Hmm. I’m not positive, but I think if the PWM is really easy to see it’s probably not 14Khz.
I got the 1.2 / 2.4 kHz thing from the header comment in a STAR firmware file. I don’t know if it’s correct. 
In any case, the shower is actually pretty bad at showing PWM. The reason is because water droplets change shape as they fall, and small droplets change shape very quickly. So, the droplets will appear to flicker even under a constant-output light.
To test PWM I recommend quickly waving a white notecard (or other thin white stiff sheet) through the beam instead. Here are a couple images showing 488 Hz PWM versus 4.5 kHz PWM:
These images fairly accurately show how the lights actually look to my eyes under regular use… but I’m unusually sensitive to the flickering.
Good images ToyKeeper. I missed your last post and was pretty confused about bdiddle’s post after that saying that someone was right. Hence my rather pedestrian response.
bdiddle… the current firmware can be confusing to setup. Please post (not here - use somewhere else like pastebin or whatever you want) examples of you the entire program you used with Phase Correct and then with Fast PWM. It’s reasonably possible that you were not actually making the changes correctly.
TK, if you don’t mind will you please share the camera settings you used to determine those PWM’s?
So many variables with shutter speed, wondering how you got there. (I can freeze a high speed object, or make a snail look fast, depends on what I’m trying to achieve) Point being, how do you do the math after determining shutter speed.
I have found that with all the road construction in my area there is always dust in the air, this dust will show PWM quite well. I would imagine snow flakes would too. (not like I have those to shoot, it was around 70º today)
I wonder what settings it would take to show the faster PWM? Might be fun trying, I’d like to see what some of mine are actually doing but need some info. 
Thanks
Edit: Isn’t it also pertinent to establish speed of the card in fps to get an accurate frame count?
Oh, um… I didn’t.
I measure PWM speed with a very small microphone and a guitar tuner app called PitchLab. When the pulsing is audible, the sound is by far the easiest way to get the frequency.
For some reason, most lights will get louder when they’re aimed point-blank at fabric. I don’t know why that happens, but it’s pretty helpful for measurement.
The pictures are just to give a rough idea what the resulting beam looks like, and how to visually check for PWM. I can only get a ballpark idea what the speed is through visual means. This is at least enough to sanity-check my sound-based measurements though.
I haven’t found airborne particles to be very useful for PWM. For example, this ice fog was shot with a 4.5 kHz PWM light. The particles were small enough to hang in the air instead of falling, and they showed no sign of flickering in the camera as the wind blew them around.
(click to enlarge)
+1
All the way from stable moonlight modes to direct drive with huge trace widths and many vias, on a one sided board.
toykeeper:
http://en.wikipedia.org/wiki/Photoacoustic_effect
try black fabric, it should be louder
thats a cool idea about the guitar tuner app.
Annother thing about the large copper area is that it reduces unwanted high frequency effects. It stabilizes ground at high frequencies with the capacitance between the layers.
That’s one of the coolest things I’ve seen in a long time. I had no idea it was caused by the dark fibers rapidly absorbing and releasing heat.
Now I really kind of want to make a firmware which will play tunes when aimed at black fabric.
Do it it would amazingly cool to be able load an midi file in to a firmware and have a hidden mode that plays the lights theme song
Not only disco modes anymore but an actual music producing disco mode 
Although i suspect something like that would take up alot of room in the attiny, but with all these new bigger attiny’s being discussed maybe someday it could be possible……
It’s cool, but with all the floppy drive symphonies and Tesla coil synthesizers, I think it’s already well-trodden territory. I think I’ll try to focus on making more useful stuff instead, like that off-time dual-pwm firmware people keep requesting.
Yeah you are right of course
But if you would just happen to run out of useful ideas, remember the one about the disco modes that actually sings
People might be surprised by an infrared flashlight with good throw doing that.
Okay, that would indeed be a fun project… infrared laser programmed to play music on objects in the distance.
I’m not aware of whether lasers can turn on and off fast enough, but otherwise it seems like it should work. Maybe I should ask over at the laser forums. A quick wikipedia scan suggests it should be feasible though.
You can’t do it with a laser, because it might hurt people’s eyes. It takes a good deal of power.
Perhaps a wide-beam laser instead of a narrow beam? Like, more focused than a SK-68 but not as much as a typical laser pointer? It’d be really cool to be able to trigger the effect from across a room.
Do different fabrics create different “songs”?
At a party, people could walk into the projected wide beam laser that is creating the disco ball effect and have their shirt/jacket/dress/chest hair/whatever sing their own specific pattern.
So let’s just say I do want to stack a zener. Can you extrapolate on the above? It’s my first time messing with a zener.
The Zener is really actually pretty easy. There’s a “limiting” or “load” resistor, that’s the 100 ohm or 200 ohm or whatever resistor that’s placed on D1’s normal location. So one end of that resistor is wired directly to the MCU’s Vcc. The Zener is also wired directly to the MCU’s Vcc as well as GND. The cathode (striped end) goes towards Vcc, the other end goes to GND.
The Zener will self destruct without a current limiting resistor. It behaves sort of like a low-resistance short to GND. That’s why we add the resistor between Vin (BAT+) and the Zener, it prevents things from getting out of control and allows the Zener to pull Vcc down to the desired voltage.
I’d either put mine diagonally across D1 and R2 (from Vcc to GND) or horizontally from D1 across to FET pins 1-3 (GND).
Post a picture with the Zener laying in place and I’ll verify whether it’s correct.
EDIT: please try and fix the strike-through in your quote of the OP. For whatever reason it’s not working and I’d rather not have anyone get confused. Stacking on C1 is not acceptable for this driver.