The Hexboard

[ssp]

3-4 mm travel? your not going to need that amount of travel, those hex systems that you link to dont even have that amount of travel. the tact switches offered as examples are perfectly suitable to do the job at hand. if you had to press one of those switches in by 3-4 mm it would be difficult to get your finger down in there with the press.

By having a small tact switch the amount of pressure required and downward stroke is such that it makes usability feaseable. however if you had a 4mm downward stroke the spacing and depth required per button makes them hard to use in application.

you have got to factor in these tolerances, spacing of buttons, stroke of the button press, the face plate width, and most importantly your fingers as you would have to factor in the width of them when you press the button down not to touch any other button at the same time.

[Ganryu]

I looked into the Sonome which is identical to the Axis 64. Key travel is 2.5 mm as opposed to 0.25 mm of the switches provided here. The problem I can imagine is that at 0.25 mm the keys will feel more like drum pads. There will be no direct feedback to pressing them. I want to feel them moving with my fingers. 4mm is probably too much, but at the size they're at 2.5 mm sounds good.

[ssp]

well look at it like this.

by the time you have designed the hex caps, had them cnc cut, designed the face plate and had that cut with any lettering etc, then bought your kits and then the switches and caps for those etc your not going to be far from getting the one you linked.

the whole point of the switches we linked is to get the build done at a relatively small cost i would seriously price up what you are going to need by doind some serious reading of the forums here and also of ucapps.de site and then compare the price for the unit and then the self build and see which is more feasable.

im not trying to put you off but you really should work things out and then add to the thread.

the best advice anyone uf can give you is to read up , read some more and then price things up.

[Wilba]

My two cents...

Cheap tactile switches with 0.25mm travel are good for control surfaces and crap for "instruments".

If there was a good source of long travel switches that you can use for this kind of project, I would have found them already (or someone else found them). Either they don't exist in the "retail" market or they do exist and if you ever find them they'll cost US$5 or more each.

The closest you'll get to cheap 4mm travel switches are Cherry MX switches. They come in flavours like linear and soft tactile. The contact point is at 2mm travel, NOT the full travel length. Maybe not ideal but you could reduce the travel to 2mm if you were creative. Caps are readily available but not in hex. It would be relatively easy to take an existing cap and chop off the actuator mount and glue/wedge into something else... or go to a place that makes custom caps.

This idea is not new, and I'm doing a similar thing using Cherry MX switches, although I'm not wasting my time with hex caps.

You should read through THIS ENTIRE BLOG before you start your project.

http://musicscienceguy.vox.com/

You might decide to copy his prototype hex keys hacked into an existing keyboard's switch matrix, or just give up searching and buy the Axis.

Like the others here, I don't want to discourage you, but building a variant keyboard is hard stuff, and expensive, if you want something that doesn't feel like a cheap toy.

[clem!]

hi all,

don't know why I missed this thread. After some earlier searches I think the cheapest way to get usable hex-buttons would be to take wood and glue it on the caps of the buttons eventualy those of an old (min. 20 years) computer-keyboard(?)have a nice feeling.

I like also this one http://musicscienceguy.vox.com/library/post/how-to-build-a-jammer.html

scroll down this is velocity-sensitive.

Best regards,

clem!

Edited December 15, 2009 by clem!

[Ganryu]

Thanks for your replies. I'm looking into that blog right now, actually.

My price limit for a single switch is 1.5 euro per switch. Above that and the price for just the switches will be in excess of 300 euro.

I will make the case for the entire thing in wood, probably.

As for button caps, I suspect I will have to order them custom made and that will probably be quite pricey.

Edit: I found switches for 0,3 euro each.

http://se.mouser.com/ProductDetail/ALPS/SPPH430200/?qs=sGAEpiMZZMvxtGF7dlGNphyanz4eCaNfhA4sIi%2fCsXw%3d

They are SPDT and not SPST but that doesn't matter afaik. It's just that I get a bit more functionality than I actually will need. One curious thing is that it says operating force is 1 to 2 N. I had a friend calculate that to supposedly a 100g. That sounds a bit much for a switch of this size...

Edited December 15, 2009 by Ganryu

[ssp]

the switches i used for my humon, the blue illuminated ones have got quite a good bit of travel to them so they would do what you want perfectly, you would have to get the non illuminated ones and then get then with the small round cap so you can glue it inside as suggested. i will find the manufacterers details for you and post it tomorrow.

[Ganryu]

the switches i used for my humon, the blue illuminated ones have got quite a good bit of travel to them so they would do what you want perfectly, you would have to get the non illuminated ones and then get then with the small round cap so you can glue it inside as suggested. i will find the manufacterers details for you and post it tomorrow.

I'm really curious about those switches.

I scrapped the ones I mentioned before. They require too much pressure. I had some plans to attempt velocity sensitivity with SPDT switches but gave up. Too much work and too difficult and I'm not even sure it can be done reliably.

[Ganryu]

Any update on those switches? I really want to know.

[Ganryu]

Sorry for bumping, but I think you have missed my PM, ssp.

[ssp]

http://www.rjselectronics.com/?Switches_LED_Illuminated:Small

[Ganryu]

Thanks!

Was it these?

http://www.rjselectronics.com/?Switches_LED_Illuminated:Small:RJS_TC011

It's quite strange, but I can't manage to find any of the illuminated switches in non-illuminated form. All I find if I check tactile non-illuminated switches is a bunch of 0.25 mm travel switches which are all useless to me. It seems I might have to get them all illuminated.

How much did you have to pay for your switches, btw?

[ssp]

i dealt direct with them so i saved a bit, try calling them, also removing the led is very easy to do its just two pins going down the side of the switch

[Ganryu]

Thanks again for your help! I guess I'll try calling them, but I'm a bit iffy about international calls hehe.

As for the switch itself, I guess the one I linked (color blue) was the one you used?

[frailn]

I've been lurking on this thread because the project sounds really, really cool. I never learned to play the traditional piano-style keyboard. So, if I'm going to learn, using this layout might be fun and unique.

Thought I would jump in with a suggestion/idea/ramble since you are at the switch-purchasing stage for the keys. The first thing I thought when I saw this project was, "I wonder if there is a way to set this up so that key-press velocity is measured and sent to the synth?" Obviously, there are a lot of synths with traditional keyboard layouts that measure velocity. Surely, it can be done on he hex-board layout?

I've done some preliminary research, nothing big. It seems to me that you could achieve this by using a switch that can record the first depression of the key, then the release of the key, sending a signal both times. Then it would take a cpu to calculate the time lapse between the two signals and determine the velocity, translate that data into part of the midi data sent out of the device.

I think that is how synths that measure velocity work. Not to sure...

This might be possible with the right switch. The other determining factors would be:

1. Can the original Core or Core32 handle the calculation?

2. If one of these can, one would need to know how to develop the code.

Anyways, this is just a random idea. But, wanted to share it before you purchased your switches. I'm still knocking around the internet to try to find switches that would work this way. No luck yet.

EDIT:

Found The Terpstra. It can be done. Just need to find out how...

Edited January 17, 2010 by frailn

[frailn]

Found some Force-Sensing Resistors at Trossen Robotics

Okay, if I add any more info, I'm in danger of hijacking this thread. So, I'm stopping now.

[Ganryu]

Found some Force-Sensing Resistors at Trossen Robotics

Okay, if I add any more info, I'm in danger of hijacking this thread. So, I'm stopping now.

Don't worry about hijacking this thread :P

Anyway. Yes I've been thinking of velocity sensitivity, but it's not a high priority for me. I've been in contact with the original designer of the Axis 64 and he uses a special kind of custom ordered switches. As far as I understand it it's done with a switch that has a break period between position on and off. You then measure this period and calculate velocity.

