I suppose that it's possible that a given LED string might not light up at all with 120VDC. This might be the case if the sum of the Vf voltages is greater than 120V. That same string could work fine with 120VAC because the peak voltage for a 120VAC signal is around 170V, and current would flow through the LEDs at the time that the instantaneous voltage is somewhere between 120V and 170V. As Entropy notes, this would vary from brand/model to brand/model of LED strings.
Come to think of it, there's another issue as well. Several years ago there was a problem with a particular brand of premium strings burning out, and some of the input connectors were disassembled. Inside were some electrolytic capacitors inside, forming (as I recall) a rudimentary low-pass filtering function. I don't remember what the cutoff frequency of that circuit was, but that circuit is going to behave differently at DC as above the cutoff frequency, which would also affect the brightness significantly.
FWIW, I plan on running all of my channels on 12VDC. I've started converting all of my AC led strands to run on 12VDC power. I found that putting a red green blue yellow and orange bulb in series drops almost all of the 12V (11.35V IIRC) plus a 33ohm resistor sets the LED current to about 20mA. Then I run each segment in parallel for a series/parallel strand, which works great with the way the LED's were wired for non-rectified AC. I just had to remove the pair of resistors that came on the string and break the series every 5 bulbs.
I _hope_ to run 32 channels off of a single computer power supply (12V 20A).
20mA/segment 12segment/string = 240mA per channel * 32 channels = 7.68A
I realize that this is unrealistic for most people who have more than a few LED strings as I do. The amount of effort to convert the strings to DC could be overwhelming.
P. Short wrote:I wonder it's time to start looking for a replacement approach to dimming strings. The current triac-based approach generates a lot of garbage on the AC power lines, which I think is the cause of some of the current 'flicker' problems, and is likely to get worse as the number of channels increases. Some of the flickering problems may be helped by using snubbers, snubberless triacs, and MT1-Gate resistors on the SSRs, but these solutions don't reduce the amount of noise going back into the AC lines.Possible approaches are reverse-phase dimming (not clear if this would be cleaner than forward-phase dimming), or an attempt at sine-wave dimming. I think that any alternative approach is going to be more expensive than the current triac-based forward-phase dimming.
Does this seem like something worth pursuing, or will the issue become moot with the increasing popularity of individually-controlled RGB pixels? At the moment I think that it's already too late to do much for the 2012 season, other than limited testing in anticipation of the 2013 deployment.
For cost concerns, Forward Phase Dimming is your best scenario especially with incandescent lights. Even to some extent LED strings powered off of AC lines. Has anyone tried using the clamp on ferrite chokes to help with EMI and transients generated by the switching Triacs? Using them will introduce some inductance to the load but probably not a whole lot more than say 8 strings wrapped around PVC to form the Texas Arch either.
If you move to DC powering of LED strings, then forms of BAM could be effective and even better control and less noise.
I can envision using a 8 pin PIC or AVR to implement a BAM algorithm for pixels. Then you just sent the pixel color and intensity levels. Two bytes plus address. This would then solve issues with patents on PWM dimming of LEDs that teh WS2801 and like chips may or may not violate. A small 0.75 inch square board can be turned in little time. Panel the layout for the best cost scenario and do a run. The only issue is to which RGB LED. Maybe something worth looking into.
just thinking out loud
james
The only 8-pin PIC which can receive and transmit data at more than about 38400 baud is the PIC12F1822. None of the other 8-pin PICs have a built-in hardware UART, so the UART functions need to be implemented in software. That part has enough throughput that other algorithms besides BAM are possible, which could further decrease the low-frequency components of the lamp modulation. Some of these other algorithms can be found in The Art of Programming by Donald Knuth (chapter 6.4 in volume 3).
