The Karplus Strong eChucK DIY Project

Blue Hell · Posted: Wed Nov 18, 2009 6:11 pm Post subject:

Oh, and I was wondering ... you pick the output from after the BBD, why not from before it? It should be the same signal with less distortion and coming in a bit earlier ... or why not tap it from after the filter, it would be the same signal again, but in this case without the dry stimulus mixed in.

Inventor · Posted: Wed Nov 18, 2009 6:17 pm Post subject:

Blue Hell · Posted: Wed Nov 18, 2009 6:22 pm Post subject:

Ok, was wondering Laughing

Dan Lavin · Posted: Wed Nov 18, 2009 7:20 pm Post subject:

Les,

thanks for posting a recording of the show. I made my wife and son listen to it while we got dinner ready tonight. (Thanks for not swearing!)

Anyway, I think with pitch control and variable gating from a sequencer or a keyboard interface, this will make a fine instrument. I think I like the bass guitar effect the best. I'm kinda looking forward to combining 3 of these and running them thru some of my guitar pedals to create some monster power chords!
-Dan

Inventor · Posted: Wed Nov 18, 2009 9:27 pm Post subject:

Dan, what a cool application for the KS board. I can see this thing is going to be creatively used in all kinds of ways.

Some notes about the project:

o We're using a TL074 opamp but I do not know if it's inputs and output are rail to rail. We really should be using a rail to rail opamp to get full range out of the OTAs.

o I'm adding a resistor in parallel with the R23 pot to reduce it's value and make a more sensitive log-ish angle-to-resistance mapping. This will make it either better or worse, but if it doesn't help we can just not populate it, so I'm just adding it and we'll see if that's a good idea or not later.

o I will add resistors and diodes in the feedback loop of the output buffer opamp to perform distortion if desired. Do not populate if you don't want distortion, and adjust resistor values to shape the distortion.

o Two people have expressed interest in populated KS boards, so I'm thinking of having E-TekNet populate some of the second run boards.

o I will space the pots on 1" centers

o I will move the lower spacer mounting holes to create a larger potentiometer and switch region for use by other boards that have yet to be designed.

o On the 2nd run of boards I want to buffer the outputs of the OTAs with rail to rail voltage followers, but there is no time or place for another chip at this point in the development process.

o There are five people who want boards for DIY, and one who wants a populated board, plus I'd like a spare, so that's 7 boards which is enough for us to make a first run of boards. I will contact E-TekNet and ask about pricing.

o The pots will be tripple-footprinted to accomodate 16mm Alpha pots, Spectrol pots, and Bourns pots. Thanks to wmonk for that idea.

o I have a request for wmonk. In the charoom, wmonk posted the available bucket brigade device chips and the number of stages in each. The were stages of 128, 256, 512, 1024, and 2048. So wmonk, would you please post your part numbers for those devices? That would be really helpful, thanks.

That's all for now, Les out!

Inventor · Posted: Wed Nov 18, 2009 11:03 pm Post subject:

Another update here. I just glued two 1/8" jacks to the breadboard's wooden base using JBWeld, which I believe is the glue that God uses because it bonds almost anything to almost anything else. It will take 24 hours for the glue to fully dry, so I have some time to type out my plans for you to read.

I'm going to use these connectors to connect the headphones out and line in of the eMac to the Karplus Strong board. This will allow me to provide any desired stimulus and CV signal using ChucK, and also ChucK will receive the output signal of the board. I'll add a microphone circuit going into the remaining channel of the line-in jack so I can talk and interface with the board at all times.

Then I can send in any desired waveforms and listen to the response. I can also record a show with this setup for you to hear. This approach lives up to the eChucK name by finally giving ChucK some modular synthesizer hardware to play with, haha. ChucK, he loves his toys don't you know?

Anyway, I'm off to code the software while the glue dries, so catch you all later.

Les

wmonk · Posted: Thu Nov 19, 2009 12:24 pm Post subject:

Inventor · Posted: Thu Nov 19, 2009 8:04 pm Post subject:

TP, thanks for all that cool info. I will add the three trimpots for level and distortion as per our discussion in chat.

I've receive a quote from E-TekNet. The charges are as follows:

10 Boards: $120.13
NRE: $89.00 - Proto Special: no NRE: $0.00
S&H: $26.22

And the 50% markup does not apply to shipping, so the actual cost per board to everyone would be:

DIY Cost: $21.00

Please let me know if you find that cost reasonable for a first-run board.

Les

edit: subtracted the NRE

Inventor · Posted: Fri Nov 20, 2009 8:42 am Post subject:

About the no NRE, I spoke with Al Fontain at E-TekNet and he clarified that our board is simple enough to manufacture that it qualifies for this special deal that they have. We can get the boards made with no NRE charge! That's a significant savings and it brings the price down to a reasonable level.

Les

Inventor · Posted: Sun Nov 22, 2009 6:55 pm Post subject:

Here are the latest board and schematic files. The input circuit from the eChucK jack will change, but everything else is final. Please have a look and let me know if there are any problems.

