Chris Radek's stuff

Home

TTT (TrueType tracer)

PalmShot

HTC PocketPutty

REALIZE for AutoCAD (G-Code export)

My CNC machinery

PCB Manufacturing

My stepper drivers

Sherline lathe conversion

Image to G-Code conversion

STL to DXF conversion

Historical IBM AT stuff

A few of my homemade clocks


Friends' sites:
Alex Bill Jeff John Michael

About this site

Coming soon?

Email the author:
chris@timeguy.com GPG Public Key

Welcome!

Recently I've found myself working on more and more projects/hacks that I want to share. I'm disorganized and my interests go here and there, which might mean this blog format will be a good match for me, since blogs are like that too. Follow the links at the left to visit the more persistent parts of the site; the content below will change.

Old programmer lore tells us that the last 20% of a project takes 80% of the time. This problem, combined with my tendency to stop working on a project as soon as it does what I want, has led me in the past to never share any of my unpolished work. I've decided to try to overcome that, not by changing my personality or work habits, but by sharing my 80%-finished projects anyway. I think they're useful as-is, and if you do too, that's great. If you don't think they're quite useful yet, please add on what you need.


Dominoes2 August 2010, 3:27 UTC


Today I made a set of double-six dominoes. I used my usual trick of cutting the entire outline full depth minus about .005 inches, and then whacking the parts with a dead-blow hammer to "dismount" them from the stock.
This leaves some mess at the bottom edge that usually has to be dealt with separately, but in this case I wanted them to have a nice chamfer anyway, so they are pleasant to handle.
I did this by setting up a work stop at the left of the milling vise to give a repeatable location.
Each domino was first mounted face down and then face up, and a chamfer cut around with a V mill.
With a little patience this gave me a nice set. This is before any hand-finishing work. Because they will be slid around when playing, polishing would be a waste of time. I will make some kind of brushed finish, and I may fill the dots with a drop of black paint.

[permalink]


spamass-milter improvement23 July 2010, 16:21 UTC

On my freebsd 8 system with sendmail running spamassassin through spamass-milter, spamassassin couldn't tell that clients had authenticated with SMTP AUTH=PLAIN via STARTTLS encryption. The problem was that the milter creates a fake Received header for spamassassin to parse, but it doesn't pass along the equivalent of "(authenticated bits=0)" or "with ESMTPSA" or similar, which allow spamassassin to know it should skip tests for dynamic IPs etc. (see this spamassassin faq).

This patch to spamass-milter (an updated version of this one) fixes it for me. This applies to the spamass-milter 0.3.1 that has already been patched by the freebsd 8 port.

Files attached to this page:

spamass-milter-smtpauth-gives-ALL_TRUSTED.patch2.3kB


[permalink]


Truetype Tracer 4.0 is released13 June 2010, 18:32 UTC

Truetype Tracer has been accumulating features for a while. It now generates biarc polylines in dxf mode, and splines/nurbs in gcode mode. This means it requires EMC 2.4.1 or newer.

Also included is the much-requested ability to fill in the letters. This is controlled with the -l flag. The default is to not fill, like previously.

Attached here is source for Truetype Tracer 4.0 and a ready-built deb package for Ubuntu 10.04 Lucid Lynx.

Files attached to this page:

truetype-tracer-4.0-screenshot-small.png26.5kB
truetype-tracer-4.0-screenshot.png39.0kB
truetype-tracer_4.0-1.dsc750 bytes
truetype-tracer_4.0-1.tar.gz16.6kB
truetype-tracer_4.0-1_i386.deb15.1kB


[permalink]


Banjo tuner24 April 2010, 2:39 UTC

The high tuner on my banjo split some time ago. An attempted super glue repair held for about five seconds, so I made a new one out of aluminum instead. I used a graver to turn the far (concave) side and a file to do the near (convex) side so it doesn't look too mechanical. I wasn't sure at first, but I think I like the industrial look of it.
[permalink]


Probing routines for EMC24 January 2010, 4:31 UTC

These are little programs I use to find hole center and diameter, and to locate top/front/back/left/right edges on a part.

Because my probe has some delay (debounce and wireless transmission) the final probe move is very slow. All the programs do fast moves first to roughly find the edge or center, and then a slow move finds the best possible result.

Beware my probe has a 6mm ball. Adjust as necessary for yours.

Files attached to this page:

P-back.ngc150 bytes
P-front.ngc150 bytes
P-hole.ngc1.0kB
P-left.ngc150 bytes
P-right.ngc150 bytes
P-top.ngc124 bytes


[permalink]


Older entries:

All older entries...


[atom feed]  
[æ]