DIY - Using a rotary Pulse Switch instead of the S1/S2 switches.

Started by GuitaRasmus, November 09, 2015, 12:52:49 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

GuitaRasmus

Duh... Sometimes I'm not too impressed with myself.  ;D

I blame the lack of sleep. Thanks, I'll try that.

CodeSmart

Hm, a bit out of topic. There are plenty of rotary encoders with a push-down switch on the shaft. Imagine what a small microcontroller with 3 digital inputs (encoder + push-down switch) and 2 digital outputs (S1/S2) plus one analog 5V DAC output (for GK VOL) could do.

One possible software implementation could be: GK VOL if just rotating the knob or S1/S2 if knob is pushed down while turning the knob. S1/S2 pulse width controlled by software. Drawback is that it would be a GK volume control without physical endpoints.

Also it would be a challenging thing to generate power the microcontroller derived from GK 7V without inducing audible digital noise.
But I got more gear than I need...and I like it!

GuitaRasmus

Quote from: CodeSmart on November 12, 2015, 03:26:57 PM
Hm, a bit out of topic. There are plenty of rotary encoders with a push-down switch on the shaft. Imagine what a small microcontroller with 3 digital inputs (encoder + push-down switch) and 2 digital outputs (S1/S2) plus one analog 5V DAC output (for GK VOL) could do.

One possible software implementation could be: GK VOL if just rotating the knob or S1/S2 if knob is pushed down while turning the knob. S1/S2 pulse width controlled by software. Drawback is that it would be a GK volume control without physical endpoints.

Also it would be a challenging thing to generate power the microcontroller derived from GK 7V without inducing audible digital noise.

Cool ideas - however I'd be happy if could just get this one function to work!  :-\

gumtown

What happens if you have just one "S" switch connected (S1 connected to rotary switch and S2 disconnected),
Rotating the knob in one direction, the GR-55 should change/increment, and not do anything when rotated the opposite direction.

Free "GR-55 FloorBoard" editor software from https://sourceforge.net/projects/grfloorboard/

GuitaRasmus

Quote from: Elantric on November 12, 2015, 02:21:26 PM
1 ) Rotate Clockwise - observe that only one set of contacts make closure - other set should remain open circuit 

2 ) Rotate CounterClockwise - observe that  the opposite set of contacts from step #1 make closure, while the Step #1 set of contacts should remain open circuit

I've just measured it very thoroughly, and the opposite set of lugs has absolutely no response, when I turn in the other direction - they are completely independent of each other, and responds like the contact/switch should. Only one set of lugs make momentary contact/sends a pulse, when it's turned in one direction. It SHOULD work.

I seem to recall something about a capacitor somewhere around the switches, that's there to eliminate contact flutter, and that perhaps it should be changed for a smaller value - however I cant seem to find any capacitors around the wires - any ideas?

GuitarBuilder

I really like the mechanical rotary switch idea and I'll be incorporating it into my next build.  The old-fashioned volume control is fine for me.  However, I think it would be nice to see which patch I'm selecting right on the guitar - might be a job for CodeSmart!
"There's no-one left alive, it must be a draw"  Peter Gabriel 1973

sixeight

QuoteI seem to recall something about a capacitor somewhere around the switches, that's there to eliminate contact flutter, and that perhaps it should be changed for a smaller value - however I cant seem to find any capacitors around the wires - any ideas?

You would have to add a capacitor to eliminate contact flutter. Maybe connecting an RC circuit through a diode can make the pulse width a little wider.

Do you have an oscilloscope? If we know what the pulse width of the switch is, perhaps we can work out a solution.

GuitaRasmus

Quote from: sixeight on November 13, 2015, 06:57:46 AM
You would have to add a capacitor to eliminate contact flutter. Maybe connecting an RC circuit through a diode can make the pulse width a little wider.

Do you have an oscilloscope? If we know what the pulse width of the switch is, perhaps we can work out a solution.

Now I'm beginning to think I'm in over my head - about the RC circuit.  ;D

My father is pretty electronics savvy, and he has an oscilloscope - I'm visiting my parents this weekend, so I'll try to get him to measure the pulse width.

Thanks again for the help - this is an AMAZINGLY geeky forum (meant in the best sense of the word).  :)

Edit: btw, if there are no capacitors there at the moment, wouldn't adding one only serve to lessen the response? At the moment, the problem is that the Gr-55 only responds to some of the switch movements - but perhaps this is the problem, that the contact flutter of the new switch, contrary to the classic momentary 3 way switch, is wreaking havoc on the gr-55 receptors, and therefore not responding?

