Lately I've been elbows-deep in some broken Playstation 3's and found myself wanting to test their cooling fans. These have a three-wire header with leads colored brown, black, and gray; you may be tempted to conclude that this is a brushed DC fan with a tachometer lead, but you'd be wrong. These are brushless fans, and the third wire is a PWM signal that you supply to control the speed of the fan. The two PS3s (both “fat” style) I've opened recently have compatible fans from separate manufacturers; one is a Nidec G14T12BS2AF-56J14 and the other is a Delta Electronics KFB-1412H.

Nidec and Delta models

It's not trivially easy to find datasheets for these fans, but no matter. If you just want to test them or need a good centrifugal blower for one of your own projects, do the following:

• Apply 12 volts across the brown and black leads; +12 V on brown with return on black. The fan will probably jump a little but it won't start spinning.
• Drive the gray lead with a TTL-level pulse train at 25 kHz from a signal generator or 555 timer circuit or microcontroller or whatever.
• Control the duty cycle of this pulse train to adjust the speed.

That's it!

Lacking an insert plate for the 10-inch Delta bandsaw (model 28-195) at work, I made a new one out of ABS plastic using our 3D printer. Here's the result:

Rough-but-ready output from the Solidoodle

The geometry is two discs stacked on top of each other with a little bit of a cutout at the edge behind the blade, which helps prevent the insert from twisting too much. I've modeled a slot a bit larger than the kerf of the blade and included it in the printed part but you could certainly leave that off and cut the slot into the insert using the bandsaw itself, making what's called a “zero-clearance” insert.

I'm starting another AVR project using my by-now-venerable STK500 and the latest 6.1 version of Atmel's Studio software. Although I'm sure Atmel would like all developers everywhere to buy STK600s, it's perfectly possible to use the older development kit with the newest software, although this combination is not as well documented as I'd like; hence I'm using this post to collect a few tips distilled from recent experience.

Regular followers of this blog or of my flickr account perhaps noticed an enormous improvment in the quality of my macro photos about a year ago; at the time I put together a primitive light tent using foamcore and tissue paper. Although the photographic results of this project were all I could hope for, the tent itself was extremely fragile, awkward to set up, and likely to collapse while shooting. So I took what I learned from my first effort and built an improved tent with a PVC-pipe frame and a bedsheet light diffuser.

I've added two new types of components to my toolkit of 3D printed machine parts: a snap-fit pivot or hinge, and a dovetail suitable for joining two parts together at a right angle.