Computer scroll wheel, bearings version

Step 5 Finished, future improvements.

All done, try it out, I find it particually useful when navigating in a 3d enviroment such as CAD programs or Google Earth.

It's also pretty good for programming, scrolling through long files quickly and accuratly, or just giving it a spin to get to the top/bottom of the file.

Future improvements:

A shaped case
Though it's not as uncomfortable as it looks, curved edges would improve it.

Buttons
If I encounter a program that requires a mouse click and scoll for a frequantly used function, I may wire one of the buttons to the side of the case
A more complex improvement would be to have a button click when the wheel is pressed down.

Smaller footprint
The reason for the big case is the size of the circuit board, if I did this again I'd choose a mouse with a smaller board, or chop the board up and solder directly to the copper traces by the chip.

Optical encoder
It'd be possible to use an optical rather than a rotary encoder, it'd be smoother (The rotary encoder has clicks, or notches) and give much less friction, so would spin for longer.
The downside would be the additional complexity oh the build. (And trying to attatch that perferoated disc to the wheel shaft, could be awkward..)

Any more suggestions for improvements?

Comments

[destructomonkey]

Nice thought with the SS rotary encoder. Much easier and more robust in this application. Next step is simply to dispense with the rest of the mouse guts entirely though. Much nicer Solid State encoders are available to substitute from as low as $1 or $2 brand new, and they can be scavenged from most newer car stereos or whatever (if it spins forever and doesn't hit a fixed stop, its likely an encoder). Other nice thing is that those usually have detent steps which is a very nice, almost necessary feature for this application. An encoder will have the two signal pins for the quadrature, and one for power in. Its that simple. No entire mouse board and other guts are even actually required unless you are looking to use them simply for the USB translation. There are other ways to do that too, or just run the signal pins directly to an actual serial port and skip the middleman entirely. Then your parts count is literally down to one shaft encoder, three wires, and one box to put it in.

[destructomonkey]

http://www.mcmanis.com/chuck/robotics/projects/lab-x3/quadratrak.html

here is a good link with schematic that will simplify things for you (i hope!!)

[kool1zero]

would you mind to throw together a schematic? i would greatly appreciate it.

your idea sounds so much simpler; no point in having all the mouse hardware if you wont be doing actual clicking with it. although a click option (im imagining something similar to clicking a joystick on a xbox controller) would be very nice too. that would be a lot more scalable (maybe a few of these on a USB hub in a larger case for multiple wheels?) though im not sure what i would use them all for, it would stil be badass looking if done right (a bank of scroll wheels???)

[destructomonkey]

I don't really have a schematic per se - I simply ran the three wires directly to a parallel port (two signal pins and one tapping the ports power pin to drive it). I took care of actually parsing the two channels changing states - and interpreting that into direction - in software. You can write that into custom code yourself or just use any one of a number of programs to do it, many free. Its pretty common to find to control things like jog wheels on CNC machines. Actually any machine control software that queries a parallel port can be configured to do it.

Look for anything meant to interpret "gray encoding" or "quadrature encoding". If you just google "gray encoding" you will get how it works really fast! Its really simple.
Before anyone gripes parallel ports are old and slow, two things;
1) They were never meant to be a data transfer method, that was hacked on later and badly. The initial reason for them to exist was to interface with and control physical machinery in real-time. At this they excel, as that MUST be parallel in nature. Things like USB meant for fast data transfer are serial, and therefore there are timing and sequence errors.
2) Parallel cards are available to this day, for those machines that don't have it, for like $10.
Any equipment hacker who wishes to interface directly with hardware needs one. Even if its an ancient design, the fact is it was designed for just this exact purpose - and any new port designed for this would have to do the exact same damn thing inteface directly and avoid virtualization layers. Just get one already. There is a reason they are still standard in the CNC industry.

[eggplanthunter]

I'm having a rather difficult time finding a mouse w/ a solid state encoder, and google isn't really helping. You mentioned a place where you can get rotary encoders, could you point me and anyone else having trouble in the right direction? Also, how would you connect the encoder to the computer? Would you have to use a serial port or could you solder the new encoder in the place of the old one? I'm a novice in DIY electronics, so any advice would be appreciated.

[destructomonkey]

Mouser Electronics has a lot of them, try searching their site or starting around page 1500 of their catalog. The rest of the electronic supply places should carry them too. You could just substitute a better quality, encoder for the original mouse one, just solder it to the same connections and use the mouse board as-is for its interface chipset. Probably the easiest if you are a novice. I use an interface that expects to be speaking to a number of quad encoders directly, (I use four) so I can just run them into a serial or parallel. Gives me a ton more control over their behavior, what they each control, and aspects like defining step resolution, etc. but it is more difficult in that it requires coding your own (if pretty simple) driver, or only using software that already understands quadrature input.

[eggplanthunter]

Let's just say programing isn't my forte, so that route is out for me... If I were to buy an encoder and wanted to attach it to the existing board, would I need a mouse that had a SS encoder to start with, or could I throw it on where the optical encoder's terminals were? Also, what kind of encoder do the mice use? (ie. binary 2-bit greyscale or greycode or quadrature, do the number of channels matter, do I need it in absolute or incremental) Well anyways, thanks for the helpful reply.

[sibrow]

This mouse worked well for me. http://www.amazon.co.uk/A4-Tech-Advanced-Double-Click-Function/dp/B000ICIN0O/ref=sr_1_4?ie=UTF8&s=electronics&qid=1220541422&sr=8-4 :the second scroll wheel is on a separate circuit board connected by wires, so you don't need to do any soldering! I also used an old floppy drive as a case, and used the motor + bearings as the wheel. Hope that helps.

[eggplanthunter]

Thanks for the advice. i did however find a mouse that worked for me, so I'm happy now.

I have pictures of it, and a link on the original version of this ible.

[w477s]

ALPS! WHERE ARE YOU LIVING I WANNA VISIT U! XD

joke ;)

[nubie]

If you wanted remove the clicks/notches you can bend the spring on the encoder. It is visible in your photo as the bit of metal with a U bend in it. I just did this with a VCR head (good bearings, very heavy, but also very tall), I snipped off the end of the plastic wheel that goes in the encoder, filed it flat, and used superglue to attach it. There is no need to glue down the encoder or position it in the very middle, I am using ribbon cable to attach and it will hold the encoder in only the axis of rotation, allowing movement to center to the axle. I just anchor the wires to the box. My ultimate goal would be designing an optical encoder so that I could grind (sadly no lathe access) off the axle to the bottom of the VCR head and weighting the bottom half from the inside with lead. One option is reflective sensing with tape or paint on the aluminum surface itself, using bits from a ball mouse. If you are interested in 3D or drafting I recommend getting yourself a SpaceOrb 360, 6 DOF optical ball! (Rotation and movement along all 3 axis) The "engineer/CAD" models are expensive, but the driver from the website works with the gaming controller, replacement heads are available from a Playstation version called the "Ascii Sphere", you can sometimes find them on stock clearance sales, I got 4 for $20 (75% of that was the shipping)

[osgeld]

i want to do this with a VCR head, but instead of a basic scroll function some improved software could make this into a very useful A/V editing attachment

[42ndOddity]

I'm thinking of ripping it all apart actually and redoing it, so I'll consider those improvements next time. I'd really like to get it to work so the knob is also the 'middle click' button. But I suspect that'll mean lots of fiddly tuning and tweaking with spring tensions. I could actually cut most of the circuitboard out, there is basically just a chip and a couple of transistors clustered close together. So my next version may be soldered directly onto the circuitboard traces. I want to keep it USB though, just for compatibility, I can currently plug it into my XP or Linux machines and they accept it perfectly. The rotary encoder actually causes a lot of friction, much to my annoyance, which may be solved by removing the spring, I'll give that a go. Thanks for the feedback!

[osgeld]

remove the spring and use some of the radio shack "precision lubricator" which is basically mineral oil with Teflon mixed in, or silicon lubricant found at most hardware stores (i get mine at ace, never wears off awesome stuff in a spray can) i mistakenly did that to my first "Microsoft optical mouse" way back when, and 1 slight movement of the finger would send the really lightweight plastic mouse wheel flying, if you rigged it up with some sort of tensioning mechanism you could replicate what i like to call the "1970's home stereo effect" where it glides super smooth and free, but also lets you have precision control

[kerowhack]

Nice machining work on the knob. I have been considering building one of these on a joystick, and maybe with a button or two on top. I know that would be starting to move out of the realm of cheap and easy, and would probably require a driver, but it sure would be nice for Google Earth and AutoCAD to control all the view and zoom with one device.