Les

KS5brd.pdf

Description:

board pdf

Download

Filename:

KS5brd.pdf

Filesize:

63.81 KB

Downloaded:

52 Time(s)

KS5sch.pdf

Description:

schematic pdf

Download

Filename:

KS5sch.pdf

Filesize:

33.08 KB

Downloaded:

72 Time(s)

Dan Lavin · Posted: Sun Nov 22, 2009 8:22 pm Post subject:

Les, That's cool that you gave Jan credit for the level shifter right on the schematic. You/he should trademark that!

Blue Hell · Posted: Sun Nov 22, 2009 8:28 pm Post subject:

thanks Les!

Inventor · Posted: Sun Nov 22, 2009 9:47 pm Post subject:

Inventor · Posted: Sun Nov 22, 2009 11:05 pm Post subject:

Jan helped me set up a BJT amplifier for the PWM CV input from ChucK. It works fairly well. The attached song repeats a few times. It is a boolean sequenced karplus strong circuit. The boolean sequencer is in ChucK and ChucK generates the stim and CV signals. Tell me what you think!

Les

Les Hall_Mon Nov 23 005253 2009.mp3

Description:

test audio file

Download

Filename:

Les Hall_Mon Nov 23 005253 2009.mp3

Filesize:

1017.14 KB

Downloaded:

108 Time(s)

electri-fire · Posted: Mon Nov 23, 2009 5:16 am Post subject:

Nice! I'm looking forward to mine.

Dan Lavin · Posted: Mon Nov 23, 2009 3:17 pm Post subject:

Les,

I finally listened to the mp3. That's what I was expecting when you announced this project! Very nice. It'll really sound good with drums and a couple more higher voices doing the chords.

Inventor · Posted: Mon Nov 23, 2009 4:35 pm Post subject:

Yes Dan, thanks, and now I've got it putting in different signals riding on the pulse - noise, triangle, and sine waves. I can send in just a signal burst or any amplitude of pulse or both, and i can vary the frequency of the added signal. It all makes different sounds, mostly percussion.

DGTom · Posted: Mon Nov 23, 2009 6:05 pm Post subject:

woah! Great stuff!!! Really tasty sounds there Very Happy

I think that price is more than fair for a micro run board & just to put it into context; to get similar sounds to that from my set-up at the moment would take 3-5 modules! So even once its built... still fantastic cost to function ratio!!

Inventor · Posted: Mon Nov 23, 2009 11:31 pm Post subject:

Inventor · Posted: Tue Nov 24, 2009 11:58 pm Post subject:

Alright, everyone has had a chance to look at the schematic and board placement without comment, so I'm going to go ahead and send the board design out on Wednesday. Also please find attached the beginnings of some documentation, comments welcome as always.

Les

KSDocumentation.pdf

Description:

Download

Filename:

KSDocumentation.pdf

Filesize:

34.46 KB

Downloaded:

148 Time(s)

KS6.pdf

Description:

Download

Filename:

KS6.pdf

Filesize:

34.32 KB

Downloaded:

42 Time(s)

Inventor · Posted: Wed Nov 25, 2009 4:04 pm Post subject:

Now that the KS board is off to the board house, I started working on the companion Boolean Sequencer board. Schematic attached. I feel that the LEDIL and CV work ok, but the stim is totally fubar. Plus the clocking scheme needs to be improved. Please have a look and comment, thanks.

Les

BS7sch.pdf

Description:

Boolean Sequencing Schematic

Download

Filename:

BS7sch.pdf

Filesize:

45.63 KB

Downloaded:

30 Time(s)

Dan Lavin · Posted: Wed Nov 25, 2009 8:02 pm Post subject:

Les,

What kind of signal does the Stim need to be? Sorry I'm fairly ignorant on K-S right now. Hopefully I'll be an expert one day...or at least a fair amateur!

Inventor · Posted: Wed Nov 25, 2009 11:40 pm Post subject:

Dan Lavin · Posted: Thu Nov 26, 2009 6:50 am Post subject:

OK Les,
That's kinda what I was thinking. So basically you need a pulse about 1 to 10ms? Is the amplitude critical or is this just a fixed amplitude pulse?

If you just need a fixed pulse, it's probably better to tap off of the clock. So the modular synth equivalent would be a trigger (and not a gate). There should be plenty of CMOS trigger circuits out there that use a resistor/cap to set the pulse width. You can fix the cap value and use a pot to vary the width. Or maybe I'm ahead of myself? That probably should be done at the voice(s)?

Let me know if I'm on the right path and you'd like suggestions on adjustable pulse width trigger circuits. I'll probably overkill it since this project is a minimal parts one, but maybe it'll get other folk's gears going.

Blue Hell · Posted: Thu Nov 26, 2009 6:54 am Post subject:

KS also makes for a nice resonator when you set the feedback a bit on the low side and the input (STIM as Les calls it) signal has a safely low amplitude.