Controller, Ribbon, Softpot, another.

frijitz · Last edited by frijitz on Sun Mar 01, 2009 3:40 pm; edited 2 times in total

So ... after playing around with a 500 mm Softpot for a few weeks I've discovered that I'm probably not going to be able to develop any virtuostic technique with it. I did find, however, that I really enjoy using it with a breath controller to get all the slips and slides that are not possible with my more traditional wind controllers. So I'm going to emphasize working with a fairly bare-bones controller operating as a kind of sliding flute. But of course keeping in mind that the usual more advanced features can easily be added at any time.

The attached schematic shows the current version of the electronics. I've tweaked this up fairly carefully, so I don't think there will be any changes in the core circuitry. It works quite reliably, although the Softpot seems to take an awful lot of pressure to operate. This leads to a bit of pitch shift due to the "squish factor" on finger-up. But it's pretty easy to compensate for this with a bit of a roll of the finger as it comes off the ribbon. Of course this doesn't matter, if the CV hold on finger-up is not used, i.e., if the VCO gets gated off. (We see a very successful unit here using that approach.)

OK, here's the geeky stuff about how it works.

A1 and associated components form a standard current source that drives about 1/4 mA into the ribbon. The CV picked off from the ribbon's touch point is buffered by A2, then lagged, scaled for 1V / Oct (high-note priority) and presented to the input of the S/H circuit built around U1 and A5. The purpose of the S/H (actually operating as a track-and-hold) is to grab and maintain the last CV produced before finger-up. The S/H itself is a feedback design employing an OTA switched current source and a FET opamp buffer. Droop is about 2 mV / sec using a 0.022 uF polypropylene storage cap.

So that's the simple part. The rest of the core circuitry is for turning the S/H on and off, which is the trickier part. This is accomplished with two separate circuits.

The crucial consideration here is the method used to detect the finger up/down transitions. In previous work, the method used was to pull the ribbon's contact strip to a power supply rail with a large resistor. This resistor must be large (>2M) to avoid loading the signal from the ribbon. For the Softpot, this method is unsuitable. The reason is that the Softpot has a large capacitance (280 pF), which causes the finger up/down signal to be delayed by a significant time interval, making it imposssible to capture the last voltage prior to finger-up before it decays significantly.

The solution to this problem is simple -- just use the usual method of detecting current in a circuit, which is to add a series resistor in the current path and measure the voltage drop across it. In the schematic, this resistor is the 8.2 k unit drawn above the ribbon. The finger up/down transition causes a 2 V change in the voltage across this resistor. This change is rapid, because there is no large resistance in the circuit, rendering the ribbon's capacitance irrelevant. Additionally, having the ribbon at low impedance avoids potential trouble with noise pickup.

On finger-down, the voltage on the ribbon's pickoff point falls from about 7 V to < 5 V. Comparator A7 detects the 6.2 V crossing and feeds the 3.3M / .022 uF lag network. The lag delays the start of acquisition by ~30 ms, enough time for the finger to squish down. When the lagged signal crosses 0.6 V, U2 is turned on. U2 is an OTA used as a fast current switch. Its output turns on the S/H switch U1, and the U1/A5 combination begins tracking the changing ribbon voltage.

On finger-up, the lag circuit's capacitor is discharged quickly through the 1 k resistor and the diode (to the right of A7). This holds the OTA currents off. However, this process is too slow for accurate acquisition of the ribbon voltage at finger-up. A special, fast, auxilliary circuit provides a much more rapid initial acquisition. On finger-up, the voltage across the 8.2 k resistor in series with the ribbon goes from 0 V to 2 V. As soon as it starts to rise, comparator A6 switches and triggers a pulse from the 555 IC timer chip U3. This pulse drives U2's (-) input, shutting off the OTAs for a long enough time to allow discharge of the lag capacitor. The key to this operation is that A6, U3 and U2 are fast switches, and they are biased near their switching points for minimal switching delay.

Finally, A8 develops a gate signal, which is high on finger down. This can gate the VCO signal (through a VCA) to produce sound only when the finger is down. Like Mr. Flicker-Fingers here.

This circuit is being provided gratis with no string attached for experimenters. But no commercial use, please, and no copying of the schematic to other sites, blogs, etc. Thank you.

Smile

Ian

Fretter_01.gif

Description:

Modified 3/1/09

Filesize:

10.63 KB

Viewed:

1883 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

RF · Posted: Mon Feb 16, 2009 9:24 pm Post subject:

Awesome - Thanks Ian!