Edit #2: In the service manual diagram for the gk3, theres a 1k resistor in series with each switch, as well as a 0.01 capacitor in parallel - is this the RC circuit you're referring to? I'm just wondering, because I've only used momentary toggle switches before, without any resistors or capacitors, and they work fine. Are these integrated on the gk kit circuit board? They are nowhere to be found on any of the wiring.

sixeight

QuoteEdit #2: In the service manual diagram for the gk3, theres a 1k resistor in series with each switch, as well as a 0.01 capacitor in parallel - is this the RC circuit you're referring to?

Yes, something like that should help. If you could try to add a capacitor and a resistor on both switches and see if it works? Well you could start with only one switch, just for trying. It may need some trial and error to find the right values of R and C. Start with 10 nF and 1k. Then increase the capacitor in steps of 10 (100 nF, 1 uF, etc). That is what I would try.

GuitaRasmus

Well, I tried.  :-\

I tried using some different values of capacitors, like the ones you mention, and it still doesn't work properly. One side of the switch (clockwise), works pretty well, as it changes up the banks, with only a few glitches. However, it doesn't respond very well to quick switching. The banks simply doesn't respond.

Counterclockwise barely works at all - it responds to perhaps every other/third click of the rotary switch, and sometimes it needs even more.

My dad will have a look at the switch, and see if his oscilloscope is able to "sample" the pulse - since it's a pretty old oscilloscope (he built it himself in the 70's) it hasn't got all the modern features.

I've given up for the moment, and installed my standard 3 way momentary switch, which works flawlessly. Go figure. I just don't get why it doesn't work, but I need the guitar for the gigs in the coming weeks, so I needed a solution.

Much thanks to everyone who helped - you guys are great. :) If you get any new ideas, please post them. :)

sixeight

To get a clear pulse, you could use a 555 timer in monostable mode:

http://www.electronics-tutorials.ws/waveforms/555_timer.html

You could then control an NPN transistor with pin 3 of the 555 timer through a resistor.
But you probably are still limited by the speed of the bounce handler of your Boss/Roland guitar processor.

ADP

Well, I've received my order and installed one. Unfortunately, I'm getting the same results as GuitaRasmus. Turning the switch clockwise, I get good response although it does skip from time to time. In the other direction, the opposite is true: I get only occasional patch changes. Counting the twenty clicks, it seems (without being too meticulous and documenting the results) that it skips changes on the same clicks, so I wonder if there is an issue with the contacts.

Also, clockwise, it's possible to change patch by simply putting pressure on the switch to start the rotation and rocking back and forth. Anti-clockwise, there is sufficient resistance to make it necessary to rotate the switch through the click. Even then, it only works from time to time.

It's a great idea, but it looks like more experimentation is needed to get it to work.

Another thing to bear in mind is that there are only enough threads for this particular switch to work with a scratchplate (pickguard) and not through the top of a guitar so I'm going to use a momentary switch for my conversion and try to make it look like a pot by putting a knob on it.

Edit: It seems that you can rock the switch in both directions with perfect patch changes every time. Because I was holding the switch outside the guitar, it was easier to rock clockwise the first time I tried.

GuitarBuilder

It sounds like the RC/diode circuit suggestion is required here, primarily because it's passive and doesn't require additional power.

Start with page 12 in the attached debouncing guide.
"There's no-one left alive, it must be a draw"  Peter Gabriel 1973

ADP

Quote from: GuitarBuilder on November 19, 2015, 02:20:51 PM
It sounds like the RC/diode circuit suggestion is required here, primarily because it's passive and doesn't require additional power.

Start with page 12 in the attached debouncing guide.

Thanks for that. I think I get the idea, but my understanding of electronics is pretty basic. I was pretty much hoping that I could just wire the switch in place and have it work. ???

GuitarBuilder

Does anyone have a schematic for a GR-55 or similar?  It would be interesting to see how Roland conditions the S1/S2 signal in the box.
"There's no-one left alive, it must be a draw"  Peter Gabriel 1973

admin

You can see the existing S1 & S2 switch debounce in both the GK-2A, GK-3 and US-20 schematics














GuitarBuilder

Thanks!  These I had already - I was interested in what happens at the synth end.
"There's no-one left alive, it must be a draw"  Peter Gabriel 1973

admin

QuoteI was interested in what happens at the synth end.

Another R/C network on each S1 & S2 line for debounce before it hits the driver transistor before the CPU


GuitarBuilder

"There's no-one left alive, it must be a draw"  Peter Gabriel 1973

Moja

Did you finally manage to get this rotary switch to work ?


Headless68

I'm hopeful it works - this is what I want on my current build project