Trackpoint for DIY keyboards?

[Sigmoid]

Hey,

I really enjoy hacking and building things, and something came up on the keyboards forum recently in an ergonomics-centered discussion - that these days you can't really call any setup ergonomic unless there is a pointing device right in the middle of the alphanumeric cluster.

Thanks to IBM's Trackpoint, that's a given on Lenovo notebooks, and on some Unicomp keyboards... It would be nice if one could be integrated into a custom-built 60% keyboard as well. So my question is, has anyone built a trackpoint-like pointing device into a DIY keyboard before, and are there known, non-ebay sources of good quality micro-joystick assemblies, either actual Trackpoints or acceptable alternatives for use in such an application?

[Ray]

http://deskthority.net/wiki/MX13_SpaceSaver_Keyboard is probably the most clean I've seen so far.

[bhtooefr]

Sprintek also sells a TrackPoint-compatible module: http://sprintek.com/products/SK8702.aspx

I don't know if it has negative inertia, though.

[Findecanor]

I have seen several other projects where trackpoint modules have been put into keyboards, but most trackpoints have been ripped from existing keyboards.
The Trackpoint modules I have read the datasheets for have been talking the PS/2 mouse protocol. People have been using Teensys (AVR ATmega32u4) for conversion from PS/2 keyboard to USB ... but while it would sure be possible to parse PS/2 mouse it might be even better if it could be read at a lower level to make the firmware code simpler.

[kbdfr]

http://deskthority.net/photos-f62/mod-a ... t4214.html
and
http://deskthority.net/f-o-r-s-a-l-e-f5 ... nt#p163082

[derzemel]

There is also the 60% Tex Yoda keyboard (with a metal case) that has been showing up on randomly on Massdrop.

It can be bought as a kit or already built.

The trackpoint modules were also available to be separately bought, but is seems that they are a bit brittle (I saw a post on Reddit with a broken one)

[JBert]

Lowpoly is still working on and off to complete his Miniguru keyboard kit, but I'm afraid he won't disclose the vendor just like that (I believe he had some trouble before).

[bhtooefr]

but while it would sure be possible to parse PS/2 mouse it might be even better if it could be read at a lower level to make the firmware code simpler.

One thing worth noting is that the TrackPoint code is actually rather complex with all of the transfer functions involved, as I understand. It'd be nice to see native USB pointing stick code, but I'm under the understanding that all of the existing USB TrackPoints - even the ones sold by IBM/Lenovo - are being converted from PS/2 internally (they're typically a complete ThinkPad keyboard with PS/2 TrackPoint, and then a controller that natively scans the keyboard, and converts the TrackPoint protocol to USB), so that they don't have to rewrite firmware from 1999 or so.

The other thing is that for a third-party to reimplement that code, they needed to license it from IBM until recently - the negative inertia transfer function patent only expired in October of last year. The PS/2 TrackPoint microcontrollers, on the other hand, included a patent license for the code on them.

[Sigmoid]

I was under the impression that the negative inertia stuff was implemented at a higher level. Good to know. The sprintek ones seem the most accessible, and it does mention "advanced motion control". Anyone happened to use it maybe?

[bhtooefr]

I believe they're actually what TEX is using in the Yoda.

In any case, the negative inertia behavior actually occurs even without a TrackPoint driver loaded, when it's merely using standard PS/2 mouse reports. (And, in fact, the TrackPoint III only speaks standard PS/2 reports, and has negative inertia. The custom protocol was really only needed for configuration, where Z-tap (which was new on the TrackPoint IV) and transfer function sensitivity are configurable on the TrackPoint IV.)

Scrolling is something handled on the host, though, not in the TrackPoint. The TrackPoint is just sending a standard middle button report, and the Windows drivers handle middle button down as a request to scroll by default. Linux support tends to treat middle button down and back up without movement as a click, and a middle drag as scrolling without a click.

[Findecanor]

Scrolling is something handled on the host, though, not in the TrackPoint. The TrackPoint is just sending a standard middle button report, and the Windows drivers handle middle button down as a request to scroll by default. Linux support tends to treat middle button down and back up without movement as a click, and a middle drag as scrolling without a click.

Whoa whoa what? That has been in trackpoint drivers all along? I have wanted that for mice!

[jip]

Not sure if I ought to make a new thread. I am working on a small project and was wondering if anyone's got any documentation on the dimensions of the sticks and the rubber caps? Been searching for quite some time to no avail.

I could only assume that different manufacturers' got their own designs; that Dell rubber caps won't necessarily fit IBM/Lenovo pointing sticks; so I guess my question then is, does anyone know which brand I should choose in terms of easily acquired rubber caps, and what are the dimensions of those rubber caps, especially the inner dimensions?