My first DIY modules - starting a Kosmo format build

[juggle]

@fredrik

The PWM control range is 0−Vcc/3, see e.g. here. The values in the datasheet are for 15 V supplies.

Ahhh! That makes sense, and I think I remember reading that in the sheet now that you point it out.

I was just looking at the reference schematic on the Electric Druid site which labeled it as 0-5v. I should have paid better attention to the datasheet itself instead of a schematic derived from it!

I did a bit of playing with my layout to convert it to board mounted pot’s…and have something that looks like it should work, has the correct measurements as far a I can tell, and passes DRC checks:

Except…the PWM CV pot is on the wrong side for my panel layout. So either I have to change my panel to put the PWM CV on the left and frequency CV on the right…which just feels backwards to me. Or I have to find a way to move that pot over to where the inputs are now…and that creates a major mess.

I also played with the kasu3340 layout a bit more…started routing it…and…ugh. That’s getting ugly faster than I expected. Might be just a bit more advanced than I’m ready to tackle just yet. Need to do a few more simpler designs first.

[juggle]

So I realized…I also had the pots on the wrong side of the PCB in that last revision. Sigh.

So…swapped them all around, moved the PWM pot which was as big of a pain as I expected. And then realized…the PWM pot was too far from the edge and the board would interfere with the PWM jack. Double sigh.

So…had to move the I/O connectors as well, and the power to the other end…reshuffle and reroute a bunch of little things…and I think it’s about good now:

Only 4 jumpers…but 3 of them are longer and more ackward than I’d like. But I don’t see any way to improve on them.

This SHOULD work though…Should give me a pot mounted version of my simple 3340

I’m tempted to add the kassu sine output to it as well…but that would require reworking my panel to fit another output jack. But…there probably IS room for the circuitry…hmmm…

[juggle]

If I didn’t know any better I’d say I’m getting better with practice

Well…mostly. I still managed to make some dumb mistakes.

Namely I forgot to put my etchant in a heated bath…so it took longer than normal…and didn’t etch quite as well. It was hard to tell when to pull it as a result and while it looked fully etched. There were a lot of “whiskers” bridging various traces Also - I didn’t quite position the board on the negative correctly and one trace almost got cut off on the right side. enough of it was still there to have continuity though so I ran with it.

Oh - also I damaged the mask along the bottom…but since it was all ground I wasn’t too worried. I gave it a bit of sharpie to help…but it still got a bit etched.

Not my best board…but should be good enough!

Ok…maybe this one is cursed. Broke 3 drills drilling it. Well, re-drilling a few of the holes by hand because I needed them a little larger.

But…not going to let that stop me:

Alright…one other mistake. I had planned on putting the fine tuning trimmer on the other side so I could adjust it with the module installed. But…not going to happen. I forgot there’s no way to solder it when it’s installed that way since the pads are hidden under it. Plus I forgot to add a hole for it in my panel anyway.

Speaking of the panel - it took just over 2 hours to print…and finished at almost exactly the same time I was ready to install the board in it I started the print just before I printed the negative to etch the board. So etching and assembling the board took almost exactly the same time as printing the panel to go with it.

It definitely makes for a more sturdy module!

I still had some debugging to do on it though. I stole the jacks from my other DIY 3340 to save time but the outputs are in a slightly different order this time so I had them mixed up at first. Super easy fix.

Bigger problem was once again PWM wasn’t doing anything. Oh nuts…not this again!

But I was able to track it down to a solder bridge due to the lack of solder mask (Soldering the pots is a bit tricky with no pads on the back…and getting the iron in there I accidentally caused a bridge.) And one trace that was still shorted out to ground due to the incomplete etch.

With those two issues fixed it works perfect!

So - now I’ve got 3 VCO’s for my case. I definitely need to finish that other power supply so I can get all these in there. I think the mega drone is going to move to it’s own case for now until I finish up the other case I’m working on.

And I need to wire up this ARP4072 and see if it works still.

So I’ll probably hold off on the Kasu3340. Try a few other boards first and work up to it. I still want a nice snare. Or a 808 clap to go with my 808 kick…of course I still need to finish the trigger outputs for the sequencer and finish the case and power supply for the sequencer so I can use it without boards laying out and janky wires to a bench supply!