Dragon's Lair · Posted: Mon Feb 16, 2009 9:34 pm Post subject:

I have been playing around with the 500mm Softpots also, but mine take very little pressure to activate, do you have more then one softpot to try with? you shouldn't have to "squish" your finger to activate it.

frijitz · Posted: Mon Feb 16, 2009 10:33 pm Post subject:

Dragon's Lair · Posted: Tue Feb 17, 2009 12:38 am Post subject:

I got 10 of them directly from Spectra Symbol, they shipped them in a tube, so they are straight and flat.
Do you have them mounted on a base? I have a 200mm and a 500mm mounted to a piece of melamine, along with a power supply and some knobs, for testing purposes. I found that if you don't stick them down to a solid surface, they can be a bit hard to use.
With this setup, I use about the same force on them as I use on the touch plate of my laptop, and have no problem sliding my finger up and down it.

frijitz · Posted: Tue Feb 17, 2009 8:57 am Post subject:

Yes, I have them mounted on bases. (Photo on Ondes Martinot thread). They certainly require much more force than my laptop touchplate. Hmmm ... Maybe we should try to figure out a way to measure the contact force so we can compare? I'm going to be very angry if it turns out that Sparkfun damaged them.

EDIT: I put a small square of balsa wood on the ribbon and stacked on coins until it switched. I takes 9 quarters. Maybe other folks could try the same experiment?

Very Happy

Ian

RF · Posted: Tue Feb 17, 2009 9:59 pm Post subject:

Hi Ian -

I placed a Digi-Key order just a day before you posted this - Sad

Just how critical is the OPA134 to the circuit? I'm curious as to why you used that particular opamp - and if there might be a reasonable sub for this circuit?

Thanks for the insight...

bruce

Dragon's Lair · Posted: Tue Feb 17, 2009 10:09 pm Post subject:

There is definitely a difference in the softpots, mine only took 6 quarters to activate.

frijitz · Posted: Wed Feb 18, 2009 7:53 am Post subject:

frijitz · Posted: Wed Feb 18, 2009 8:01 am Post subject:

RF · Posted: Wed Feb 18, 2009 8:57 am Post subject:

Thanks Ian -
I appreciate the response. The OPA134 isn't something I have in the parts box here - so I'll try some others. I'll likely order some OPA134's next time I order parts and then try those, too.

Thanks again

bruce

Dragon's Lair · Posted: Wed Feb 18, 2009 11:09 am Post subject:

That does look interesting, it would work well with this -

http://www.wiseguysynth.com/larry/ribbon/ribbon.htm

Damn! another project to put on the pile Shocked

Scott Stites · Posted: Thu Feb 19, 2009 7:29 am Post subject:

What an elegant design, Ian!

Maybe you did get a bum SP. I haven't tried the quarter test on mine, but I'd say I use pretty much the same amount of pressure I use on my computer touchpad. That's not to say it's as sensitive to the touch, but there's not any bearing down, either. Finger-weight is where I'd put it.

I liked your cotton gloves idea. I had a mental image of a man in tuxedo standing before a ribbon controller as the conductor raises his baton....

Cheerios,
Scott

frijitz · Posted: Thu Feb 19, 2009 8:01 pm Post subject:

Thanks, Scott. This is still pretty experimental. Time will tell if it is adequate, or if a bullet-proof design like yours is needed.

I started from the J. Simonton design, and was surprised to see that it didn't even come close to working. It took some head-scratching to realize that his wire contact has negligible capacitance vs a quite high value for the SP, and that that was a serious problem for acquiring the finger-up voltage. Switching over to the series-resistor sensing approach fixed most of the problem. (It might tighten up the timing in your system also.)

Adding the side-chain speedup blaster got the acquisition time down to a pretty small value. The main voltage still needs a bit of lag, which may bother some people. I think, though, that the glide time is comparable to the transition time between notes in acoustic instruments.

Actually, I'm enjoying using the controller in the Vangole mode, gating the VCO off on finger-up.

As far as the SP bouncing, I think I have a better understanding for what is happening. Since the blaster strobes the system hold for a finite time after the first finger-up signal is detected, there cannot be bouncing on the signal in the usual sense of a switch bouncing between on and off. But there is still some jumping around of the pitch if the finger is pulled off slowly. So I now think that remaining flakeyness is due to an intermittent change in contact resistance. This can be mostly avoided with clean fingering, I think (hope).

