Hi all,
A few weeks back, I wanted to grab a strobe light and show my kids the water-drop and stopped-fan effects. I remember doing this as a kid and having it work great.
Well, I guess new COTS strobes suck. Even with the fan on low speed, the effect just wasn't right. The pot swept from maybe 3hz to maybe 20hz.
So I started looking around for some ideas. I have a few Arduinos, but that's overkill. I remember using 555 timer strobes for RC airplanes a few years back.
I went to Radio Shack and picked up a new breadboard (the old one was melted from a solvent spill) an "assorted resistors" pack, a case of various LEDs, an "assorted capacitors" pack, and some 555 timers.
So, for starters, the "assorted resistors", LEDs, and breadboard are awesome. The breadboard is of the 2/5/5/2 type. I hot-glued a 20mm foam sheet to the bottom, cut out a space for a 9v battery and power terminals, and verified the posts were "hot". Then I jumpered from the posts to the power rails.
When I opened the "assorted capacitors" pack, I knew things were not going to go well. Just that sense of dread started to build. It had exactly 1 0.01µf in the bunch. The rest were down in the pf range. Like *way* down. So, I headed back and raided the parts bin for all the 10µf and 1µf I could find.
Next, I got online and played with a few simulators and looked for people doing similar work. Why re-invent the wheel.
I found this site:
http://people.cornellcollege.edu/dsherman/waterdrops.html
Which led me to this diagram:
http://people.cornellcollege.edu/dsherman/pcboard-LED-strobe.pdf
Now, maybe this works in theory. I can tell you for sure it *does not* work in practice. It also *does not* simulate.
So, I played around with some standard values and finally ended up pulling the TI datasheet. It tells me that the 555 should work in astable as long as the R1 and R2 values are >=1k. More playing in the sim and I had a passable astable circuit that used R1=1K, R2 from 689 (out of spec, but sims well) to maybe 80k and C1=10µf. In the sim, it gives me >1hz to 60hz with no problems.
Technicaly, the datasheet said that for "reliable" operation, R1 and R2 need to be >=1k. I don't know why my design works out of spec while dsherman's design doesn't work.
Anyway, once I had the astable portion working, it was time to add a simple monostable circuit.
Again, it's technically out of spec. Really, *way* out of spec. But I started with R1=1000 and found that the pulse was too long. The LED was bright, but the fan blades would "blur" under the light. On my application, I replaced R1 with a 10k pot and started cranking it down until I could see a clear snapshot of the effect I wanted. Somewhere between 200µs and 100µs seemed about right. With a 10µf cap, every 100 ohms gave me about 100µs of delay.
The down side is that my LEDs don't have time to ramp up to full brightness.
Anyway, here is the circuit with a single LED.
Here is the circuit with a 2222 inserted so I can chain as many LEDs as I need.
At some point, I think the sim breaks down and starts doing weird things. I have tested this on a breadboard and it works fine.
If anyone has suggestions for improvement, please let me know.
A few weeks back, I wanted to grab a strobe light and show my kids the water-drop and stopped-fan effects. I remember doing this as a kid and having it work great.
Well, I guess new COTS strobes suck. Even with the fan on low speed, the effect just wasn't right. The pot swept from maybe 3hz to maybe 20hz.
So I started looking around for some ideas. I have a few Arduinos, but that's overkill. I remember using 555 timer strobes for RC airplanes a few years back.
I went to Radio Shack and picked up a new breadboard (the old one was melted from a solvent spill) an "assorted resistors" pack, a case of various LEDs, an "assorted capacitors" pack, and some 555 timers.
So, for starters, the "assorted resistors", LEDs, and breadboard are awesome. The breadboard is of the 2/5/5/2 type. I hot-glued a 20mm foam sheet to the bottom, cut out a space for a 9v battery and power terminals, and verified the posts were "hot". Then I jumpered from the posts to the power rails.
When I opened the "assorted capacitors" pack, I knew things were not going to go well. Just that sense of dread started to build. It had exactly 1 0.01µf in the bunch. The rest were down in the pf range. Like *way* down. So, I headed back and raided the parts bin for all the 10µf and 1µf I could find.
Next, I got online and played with a few simulators and looked for people doing similar work. Why re-invent the wheel.
I found this site:
http://people.cornellcollege.edu/dsherman/waterdrops.html
Which led me to this diagram:
http://people.cornellcollege.edu/dsherman/pcboard-LED-strobe.pdf
Now, maybe this works in theory. I can tell you for sure it *does not* work in practice. It also *does not* simulate.
So, I played around with some standard values and finally ended up pulling the TI datasheet. It tells me that the 555 should work in astable as long as the R1 and R2 values are >=1k. More playing in the sim and I had a passable astable circuit that used R1=1K, R2 from 689 (out of spec, but sims well) to maybe 80k and C1=10µf. In the sim, it gives me >1hz to 60hz with no problems.
Technicaly, the datasheet said that for "reliable" operation, R1 and R2 need to be >=1k. I don't know why my design works out of spec while dsherman's design doesn't work.
Anyway, once I had the astable portion working, it was time to add a simple monostable circuit.
Again, it's technically out of spec. Really, *way* out of spec. But I started with R1=1000 and found that the pulse was too long. The LED was bright, but the fan blades would "blur" under the light. On my application, I replaced R1 with a 10k pot and started cranking it down until I could see a clear snapshot of the effect I wanted. Somewhere between 200µs and 100µs seemed about right. With a 10µf cap, every 100 ohms gave me about 100µs of delay.
The down side is that my LEDs don't have time to ramp up to full brightness.
Anyway, here is the circuit with a single LED.
Here is the circuit with a 2222 inserted so I can chain as many LEDs as I need.
At some point, I think the sim breaks down and starts doing weird things. I have tested this on a breadboard and it works fine.
If anyone has suggestions for improvement, please let me know.