[eric]

Nice.
Two hours for a build definitely beats waiting two weeks for PCBs from China…

@juggle

(Soldering the pots is a bit tricky with no pads on the back…and getting the iron in there I accidentally caused a bridge.)

You can use “longer” pads for the pot-pads, like those on ICs footprints.
Should make soldering easier.

[Farabide]

Very impressive!
Side chain the triangle and make a mini strip for sine wave.

[juggle]

@eric

Two hours for a build definitely beats waiting two weeks for PCBs from China…

Yeah. I had more time into doing the PCB layout changes than making and assembling the board. Though having a proper two sided board with plated vias would solve the issue of the trim pot soldering to get it front facing. I’m not sure longer pads would really help with it…the way the trimmer is shaped it’s just hard to even see under it to where the pins are. I thought about heating the pins from the back but I couldn’t even get solder in from the front because the body of the pot just shields everything.

Longer pads would help with the normal pots. If I had seen existing footprints with longer pads I’d have used them - I use them everywhere else they’re available. But I’m still figuring out how to edit footprints so didn’t want to take the time to make custom ones with longer pads.

@Farabide

Side chain the triangle and make a mini strip for sine wave.

Before I even dug in on making more VCO’s I started looking at making a simple tri-sin module by itself. But didn’t want to use up another power connection and space in the rack just yet Might do it once I remove the megadrone though…

[ElectricalMusineer]

@juggle These look super good! Do you print the face plates with 2 colors of plastic or paint the letters afterwards?

[juggle]

@ElectricalMusineer

These look super good! Do you print the face plates with 2 colors of plastic or paint the letters afterwards?

Thanks. I have a Prusa Mk3s with MMUs2 so I do two color prints to get the letters. Though I occasionally add some with sharpie when I forget them or mess them up and don’t want to take the time to print a replacement.

[juggle]

So…I need some more power supplies. Seeing the video from Moritz Klein got me thinking. I just used the FC power boards for my last two power supplies…but they’re $12 each and another $10 or so for shipping. I want 3 more power supplies (2 for my new case and one for my bench) so almost $50 just for 3 PCB’s seems a bit steep.

Since I’ve been having fun etching my own boards and power supplies are a fairly simple circuit…I did a bit of playing in kicad this evening:

(Note - for some reason the 3D rendering in Kicad is making these bigger than they should be. The actual board outline ends where the copper pours do.)

Yeah…I took a lot of inspiration from the FC board. Not just for the circuit (though I eliminated the LED’s, 5v, and eurorack connector since I have no need or desire for them) but for the layout as well.

Though - since the boards I’m etching are 70mm x 100mm I thought I may try to see if I could fit this into 35mm x 100mm so I could do two of them on one board.

Well…not quite

Electrically sure it fits. But…it’s REALLY tight…and there isn’t room for the heat sinks on the regulators. And the terminal blocks I wanted to use really take up a LOT of space.

If I eliminate the terminal blocks I can probably make them fit.

But…

70x100 blank boards are cheap…so I could just do one per board and have plenty of room. Or I could just send these off to JLC…$7 for 5 delivered is a lot better deal than the FC boards (not that I have any problem with them or begrudge them making a profit - far from it…I gladly bought my last two supply boards from them.)

Also…I’m tempted to try these as an “easy” dual sided etch. My initial layout has all of the ground on one side. But - soldering to it under the caps would be quite difficult! And I don’t yet have a way to plate vias.

I could easily just add a few vias in a few spots and solder some cut off leads on both sides to tie the top ground together…but…I think if I etch them myself it’s just more bother than it’s worth to go dual sided on this.

I’ll probably just spread things out and fill a 70x100 board with it then etch it myself. I ordered enough parts from Tayda today to build 3 more of these (for less than I spent for my 2 original blank FC boards) so I’ve got a few days to think while I wait for those.

[juggle]

Ah…I see why the board outline was wrong. I had it on the Eco2.User layer instead of the Edge.Cuts layer. Oops.

And…yeah…piece of cake fitting this on 70x100 with no jumpers:

Have a few days to think about optimizations and aesthetics though since I’m waiting on Tayda. Like…I’m not huge on the regulators being hidden inside the caps like that. Not great for airflow around the heat sinks. So will probably shift things around quite a bit before I etch anything.

