The Karplus Strong eChucK DIY Project

Dan Lavin · Posted: Thu Nov 26, 2009 6:56 am Post subject:

oops...I didn't read your second paragraph.

I think you could generate a pulse and use an op amp to mix in a little audio signal. Any idea of how much modulation we're talking about? I'm guessing in the mV range instead of volt range.

Inventor · Posted: Thu Nov 26, 2009 7:23 am Post subject:

Dan, thanks for your questions, stuff like this makes it fun for me.

The input pulse amplitude controls the output amplitude linearly, so send in softer pulses for softer sounds, etc...

To patch in audio, just send it in on the Stim input. I have an extra opamp to provide the voice / audio input if you like.

Les

Inventor · Posted: Thu Nov 26, 2009 7:50 am Post subject:

Tell you what, let's have a little fun here. Howard won't like it, but how about a little competition?

Let's say that everyone who gets a Karplus Strong board does some experimentation and makes design suggestions. After some time we all vote and the winner gets a free second run circuit board!

I see this as a way of boosting the energy level of your brain power, nothing like a little competition, eh?

Les

Dan Lavin · Posted: Thu Nov 26, 2009 8:03 am Post subject:

Les,

OK, so the amplitude of the stim pulse is critical...if you want harmonics not to be static. Desirable but not critical. But to make it interesting, it should be done.

Right now, you've got it tied to the CV out, which makes it permanently tracking the pitch. This will give it a more acoustic feel since harmonics generally increase in acoustic instruments as the pitch goes up. However, it may be desirable to not track pitch. The stim should have a mixer so that pitch, velocity, noise, LFO, etc can be mixed into the signal.

My power chord guitar idea would definitely require a velocity input. But maybe I should just shut the heck up so I have a chance to win your contest!

Inventor · Posted: Thu Nov 26, 2009 12:09 pm Post subject:

Dan, I sure hope I didn't tie Stim to CV, that wouldn't work at all. The board has jumpers which when populated short the two CV inputs together. Also if the eChucK circuitry is present, a PWM on the right channel sets the CV level for the current note. Stim comes in on the left channel directly into the summing amp. So that's the deal on that...

Les

Inventor · Posted: Thu Nov 26, 2009 6:44 pm Post subject:

Inventor · Posted: Thu Nov 26, 2009 7:24 pm Post subject:

In other plans about the BS board, the clock and reset need to be synchronized among many boards. It needs to be set up so that many boards can all receive clock and reset from one board, or run independently. I think we need a switch for this, not sure - you tell me.

Les

Inventor · Posted: Thu Nov 26, 2009 8:53 pm Post subject:

I generated a bill of materials (BOM), attached. This is just the default BOM that bom.ulp generates, and I have not added DigiKey part numbers. Anyway, it's a BOM.

Les

KS7BOM.txt

Description:

KS7 BOM

Download (listen)

Filename:

KS7BOM.txt

Filesize:

4.36 KB

Downloaded:

402 Time(s)

Dan Lavin · Posted: Fri Nov 27, 2009 6:48 am Post subject:

Les, Yes you will need a switch for the clock and reset signals on each of the boards. I like the idea of "Master" and "Slave" settings for the switches, but nomenclature is up to you.

Inventor · Posted: Fri Nov 27, 2009 10:12 pm Post subject:

Inventor · Posted: Sat Nov 28, 2009 12:02 am Post subject:

Well I put in a clock selection switch, but there are still details about how these things will be wired up. Specifically, in the eChucK configuration we need some jumper wires that have a low pcb footprint. I was thinking of headers. We would need to daisy-chain the reset i/o and clock in and out lines somehow. I think the cable needs a twist in it or something, but I really haven't thought it out.

For the stim I'm thinking of using the two spare TLO74 opamps as comparators and then only when stim is high AND clock is low do we generate a stim pulse. There is no room for an AND gate, but I can make one out of RDL or RTL or LEDIL or something. We'll see how that goes.

Any thoughts? Suggestions?

Les

wmonk · Posted: Sat Nov 28, 2009 5:35 am Post subject:

You don't have to use the switch if you use switched jack sockets. You need a clock out jack on board A, and a switched clock in on board B.

Inventor · Posted: Sat Nov 28, 2009 11:42 am Post subject:

wmonk · Posted: Sat Nov 28, 2009 12:02 pm Post subject:

The jacks don't have to be board mounted. Then you only need three holes for a jack, and one can choose which type of jack he wants.

