In order to choose the best options for various aspects of the projects, we need to do some testing.
Some initial testing of a few miscellaneous LEDs showed that the color mixing and spreading of light was really bad.
- Test the color mixing of the LEDShoppe LEDs
- Done: it sucks
- Assume that the diffuser will be placed about 6 in. away from the LEDs. Does the light spread out enough so that there are no strong hotspots on the diffuser?
We need to know how many UV LEDs will be necessary to make blacklight conditions. We also need to know how many (and what type) are necessary for strobes.
- Test UV LEDs from LEDShoppe and figure out how it will scale.
- Done: too weak
- Test white LEDs from LEDShoppe. There are lots of ultra-high intensity options. It may be better (and cheaper) to get just a few of the ultrabrights instead of a ton of the LEDShoppe ones. Since these ultrabrights are in far higher production quantities, they don't have nearly as high prices as the RGB chips, which are only manufactured by a few companies.
- Done: also too weak
Conclusion: use power LEDs (3W emitters) instead. There are far fewer parts to assemble and they mix colors much better.
We'd like to support transfer speeds as high as possible for the distances we require.
- Using the SN75176B, determine if we can transfer data at 1.0Mbps (which is the highest speed that we can handle with the default 8MHz oscillator with 0% error) the full length of the wire (do we need 8 SN75176Bs in order to fully simulate the scenario?). If we have too many errors at this speed, drop to 500kbps and then 250kbps. The output of the chip should connect to two AVRs since this is what the final design will probably do.
- If we can't transfer at those speeds, we'll probably have to use an external crystal oscillator to transfer at 115200bps.