[eric]

Some suggestions :

[analogoutput]

This might be silly, but I’d consider putting the big caps in the middle and the regulators on the outside for better air flow to the heat sinks.

Also, the four diodes on the right with three facing one way and one facing the other strikes me as an assembly mistake waiting to happen.

Also also, you do know there’s a single sided PCB layout for a wall wart ±12V supply here, right?
http://musicfromouterspace.com/analogsynth_new/WALLWARTSUPPLY/WALLWARTSUPPLY.php

A tad too long for your blanks but maybe it’ll help your design?

[juggle]

Great feedback guys! I honestly just tossed that together before bed and didn’t even give it a close look myself yet. Well, I did glance at it and noticed that one rectifier diode was reversed in the circuit which I fixed before taking those screen shots.

I actually was thinking about putting the big caps down the middle and moving the regulators outside so they’d get more airflow - that or stacking things with the caps all on one side and the regulators on the other. Or doing two vertical rows of caps and having gobs of room for the regulators/sinks to the right.

The three facing one way and one the other bugged me as well - but I was thinking about bigger layout changes to address some of the issues eric pointed out so wasn’t that worried about it yet. I was also toying with eliminating one of those since Moritz pointed out in his video that 7912’s have a built in protection diode (which I had never heard before and wanted to confirm before removing it.)

I did see the MFOS design but wasn’t a huge fan. I didn’t like the regulators being flat against the board (I like having some heat sinks on them not just trusting the ground plane to be enough of a sink) and the positioning of the output pads rubs me the wrong way. But it would definitely go under the list of inspirations for doing my own.

Plus…I can read a schematic for a week…but redrawing it and laying out a board for it seems to help my brain get wrapped around it WAY more than just reading it over does. So it’s a learning experience both as far as circuit operation and more practice at board layouts (as well as a chance to play with chunkier traces.)

[analogoutput]

Agreed, I’m taking the MFOS circuit and drawing it up, with changes — as you say, vertical regulators with heat sinks, and a barrel connector for the input and terminal block for the output, and polyfuses and LEDs. And getting it under 100 mm (but wider than Wilson’s).

[juggle]

I considered adding a barrel connector for the input. But…since the supply is inside my case I decided I’d rather have the barrel connector mounted to the outside of my case and then wire that to my supply boards - so I put the terminal blocks on for input.

Right now I just have my supply wire running through a gap between modules and want something cleaner.

I’d like to add a fused/switched IEC320 jack so I can just use standard PC power cords to power the boxes…but I don’t’ want to hack the wall wart, or just mount a wall wart inside the box. I’m half tempted to do my own transformer…but…I’m also smart enough to know that only being half tempted and unsure of myself is good enough reason NOT to take that path

So a small panel with the barrel jack on the outside seems like the best option for now.

[analogoutput]

My rationale was that if using a couple boards in parallel off a single wall wart, using the barrel jack makes it easy to disable one if necessary.

[juggle]

I was thinking about using a pluggable terminal block (Phoenix 1757019) for the best of both worlds - keyed so no risk of swapping polarity (not that it matters on the AC side) screw terminals for wire attachment, and pluggable so can be connected/disconnected.

They’re used on my RAMPS and Smoothieboard 3D printer/Laser/CNC control boards and seem like the best solution for this. (as long as they’re sized correctly, RAMPS is notorious for having too small of a connector on the main power and melting them down. But the heaters in a 3D printer draw quite a bit more power than we’re dealing with here!)

[analogoutput]

My current iteration:

(100x61 mm)

Edit: Swapped barrel connector for pluggable terminal block. Thanks for that tip!

[juggle]

Very nice. A lot more thought and effort than I’ve put in so far for sure!

I like the addition of the polyfuses…I admit LED’s are handy for debugging…but I have a DMM for that and since the supply is hidden away I figured it wasn’t worth the effort of including them.

[analogoutput]

The LEDs and their resistors also are a load which (at least on the negative side) is needed for testing with nothing connected to output. Won’t be seen much once there’s a full set of modules in place, but I like having them until then.