Inventor · Posted: Sat Nov 28, 2009 1:00 pm Post subject:

Inventor · Posted: Sat Nov 28, 2009 1:39 pm Post subject:

Oh, and by the way I've received payment for four of the ten boards and I need one for myself, so there are only five left on a first-come first-serve basis. This should work out just about right I think, no boards to spare and everyone gets one, but place your order early to ensure you get one.

Les

wmonk · Posted: Sat Nov 28, 2009 1:41 pm Post subject:

Perhaps you can do that with other components that have to be 'panel mounted'. Saves some space and in this case it would be nice to implement all the needed features.
You didn't post the layout you have now for the BS, but I can imagine there are some other space-saving tips. Like using vertical mounted trimmers instead of the standard horizontal mount.

And for spacing: use standard 0.1" header spacing for panel mount components.

Keep up the good work! Smile

Inventor · Posted: Sat Nov 28, 2009 2:00 pm Post subject:

wmonk, here are the updated BSG schematic and board layout. As you can see the board is very tight, but there is some room for a few 0.1" vias to connect to those fancy switched jacks.

I believe that the counter, the LEDIL, and the CV circuit are fine. The clock and reset circuit needs attention as does the Stim circuit. Comments welcome.

Les

BS7sch.pdf

Description:

BS7 Schematic

Download (listen)

Filename:

BS7sch.pdf

Filesize:

46.04 KB

Downloaded:

251 Time(s)

BS7pcb.pdf

Description:

BS7 Printed Circuit Board

Download (listen)

Filename:

BS7pcb.pdf

Filesize:

81.47 KB

Downloaded:

224 Time(s)

Inventor · Posted: Sat Nov 28, 2009 2:06 pm Post subject:

Does anyone have a manufacturer and part number for those switched jacks? I can't seem to find one in DigiKey search.

Les

wmonk · Posted: Sat Nov 28, 2009 2:13 pm Post subject:

What about using rectangular LEDs (2 x 5mm)?
Posted Image, might have been reduced in size. Click Image to view fullscreen.

That would save some rows of space!

I use Switchcraft 112A 1/4" clones as jack. Always use switched type, and most of the time I don't use the switch, but hey, then they all look the same! There are many types available.

Inventor · Posted: Sat Nov 28, 2009 6:52 pm Post subject:

OK, thanks to wmonk's help here and in the chatroom, we have better LEDs, a switched jack solution to clock/reset synchronization, and some other features. Still working on it though...

Les

Inventor · Posted: Sat Nov 28, 2009 8:13 pm Post subject:

Here are the BS8 schematic and pcb pdf files for your review. The reset switch needs to be changed to a SPST but aside from that it's about done. Or so I hope. I guess I'd better prototype it before sending it out, huh? Oh well we have time, no rush. Please review the documents and offer your critiques if any, thanks.

Les

BS8sch.pdf

Description:

BS8 Schematic

Download (listen)

Filename:

BS8sch.pdf

Filesize:

39.74 KB

Downloaded:

261 Time(s)

BS8pcb.pdf

Description:

BS8 Printed Circuit Board

Download (listen)

Filename:

BS8pcb.pdf

Filesize:

72.97 KB

Downloaded:

249 Time(s)

mrcold · Posted: Mon Nov 30, 2009 2:02 pm Post subject:

hey les.. just listened to your Tuesday Show... Nice work!

looking forward to more!

Inventor · Posted: Mon Nov 30, 2009 2:58 pm Post subject:

Thanks Cody, glad you liked it!

Les

Inventor · Posted: Mon Nov 30, 2009 10:03 pm Post subject:

By the way, on the BS board I have set up three sizes of DIP switches for the logic matrix: 12, 8, and 4. Each has two terms, so that the total number of terms is 12, 12, 8, 8, 4, and 4. In contrast, I could go back to the earlier way of doing it with six equally sized DIP switches of 8 bits each.

The trade-off here is that with only 8 bits, and clocking at 4 Hz, we get a 64 second song which is kind of short. Using all 12 bits we get a song 16 times longer, or 16 minutes. That's the timing, and the second trade-off is the flexibility. With the existing arrangement there is not as much that you can do with those 12 bits of song compared to six 8-bit terms. So that's the compromise.

Do you think I should leave the board as-is or change it to six 8-bit terms?

Les