As far as the SP response, I guess it's not so much that it takes a huge force, more that it is hard to maintain the pressure while sliding around. I'm using the paper tape over the stripe now, and that helps a lot. Moving the fretboard back a bit and playing with flatter fingers is also helping. So maybe just need more practicing.

Ian

Minimalist · Joined: Nov 21, 2008 Posts: 9 Location: NZ

Thanks for posting this Ian

I'm thinking of having a go at building this and I've a few questions ( I'm a electronic novice so bear with me ) . What would you recommend for the PNP transistor next to A1, something like a 2n3906 ?. What value is the trimmer between A3 and the 33k resistor ? and is this a +15/-15v circuit ? . So thanks again for posting this and any help is appreciated.

Marcus

frijitz · Posted: Sun Mar 01, 2009 3:57 pm Post subject:

Mini:

Yes, 3906 is fine for the PNP -- it's not critical.

The trimmer can be 2k or 5k -- just a divider, so doesn't matter too much.

I use +/-12V supplies, but +/-15V should be fine.

All:

I've modified the circuit slightly and replaced the diagram in the first post.

I noticed that my 3.6V zeners don't seem to be very accurate. The 12k resistor in the I source may need to be selected to give the correct current. Just adjust it until there is about 5V at the top of the ribbon.

Also, I replaced the 5.6V zener at the input of U2 with a 4.7V unit, because of the 5V maximum input to the OTA.

There are several other changes in R and C values. The most important is the RC of the slewing filter, which I increased to 3 ms. You can experiment with this.

Very Happy

Ian

Minimalist · Joined: Nov 21, 2008 Posts: 9 Location: NZ

Hi Ian

thanks for the generously supplied clarification, one last amateur type question, I'm a little confused by the Extra arrow, (is that biasing?) on U1 & U2 the LM13700, could you add the leg numbers for the LM13700?

thanks again
marcus

frijitz · Posted: Tue Mar 03, 2009 10:24 am Post subject:

tooheys · Joined: Dec 04, 2009 Posts: 2 Location: Geelong, Australia

Hey Ian, this designs is impressive - & I'm really keen to try it out with a home-brew ribbon controller meets ondes martenot design.

My hotpot is hurtling to me in the post - now I'm trying to layout the parts on a pcb in EagleCad. Only the 5 ICs are creating havok when trying to arrange them in a single-sided way for DIY etching. So I was just wondering if u had a recommended layout position to minimise crazy amounts of jumper wires?

Or just a pic of your circuit in action

Also - for the caps - any non-polar caps work? - i.e. just ceramics or should some/all of them be film caps?

Cheers mate! I just thought I'd give this design ago rather than the more indepth appendage - as I'm really only after a simple ribbon part

now to work out how to make an optical button controller like the ondes

frijitz · Posted: Sun Oct 21, 2012 9:35 pm Post subject:

I'm resurecting this thread, since I have been working on my ribbon controller ideas again. Here is an improved version of the original circuit. The main change is that one of the comparators was redundant (A7), so I took it out. There are also some changes in component values.

Ian

FretterVer2a.gif

Description:

Filesize:

18.27 KB

Viewed:

1092 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Benjamin AM · Joined: Nov 04, 2010 Posts: 83 Location: Boise

If one was to use a 100k ribbon would he want to run a resistor in parallel ??25k??

frijitz · Posted: Wed Nov 07, 2012 5:01 pm Post subject:

Benjamin AM · Joined: Nov 04, 2010 Posts: 83 Location: Boise

I'm experimenting with ESD ribbon taken from foot ground straps. THey seem durable and were cheap on Ebay.

frijitz · Posted: Wed Nov 07, 2012 8:47 pm Post subject:

Benjamin AM · Joined: Nov 04, 2010 Posts: 83 Location: Boise

Hi Ian,
I finally got around to getting this circuit on a breadboard. I eventually ended up using a softpot instead of using ESD ribbon. I noticed that the CV output ranged from 0V to 4V. To change this I adjusted the value of the 120k resistor attached to A4's negative input. Doubling the value(changing it to 240k) dropped the output to a range of -2V to 2V as the schematics indicates. I was curious if this sounds right.
Also, I was curious if other ribbon controllers designs that use the softpot exhibit the phenomenon that has been deemed as the "squish factor", that is, pitch bending on finger pull off. Not that this is a problem, I found that the effect to be controllable with said roll off technique, and even at times(for the few hours I've spent playing the controller) desirable.
Anyhow, thank you so much for contributing this design, I look forward to the spending many hours playing this thing.
Edit: I forgot to mention that I used CA3080s in place of the LM13700. Seems to be no problem.