19 January 2007

Last weeks problem with the rope light, Unfortunately the 24 bit route to a solution did not work, as I have since measured the number of bits to be more like 1937, but by sending one and one only at a time, a pattern emerges. The pattern seems to be, each LED is allocated 3 bits, red, green & blue and for 250 LEDs as in the sample rope still equates to 750 bits, all sent serially with a clock, then an enable pulse at the end. This format at least appeared to be repeatable and reliable, however the customer really wanted DMX control of the rope.

So a new problem.

After some searching on the WEB I found the full DMX protocol and with the loan of a simple DMX controller, I have now managed to read eight dimmer levels into registers on a PIC, and a rock bottom PIC at that (16C55), no serial in hardware, but at 16Mhz it appears to reliably read these commands.

Fun as it is I must now concentrate on a very real job of the LEDLIFT26 and the problems of giving it a CE mark, though I think this is more an exercise in paperwork, than any real technology.

Do I have a cup of tea, or a brew ???

This is the question !

After 20 mS of serious thinking I went for the brew yet again.

It maybe detrimental to my health but is easier to drink and does not require filling the kettle or tea bags or washing up or even any leccy etc.

It is also very filling at 1000 odd calories a tin.

It therefore negates the need to eat for many hours, however it does have a down side of rendering one completely unable to think, talk coherently or do anything, even walk in a straight line.
But these are problems we learn to live with !!!

In the left corner a setup to read DMX stuff, using the Rice17 PIC in circuit emulator.

And on the right, a scope trace of 7 channels coming in. Top trace is all DMX channels at full brightness, bottom is an output of my sample points of the data