audio signal :stuck_out_tongue:

2 Likes
14

It’s funny tayda has loads of 3pdt stomp switches but only the on/on mini toggle

15

Can’t wait to see what u do with the faceplate. Did you get @analogoutput ‘s Kosmo hole library from his GitHub? Super helpful.

You’ll have to sort out how the build is laid out, with two boards or wired jacks, and tweak the schem before u start laying out the pcbs

3 Likes
16

Yeah, I can get them easilly enough as a toggle from Farnell here in the UK (they are also based down the road so before Covid I could stroll along and pick up this kind of thing from the counter which was awesome), but they cost a bomb for a micro toggle, like £5.50 +VAT each. Single most expensive part in the whole thing, and that includes getting a custom panel from JLCPCB

2 Likes
17

Yeah, I am hoping I can import artwork from something like illustrator for it, but that is the next thing to work on tomorrow now I have this schematic nailed down-(ish) and waiting for parts. As for the hole library, no I have not! I will swing by and pick it up

For this one I am hoping that I will be able to get it on a single 5cm wide board with connections for the jacks a-la Sam’s early modules. If/When I get round to doing its big brother I will likely go with two boards, one for the guts and another for the pots

1 Like
18

@d42kn355 is a whiz at the art import game and has done most the tricks

19

and as a side note PLEASE indicate untested on your schematics when you show them in these threads so idiots like myself don’t take them as gospel and try to build them as is and wonder why they didn’t work … .

4 Likes
20

Lol, I figured that would have been clear from the OP but you have a point, I’ll make it more obvious now… :wink:

Did you try to breadboard it? I’d love to hear how it turned out, where the bugs were etc!

1 Like