[frailn]

Ganryu - not sure if you are located in North America, but if you are, check out the deal at Ponoko on 3mm black acrylic sheets.

Maybe this could work for your key/button building?

[Therezin]

It seems to me that you could achieve this by using a switch that can record the first depression of the key, then the release of the key, sending a signal both times.

This isn't how velocity sensitivity works. Keyboards have 2 switches per key, a known distance apart that are pressed one after the other. The time difference between the 2 switches being triggered is used to calculate the velocity, which is sent as the second byte of a MIDI Note-on message. When you release the key, the keyboard detects this and sends a MIDI Note-off.

[Ganryu]

Ganryu - not sure if you are located in North America, but if you are, check out the deal at Ponoko on 3mm black acrylic sheets.

Maybe this could work for your key/button building?

I'm not in the US. I'm in Sweden. Secondly I can't use that service anyway. I will need custom molded plastics, unless there is another option. The keys must not be flat. Flat keys will be useless to me as I won't be able to play it without looking. The keys must have a shape that allows me to know if my finger is in the center of the key.

This is also why the low-travel keys do not work. There will not be enough tactile feedback. I need to feel atleast some movement for the keys to be playable.

Thanks for your reccomendation anyway, though.

[Ganryu]

I have found switches that are perfect for my needs. They will go at 0.4 euro per piece, which is acceptable:

http://www.cherrycorp.com/english/switches/key/mx.htm

These are desktop keyboard switches and I can live without using custom switches for velocity sensitivity or anything like that.

Now I just need some way to make a suitable blueprint. Is there some free cad apps for doing this that are actually useable?

[Ganryu]

I have almost all the parts I need available now. I haven't ordered them, as I need to plan this thing out before I start working on anything.

My current stage is trying to figure out how to handle the coding elements of this project. I will need, ideally, 192 inputs. On the old PIC model it seems skill with assembler is necessary for such a project, but the STM32 seems to support C instead so I would prefer to work on that one. From what I have heard from other people the 128 input limit on the MIDIO128 is a result of the software itself, and adding more theoretical inputs is possible but it will slow down processing speed.

[Wilba]

I have found switches that are perfect for my needs. They will go at 0.4 euro per piece, which is acceptable:

http://www.cherrycorp.com/english/switches/key/mx.htm

These are desktop keyboard switches and I can live without using custom switches for velocity sensitivity or anything like that.

Now I just need some way to make a suitable blueprint. Is there some free cad apps for doing this that are actually useable?

I swear I suggested using Cherry MX switches before.... surely I did.... Oh yes, I did...

The closest you'll get to cheap 4mm travel switches are Cherry MX switches. They come in flavours like linear and soft tactile. The contact point is at 2mm travel, NOT the full travel length. Maybe not ideal but you could reduce the travel to 2mm if you were creative. Caps are readily available but not in hex. It would be relatively easy to take an existing cap and chop off the actuator mount and glue/wedge into something else... or go to a place that makes custom caps.

This idea is not new, and I'm doing a similar thing using Cherry MX switches, although I'm not wasting my time with hex caps.

You can add extra contacts to make them velocity sensitive, although this is fiddly and takes a long time. Here's my first "proof of concept". I've since improved the design a little.

[Wilba]

... and here's one I prepared earlier :wink:

... and even when I thought I was being original... someone else has used Cherry MX in a music keyboard:

[Ganryu]

Cool :o

Thanks for your reply! I've contacted a couple of companies about making the hexagonal keycaps, but looking at the Janko it seems I should be able to make my own in similar style. The only thing is I don't want them to be completely flat. That will make playing too difficult.

As for the cherry mx switches I don't know but yes I think you made a post in this thread before and for some reason I can't fathom I dismissed the suggestion. They look perfectly fine to me, now that I see them. The only problem is that they are a little too big, but I think I can live with that.

I've been thinking of using the Cherry ML switches as an alternative. They are definitely smaller and would get me below 20mm distance between the keys.