qHack Development

Findecanor · 07 Aug 2014, 01:02

I looked through my box of mouse parts and compared against some keyboard parts ...
My only scrollwheel that would fit in a keyboard row is 17 mm in diameter and has a rotary encoder that is 13 mm high from the PCB. The encoder fits just under the roof of a 2.75u DSA keycap when you press the key.
However, the wheel's axis would be in the way for the keycap's edge, and when not pressed it would be so sunken that I think it would be too awkward to use.

... But how about custom-making the wheel? I am thinking of a narrow bracket between the keys that holds up a scrollwheel at a comfortable height. The wheel could have spokes a little bit like the encoder wheel in an old ball-mouse, read optically by a sensor on the PCB below the wheel's axis. Because the wheel would be elevated, there could also be more room for microswitches in front of the wheel.
I am thinking that the wheel and the bracket could be laser-cut from black acrylic. A typical scrollwheel has only 24 steps per revolution, so the precision does not need to be very high. Maybe it would suffice to have 24 holes in three different hole sizes and a single light sensor connected to the AVR's analogue input.

webwit · 07 Aug 2014, 01:16

Talking about custom made, is it an option to 3d print the space bars? Maybe it is viable because these keys don't need labels. But I don't know about precision, shrink, durability and (matching) coloring when it comes to 3d printing.

Daniel Beardsmore · 07 Aug 2014, 01:21

Muirium wrote: ↑Exactly. Ratchet scrolling is a scourge. It's another axis, treat it that way, good and smooth.

Sure, now try pressing down on a detent-free wheel as a button. It's horrible. I would eventually have got used to finding the precise amount to move the wheel to not lurch past where I wanted to be, but middle-click was a write-off. In Windows you get to choose the amount to scroll when rotating the wheel but, guess what, loads of programs hard-code it to three lines instead of asking the OS for a user's preference :(

In the case of the qHACK, though, it appears to have its own middle button, and I presume nobody intends for the wheel to be pressed.

lowpoly · 07 Aug 2014, 01:26

Alps EC05E series:
http://www.alps.com/prod/info/E/HTML/En ... _list.html

Alps EC10E series:
http://www.alps.com/prod/info/E/HTML/En ... _list.html

Findecanor · 07 Aug 2014, 01:38

I may have got carried away a bit in my previous post...
Anyway, a scroll wheel still needs a bracket to hold its axis on the opposite side from the rotary encoder or else using the wheel would put stress on the encoder's solder pads.

How about increasing the width of the scrollwheel's space to a full 1 key width? A wheel, a rotary encoder and a bracket may fit inside that width.
Then make the Back/Space Bars be 2.25 and 2.75 with regular 2u stabilisers, because those keys are regularly available as Shift keys and they would give more space below the keycaps for the mouse switches. If someone wants Space Bar-profile (to fit whatever other profile he is using) then he could cut up a space bar and glue that on top the stems from another key.
I think I would cut up a space bar into three parts: left, right and middle, where the middle would have a cut-out for the wheel. I think it would look great to have the scrollwheel mechanism's housing flush with the Back/Space bars. Someone else might want to put the scrollwheel inside a one-unit key.

BTW, has anyone considered some other mouse switch than Cherry DG? There is the Cherry DH series, which would even fit between a PCB and a plate. The common square microswitches are also available with right-angle housing, but they would need some backing (from a custom-made bracket, maybe ...

) so as to not put strain on the solder pads.

7bit · 07 Aug 2014, 08:31

19.05-3.3*2 = 12.45 mm.
This should be enough to fit between 2 switches in a 1 unit gap!

Findecanor wrote: ↑...
How about increasing the width of the scrollwheel's space to a full 1 key width? A wheel, a rotary encoder and a bracket may fit inside that width.
Then make the Back/Space Bars be 2.25 and 2.75 with regular 2u stabilisers, because those keys are regularly available as Shift keys and they would give more space below the keycaps for the mouse switches. ...

That is already my plan.

Findecanor wrote: ↑BTW, has anyone considered some other mouse switch than Cherry DG? There is the Cherry DH series, which would even fit between a PCB and a plate. The common square microswitches are also available with right-angle housing, but they would need some backing (from a custom-made bracket, maybe ... ) so as to not put strain on the solder pads.

.43 units, instead of 0.672. But the switches will get into their way as well.

7bit · 07 Aug 2014, 09:33

Here is what I have with still the Cherry DG switches:

: qhack_mousebuttons.png (47.75 KiB) Viewed 6155 times

The upper-row of solderpads are for sideways mounting, so buttons are pushed from the front, and the lower-row is for upright mounting (pushed from top). The later can be moved up a little bit.

matt3o · 07 Aug 2014, 10:08

webwit wrote: ↑Talking about custom made, is it an option to 3d print the space bars? Maybe it is viable because these keys don't need labels. But I don't know about precision, shrink, durability and (matching) coloring when it comes to 3d printing.

3d printing is a bit expensive if you want high definition. We could make a prototype and then resin mold them, though

7bit · 07 Aug 2014, 19:48

Am I right when I assume that there are 3 wires (A, B and C)coming out of it which have to be connected to the controller?

Also: What does "detent torque" mean? Is that a bad thing, like ratchet scrolling?

matt3o · 07 Aug 2014, 20:40

don't quote me on this but C seems ground, A and B are the signal pins.

never heard of "detent torque" for encoders. if it is similar to stepper motors, might be the distance that is not registered from one dent to the next? So the smaller the better. It's nothing I would worry about since you don't need an incredibly high definition.

