New project: Sawyer...
 
Notifications
Clear all

New project: Sawyer Star (LED array)

54 Posts
13 Users
0 Reactions
33.2 K Views
dmcole
Posts: 2196
Topic starter
(@dmcole)
Noble Member
Joined: 19 years ago
[#10042]

There is a 24-inch, clear-acrylic sheet, five-point star with 60 LEDs on the wall over my desk; it's been there for a year (and lived in the garage for six months before that). After working on this project for much of 2008, the goal for 2009 was to be able to drive this star (and many more like it) with DMX 512 via Vixen. It -- and its partners -- would go in windows to supplement the exterior lights.

This is a project that I'm adapting from one by David Thorpe at http://www.unusualelectronics.co.uk/projects/star/ . Thorpe provides the schematics, much theory and a construction manual, but he doesn't share his source code.

I have always believed that there would be some interest in a DMX-controlled, multi-LED device similar to Thorpe's. While this is specifically a star, the circuit could be used for a variety of lighting fixtures, including snowflakes and wreathes. It could also be a mega-tree topper as well as a window fixture.

The general electronics theory is that it is a matrix array, with a ULN2803 Darlington chip controlling the sink and the PIC 18f2620 microcontroller handling the source.

I spent much of 2008 trying to write code for the 18f2620 that would substitute Thorpe's rotating routines with a DMX receiver system that would have 16 channels. The user, via Vixen, would bring high two channels in order to control any single LED.

But my ambition outpaces my skills. I kept hitting technical roadblocks, mostly from my inability to write compact, efficient code (as well as understanding the compact, efficient code of others) coupled with a general lack of electronics skills. And while I can muddle about in C, I look at assembler code and my eyes go cross.

After that setback, I got a bit further by building Henne's DMX transceiver with the LED array option -- http://www.hoelscher-hi.de/hendrik/english/matrix.htm . Henne uses a similar circuit design as Thorpe, which led me to believe I was on the right track.

Unfortunately, after building the circuit on a breadboard, it didn't perform as expected. Individual LEDs weren't controlled with it -- when two channels were set to high, 16 LEDs lit up. Adding channels increased the brightness of those 16 LEDs and dimming never actually worked.

While Henne shares much of his code, this particular set he provides only as compiled HEX, so I couldn't really debug it.

(I also tried to build a strip-board version of the circuit, which I never got working. I further spent a lot of time trying to learn a variety of schematic- and PCB-drawing software applications, to little or no avail.)

So, for 2010, I would like to propose that a group of people develop what I'm now calling the Sawyer Star, after the master of getting other people to do your work for you, Tom Sawyer.

Despite the name, I won't sit by the side of the road and munch an apple while you whitewash Aunt Polly's fence. I will:

*Build the circuit on a breadboard to assist in software testing and proof-of-concept.

*Send out to a board house and purchase prototype boards, building up a set to make certain they operate correctly.

*Should there be sufficient interest, manage a co-op for PCBs and/or components.

*Develop a construction manual that will outline how to build both 18-inch and 24-inch stars, using either 5mm LEDs or 10mm LEDs. Additionally, if there's interest, I'll develop construction techniques for acrylic snowflakes and plastic or real wreathes.

*Develop/maintain the BOM; provide alternate components as necessary.

*Co-ordinate and manage day-to-day questions, user complaints, etc.

Personally, I would prefer that the microcontroller be Atmel, as I am much more comfortable with the AV-GCC development chain, but if no Atmel developers come forward, I would consider going in the Microchip direction.

What is needed -- and this can come from multiple volunteers -- is:

*Someone to design the circuit. While the Thorpe and Henne circuits are certainly building blocks, there's much open to debate here. Processor -- ATTiny2313? Another processor? Should it really be a UNL2803 or should it be a TI TLC5940 (or a TLC5920)? Should it operate off of household current (on-board 110v/220v with built-in transformer) or from a nine-volt or 12-volt wall wart? (Or a nine-volt battery, as per Thorpe?) Should the output termination be screw terminals? RJ45s? Heartbeat LEDs or not? What sort of jumpers need to be included? An ICSP header?

*Someone to convert that circuit to an efficient printed circuit board. Should it be one-layer for easy self-etching? Does it need to be two-layer?

*Someone to write the code. C or assembly? DMX control is obvious, but can we have switchless DMX channel assignments or do we need to provide for physical switches? Should we provide some sort of functionality similar to Thorpe's, so that the device can operate in a stand-alone manner without being attached to DMX/Vixen? Should we provide for a way to control all of the same colors with only one channel?

This would be an open source project, which would mean a sharing of sources but a retention of individual copyright.

I would anticipate a certain amount of give-and-take between the various participants, with the goal being to create the most appealing device. We would do this development right out in the open, with the participation of anybody who'd be interested.

So, what do you say? Wouldn't painting a fence be fun today?

Thanks.

\dmc

PS: In response to this posting, I'm not only looking for volunteers to handle the three sets of technical tasks, but also general commentary on the usefulness of the design and whether this would be a board that you, a Christmas light nut, want developed.


53 Replies
Posts: 6728
(@p-short)
Illustrious Member
Joined: 20 years ago

Re: New project: Sawyer Star (Charlieplexed LED array)

Does the RPM DMX-dimmer handle DC as well as AC outputs? If so, I think that board (DMX_dimmer + Grinch) would do most of what you want.


Reply
Matt_Edwards
Posts: 2664
(@matt-edwards)
Famed Member
Joined: 18 years ago

Re: New project: Sawyer Star (Charlieplexed LED array)

Yeap,
RPM's design uses PWM for both AC and DC. the design using Tiny2313s are limited in the number of channels they can control compared to the Mega88 (Grinch Dimmer), but in this case no problem.
For a fast and nasty prototype grab 2x DMX8 DCSSR boards.

A DMX16 DCSSR is either on the way (RPM?) or could easily be hacked together quickly ( I can help here with code and Schematic).

However I am wondering how easy the Star will be to program?

Heres a thought:
I wonder if it would be better to have predetermined sequences in the uC that are triggered by a specific DMX level, or channel.

Cool project
Matt


Reply
Posts: 6728
(@p-short)
Illustrious Member
Joined: 20 years ago

Re: New project: Sawyer Star (Charlieplexed LED array)

NM, to me it looks like RPM's design(s) are the way to go.


Reply
Posts: 597
(@entropy)
Honorable Member
Joined: 19 years ago

Re: New project: Sawyer Star (Charlieplexed LED array)

Don't think you can do Charlieplexing using a ULN2803 - Just matrixing.

You should be able to tweak the Rainbowduino design to do DMX reception. I'd ditch the MBI5168s though - they're very difficult to source.

Wiring to the LEDs of the star will be a royal PITA, probably the hardest part of the project!


Reply
Page 1 / 11
Share: