If computers were like the Catholic religion...

[giza]

Per nostra Pentium Quatro cum 2 gigahertzum e bus de cuatrocentum megahertzum...
Deo Gratie...
Per nostros Quinientum Doce megabaitum de RAMus...
Deo Gratie...
Per nostra GeForsum Duo Mu Omega cum centum ventiocho megabytum de memoria Delta Delta Rho...
Deo Gratie...
E por nostro casum de aluminum con sweetum modus e infinitum blinkenlightenus...
Amen

Comments

Hmm.

[cjthomas.livejournal.com]

Pray for an upgrade, if it's a GF2MX };>.

On the subject of inifinte blinkenlights, it occurs to me that it'd be very easy to make a blinkenlight panel peripheral. Does anyone sell this yet, or have I struck virgin territory?

(I'm thinking LED array and a microcontroller, attached to a parallel port.)

-Deuce

[silverscalenaga.livejournal.com]

AMEN! *giggle*

Re: Hmm.

[shockwave77598.livejournal.com]

64 LEDs in an 8x8 grid. One 8bit port of the MCU drives a high side driver and one port drives a low side driver. Have a memory map of 8 bytes in an array in Ram. At 240 hz, set the low side driver and output the corresponding byte from the array to the high side driver. This will generate a 30hz refreshed display.

At the end of every scan (30hz) Xor one randomly chosen bit in the 8byte array. Very simple, cheap, and attractive. You can also do more clever things with more complex code.

And yes, I could build this. Prototyping isn't cheap however.

[aericwolf.livejournal.com]

I think it's very sad that I was able to read that... on both the geeky side, and the religious side.

Re: Hmm.

[cjthomas.livejournal.com]

64 LEDs in an 8x8 grid. One 8bit port of the MCU drives a high side driver and one port drives a low side driver. Have a memory map of 8 bytes in an array in Ram. At 240 hz, set the low side driver and output the corresponding byte from the array to the high side driver. This will generate a 30hz refreshed display.

I was thinking just an Atmel "tiny" microcontroller driving two shift registers for the row and column drivers. Serial loading of the shift registers is fast enough to be imperceptible even if you use the internal clock on the microcontroller. But there are many ways of building this circuit.

Unfortunately, being a lazy bastard, I won't be in a position to work on electronic widgetry until fall, if then.

Re: Hmm.

[shockwave77598.livejournal.com]

I've never used that part. I have a few 68hc912b32 flash-based parts though and they have tremendous IO for a little 80 pin part. But yeah, there are serial input high and low side drivers as well. (speed up trick - only shift the column driver once to drive the next column rather than loading a whole new pattern. In a small array like this it's not an issue - when the array reachs 256 columns though, you start to look for ways to speed things up. :)

Re: Hmm.

[cjthomas.livejournal.com]

I've never used that part. I have a few 68hc912b32 flash-based parts though and they have tremendous IO for a little 80 pin part. But yeah, there are serial input high and low side drivers as well.

This is a little 8-pin part :). I wanted something I could solder on vector board, as opposed to something I'd need to fabricate a printed circuit board for. Very small amount of ROM, no RAM (just registers). Still surprisingly handy. Designed to be used (in a pinch) with no auxiliary components at all.

Drivers I'm considering are your standard 7000-series shift registers. They should be able to source and sink enough current on each pin for an acceptable display. If not, I'd add a transistor for sinking on each row.

All of this is academic, of course, until my thesis is out of the way (doing a good job of eating my life so far).

[shesaspy2.livejournal.com]

yeah! whatever that says..