7bit · 07 Aug 2014, 20:50

So it needs to be connected to GND and 2 pins at the Teensy.

If we use these, then they should be mounted on the mount plate, because on the PCB itself they would be too far down.

So it would be sufficient for now to add 3 pads to solder the cables from that scroll-wheel thing down to the PCB.

Findecanor · 07 Aug 2014, 21:22

The rotary encoder is thin enough, but I don't think it is high enough to mount on the PCB. Imagine that with a SA space bar...

lowpoly · 07 Aug 2014, 21:24

7bit wrote: ↑Also: What does "detent torque" mean? Is that a bad thing, like ratchet scrolling?

Probably. Maybe you have to mod it and if you're lucky this can be done without opening it.

You also have to check height. With this encoder the middle of the axis is at 4.5mm over pcb top.

With the Miniguru, top of pcb to table ground is 4.7mm in that position. Maximum wheel radius would be 9.2mm then or 9mm without scratching the ground. You already have to cut the case here. Gives you a maximum height of 13.5mm over pcb. Spacebar height is about 20mm over pcb (Cherry cap). So you're 6,5mm below spacebar top.

18mm would be a good diameter for the wheel though. I think you need a 13mm encoder. The Alps ones I posted are too small.

lowpoly · 07 Aug 2014, 21:35

Findecanor wrote: ↑The common square microswitches are also available with right-angle housing, but they would need some backing (from a custom-made bracket, maybe ... ) so as to not put strain on the solder pads.

The ones I have here have 4 through hole solder points, 2x case and 2x switch. Pretty stable.

7bit · 07 Aug 2014, 21:58

As I said: I want to put it on the mount plate. If necessary, it can be lifted further up.

2x case means the wheel and 2x switch means there is the switch in the scroll-wheel?
I thought that would come extra:

See the little Omron switch at the left side of the wheel?

lowpoly · 07 Aug 2014, 22:01

No, Findecanor meant these (I guess):
http://www.aliexpress.com/item/6-x-6-x- ... 49625.html

As I said: I want to put it on the mount plate. If necessary, it can be lifted further up.

That could help. Or complicate things. Because the encoder is THT and the contacts seem to be part of the construction. Like the push button switch. How are you going to attach it to the plate? How do you connect the contacts to the pcb? You may need several small parts when a 13mm encoder could do. The wheel, axis and support are enough to manufacture IMO.

7bit · 07 Aug 2014, 22:04

I never get why people photograph their dirty fingers!

How much force do they require?

Also, I would need a datasheet to support these on the PCB.

lowpoly · 07 Aug 2014, 22:11

They require more force than the Omron or comparable micro switches. I can send you some no-names in two weeks.

If you need a datasheet you have to find them on Mouser, digikey, etc.

7bit · 07 Aug 2014, 22:14

The Omrons I have are 75 g, same as the Cherry DG switches.

If they are higher force, I don't see a reason to use them.

lowpoly · 07 Aug 2014, 22:40

Force will be higher than 75g.

7bit · 09 Aug 2014, 10:08

I've added 2 pads for the scroll-wheel encoder.

I've got 4 pads at the keyboard-Teensy left for indicator LEDs.

One goes to the normal Caps Lock position (left of ASDFG) and another one to the leftmost bottom row modifier. I plan to add an LED to the top right key and another to the bottom right.

Any other position would have to be hand-wired.

jacobolus · 09 Aug 2014, 10:59

lowpoly wrote: ↑Alps EC05E series:
http://www.alps.com/prod/info/E/HTML/En ... _list.html

Alps EC10E series:
http://www.alps.com/prod/info/E/HTML/En ... _list.html

Do these say how many counts they have per revolution? Is that what “number of pulse” means? 12 counts/revolution seems staggeringly low.

Personally, I want a control which spins smoothly and has at least 200 counts/revolution, or ideally more, so that I can use it for something like analog input. (And decide in my own logic if I want to use discrete steps.)

7bit · 09 Aug 2014, 22:48

: qHACK_PCB_003.png (1.95 MiB) Viewed 5948 times

7bit · 16 Aug 2014, 14:44

Version with only one controller:

Spoiler:

7bit · 15 Jan 2015, 22:14

The latest version:

: qHACK_PCB_001_front.png (740.59 KiB) Viewed 5698 times

: qHACK_PCB_001_back.png (624.33 KiB) Viewed 5698 times

scottc · 15 Jan 2015, 22:15

Yessss, I can't wait for the qHACK! It's going to be amazing!

7bit · 15 Jan 2015, 22:26

I'm quite sure this will work out of the box!

However, I will order a prototype series first!

Anybody interested in testing?

scottc · 15 Jan 2015, 22:36

Very!

Muirium · 15 Jan 2015, 22:40

I'll give it a shot, if you'll have me in again. The HyperMicro is nice, but I'm as awkward on a 40% as other people think they are on 60%s! Anyway, I've switches and diodes, and Teensies soon enough.

Scott! You need one? Pro Micro no go for this.

scottc · 15 Jan 2015, 23:03

Sorry, Mu - I've already got two Teensies burning a hole in my parts box! I need one for the HyperMicro, one for this, and at some point in my life I'll probably finish my custom board, but I'll likely use a Pro Micro on that if I ever get around to it.