Dark current.
One spec that isn’t published or measured for pixels is what I call “dark current” - the current needed per pixel when all dark, ie: (0,0,0) RGB values. This partly depends on the chip, but it also depends on the circuitry around the chip, eg: on the PCB supporting the chip. One potential power hog is the voltage regulator supplying logic power to the chip, especially when operating at 12 or 24V - sometimes this is a cheap & small but power wasting Zener shunt regulator on the PCB for example.
Why does this matter? For some cases it won’t matter at all. For a battery powered system it could be significant, particularly if (1) you plan to have a low average brightness to conserve battery, only to find that the dark current is high so you draw more current than your power budget allows even when the lights are dim or off. Or (2) if you plan to have the pixels powered up but off most of the time. I was thinking to put a bunch of them around the edge of the living room for mood lighting, but in the simple case I would be drawing dark current 24/7.
I found that the 12V 2801 pixels I was thinking to use draw about 5ma/pixel (1 A for 200 pixels). The 12V TM1829 pixels I was considering have a dark current of about 6ma/pixel.
For long periods of totally off, the dark current can be mitigated by using a power FET or a relay on the power lead and cutting all power to the pixels when not needed (case 2 above). This doesn’t work if one wants long battery life at low average brightness outputs tho (case 1 above).
One warning - some pixels still light up using parasitic power from a still powered data line, tho they may be dimmer and less reliable. This might overload the first pixel’s logic circuitry, or the controller’s output data line tho. So be sure to drop the data lines low or disconnect them too, if you cut the power line to the pixels.
Dan, if you do create the table of pixel characteristics, please include a dark current column, and we can try to collect measurements for everybody’s reference. We will need to distinguish between 5V and 12V pixels using the same chips in some cases, so maybe a column for dark current at each voltage. We can all contribute measurements.
We’ve been collecting/including dark current notes on the google code page - https://code.google.com/p/fastspi/wiki/ChipsetOverview - chipset power draw is what the chips pull without led light on. Ill ad the details for the 1829s when not on the phone.
I’m kind of discouraged that so far I’m not seeing any 12V pixel strings that combined high PWM frequency with low dark current.
For pixel strips, I don’t yet know about the LPD8806’s dark current, but that chip controls 6 channels or two RGB pixels so it’s not used in pixel strings (and I don’t find any LPD8803 based strings). (If the LPD8806 turns out to have low dark current then perhaps strips based on it would have both features but I’m looking for strings)
Using 5v pixels strings might help, but there are problems with voltage drop. I’m looking at a 50+ foot linear run with pixels every 6" or so - no problem with 12V pixels, problem wih 5v pixels - might need a heavy gauge power injection lead woven in to supply the far end. Or carry 120VAC along with the pixels and put in a local 5V PS every 50 pixels - cumbersome and less safe
I was wondering too if that was a factor. Using the same technology, faster clocks tend to use more power - at least for some parts of the curcuitry. Comparing 5V chips with different PWM rates might give some evidence of whether faster PWM leads to higher dark current.; 12V operation adds the additional factor of how efficiently the 12V supply is being regulated down to logic levels. But there are also questions about how the PWM is implemented; eg: using a 256x clock and digitaal compaitor, or using a sawtooth wave and analog comparator. It would be nice (for my purposes) if dark current was one of the standard specs, and if there was at least some competition to reduce it.