This is in fact a common way of making LED “throwies” - just a 10mm LED, a CR2032 coin battery and a neodymium magnet all taped together..
Ignore this post.
I got LEDs because I didn’t want to have to use 1800W current. I was trying to get something completely safe.
I don’t know how to get an incandescent bulb any more. Fluorescents don’t dim well, so I’m at LEDs.
I wasn’t talking about using bulbs intended for household 110V current. But rather low voltage flashlight bulbs like these: Amazon.com : incandescent flashlight bulb.
Amazon, search “incandescent flashlight bulb”. But the chain hardware stores have them, too.
ETA: Damn you, @LSLGuy! 
This won’t electroplate the electrodes in the ohm meter?
Not if the meter is the excitation source; it sources just a few hundred microamps for the resistance function. Plus your test probe tips likely have a nickel plating on them, which is very corrosion resistant. (The tips of your fingers have a layer of saltwater and acids, yet touching the tips of the probes won’t cause them to corrode.) If you’re concerned about it, simply use a couple stainless steel probes in the jar of water (stainless steel bolts would work fine), and connect to them using alligator clips.
60 Hz is well above the flicker rate. The light would look steady. The standard for film movies was 24 frames/second.
The flicker fusion threshold is highly dependent on circumstances and is usually higher (sometimes way higher) than that. You probably will have seen 60 Hz flickering in the wild. In the right settings the fusion threshold can be north of 1 kHz for some.
For the experimental set up proposed above, there are many things working toward higher fusion thresholds. The type of light source offers no intrinsic persistence; the intent is to run it at non-design currents; the source will be only half-rectified AC, leaving the LED off more often than on; the LED is a visual point, meaning persistence in the visual system doesn’t help during eye motion (voluntary or otherwise); the room lights will probably be dimmed or off; and the audience appears to be a room full of teenagers, for whom flicker fusion rates will be highest.
I hedged in my original mention of this, but to be honest, I would definitely bet that the flickering would be noticeable in this scenario.
If you are waving the light around, yes you can see stroboscopic effects. That seems more like fun than a problem.
If the diode nature is refusing the AC in one direction, why would the light be off more than it’s on?
Also, I just did this setup with a transformer, using the AC side. We didn’t notice any flickering. I put one leg of the LED into an alligator clip, and then the end of the leg into the water. The other clip was put in the water, and later pencil graphite was put in the clip and then into the water.
It very much occurs to me that with an LED as the connector between the clips, what we have is pulsing DC, not AC. Even so, we couldn’t notice any flickering.
Questions:
- What is the black(?) solid coming off the negative side? The LED legs are iron-based, so it’s an iron oxide?
- The positive side is producing H2 gas which isn’t corroding the steel parts, so it escapes?
- I used NaCl, and the water went yellow. I tried to smell Cl2, but couldn’t. My student had theories about the yellow being Prussian blue dissolved, but I was too busy and skeptical to really retain the idea. Any theories about the yellow water?
BTW, I manages to tape pencil “lead” to an LED leg and put that in the water to avoid more iron corrosion. Just a note to anyone trying something similar.
It’s off when the voltage is the wrong polarity, and it’s also off when the voltage is the right polarity and not enough to overcome the forward bias of the diode, and it’s effectively off when the current is flowing but not at enough of a level to make much light.
A good point. You can use two LEDs in parallel, wired in opposite directions, to recover the chemistry-mitigating effects of AC (even though it sounds like you’re getting some fun out of the electrochemistry with the half-wave rectified AC).
Nope, I don’t get it, about putting LEDs in parallel. The current can only go through one direction, and it’s the same direction for identical LEDs. Can you be more specific?
The key point is “wired in opposite directions”, so that as the AC voltage switches polarity one or the other of the two LEDs will always be forward biased - when one is forward biased (“on”) the other is reverse biased (“off”).
Nope, I don’t get it. Identical LEDs are biased the same direction, and need the flow of electrons going in the same direction as each other. Placing them somewhere physically doesn’t mean that the current will be heading in the correct direction for one and not the other.
LEDs have two terminals, an anode and a cathode. When the voltage on the anode is higher that that of the cathode, the LED is said to be forward biased and if the forward bias is large enough the LED will conduct current and emit light. When the voltage on the cathode is higher than the voltage on the cathode the LED is said to be reverse biased and will not conduct current and remain dark. (I am simplifying things a bit here ignoring second-order effects).
If you connect two diodes so that the anode of one is connected the the cathode of the other and vice-versa (anti-parallel) and apply an AC voltage to the pair then for one half of the AC cycle one diode will be forward biased and the other reverse biased, while for the other half of the AC cycle their roles are reversed.
@Marvin_the_Martian’s clarified it, but in case it helps:
.--->|---.
.-----/\/\/---| |----.
| `---|<---' |
(VAC) |
| |
`---------------------------'
(And just to be sure: “in parallel” is a technical term, not just a physical description.)
Connecting LEDs in antiparallel is a common way to power LEDs using AC or bidirectional DC. You can even purchase LEDs that contain two dies on the same substrate and internally connected in antiparallel. These LEDs are more commonly called “bidirectional.” Each die can be the same color (e.g. red-red) or two different colors (e.g. red-green).
Something else to keep in mind: while an LED is a diode, you should never use it as a diode, because the reverse-bias breakdown voltage is really low, perhaps as low as 5 V. This is why a regular silicon diode is always connected in antiparallel to a standard LED when the LED is powered from AC; it will guarantee the max reverse-bias voltage across the LED is less than 1 V or so. The only exception I can think of (when it’s O.K. to use an LED as a diode) is described in the previous paragraph, i.e. connecting two LEDs in antiparallel. This is because the reverse-bias voltage on each LED will be equal to the forward-bias voltage of the other LED, and the forward-bias voltage of an LED is usually pretty low (less than 4 V typically).
My go-to is Mouser.
They are the default for most of the Broadcast television engineers I know as well. Figure out what you need, they’ll have it.