But for the 'real cheap' bulbs, they likely (due to being "real cheap") have either a half wave or full wave rectifier (i.e., no switching PSU) which results in the LED's having a flicker at power line frequency (either 50hz, 60hz, 100hz, or 120hz depending upon which combination of line frequency and full/half wave rectifier is present).
Unfortunately, Powerwalls and other similar products are made with built-in inverters and connect only to AC, there is no DC tap. And there aren't any standards around DC wiring and small appliances, so it isn't likely to get traction.
I'm mentioned before (https://news.ycombinator.com/item?id=21109247) that I've seen a DC installation at a friend's house, he used 12V cable lights and DC bulbs, so it seems to work.
Imperceptible if you are just looking at it, but in a dark room if you move something quickly in front of it you can see the strobe effect.
In the CATV industry, we used to send 110 analog channels across roughly 800MHz of spectrum (54MHz to 860MHz). Other than background and ingress noise, a major source of noise (effectively) was distortion - non-linearities in the amplifiers that kept the signal at a reasonable level. Composite Second Order (CSO) and Composite Triple Beat (CTB) distortions were two values that equipment transporting these signals would generally call out - they are analogous to the summing and differencing of two signals and three signals (respectively) as described above for the lightbulb, but imagine doing that with 110 different frequencies simultaneously.
Sorry...I just noticed you are talking about filament LED and not old fashioned analog filament bulbs.
The OLED displays in current smartphones also exhibit a visible strobe effect at lower brightness settings. I figured I’d get used to it but ended up returning my iPhone 11 pro after a week because it was driving me nuts. I’ve met a few people who perceive it as well but most seemed to have no idea what I was talking about.
Filament LED bulbs (the ones that look almost exactly like clear incandescent bulbs with a visible filament when turned on) do this because they have to fit their electronics in the tiny metal base. Room enough for some diodes (and maybe a capacitor), but usually not a full set of high frequency switching electronics like in most other LED bulbs (although I've seen some filament bulb tear downs that might show more advanced electronics). But again, doesn't bother me due to their high duty cycle and relatively okay refresh rate.
(At least, this is all my impression. Please correct me.)
https://thebark.com/content/heres-what-dogs-see-when-they-wa...
Even if that 75Hz number were derived from deep understanding and measurement of the entire visual pathway, there are variations in perception for people, there most likely are for dogs as well.
Perceiving flicker has a higher threshold. I can perceive 60Hz flicker in my peripheral vision easily enough.
Other sources online (you can find lots of them but I didn't see an obvious authoritative article) suggest that 16Hz is the flicker fusion rate for motion in humans.
Flicker fusion for continuous brightness (CFF) is somewhere around 30Hz per https://www.nature.com/articles/s41598-017-15034-z and it rises with brightness - this paper is about testing how it changes in different circumstances.
Most of the reason you don’t is because filmmakers know very well what their “maximum pan speed” is and they stay as far away from it as possible.
1000Hz I find hard to fathom, though. Would want to see what these signals look like.
Classical 50/60Hz lighting looks like a stroboscope to chickens. There is a whole field of research on the effect of light regimes, light color and flicker on poultry farming efficiency. I couldn’t find a scientific link, but see https://agrilight.nl/lichtadvies/pluimvee/?lang=en&-en
Are all these cheap SMPS driving our pets insane? (or just annoying them? or neither?)
I wonder if I could rent some super-high-quality mic from a local sound reinforcement studio and look at a 20-60kHz FFT ... or do fancy mics not really go beyond human hearing?
The best flicker-free lights are ones that have low ripple constant current power supplies (somewhat expensive). LED lamps intended for use around rotating machinery used to all be this way until someone figured out that if the PWM frequency was wildly unstable it would prevent the problems with strobe lights around spinning things at a fraction of the cost.
What does it take to make them not flicker? I.e. how much of a cost increase is required?
Honestly it's starting to feel like a most cost effective long term solution is to build homes with a master AD/DC converter (or several depending on sq/ft) and simply wire 12v outlets and light fixtures. The energy savings of swapping all incandescents or compact florescence bulbs with cheap LEDs with cheaper power converters is starting to scare me knowing just how shoddy and fire prone some of these power converters can be.
Let say it is called CLED,
And it requires the following features / spec before it could be called CLED. From PWM, Colour Correctness, Energy Efficiency, heat etc. The different between a low price light bulb and higher priced is relatively minimal , the difference in LEDs is massive. To be point they should not even be allowed into the market in the first place.
Philips' discount brand (Attralux) is even cheaper, I recently bought more of these 8W 800lm for less than €1 a piece. For now I cannot tell the difference.
Just white. We run whatever cheap bulbs in my house at the moment, but also tend to do to a lot of "slow mo" videos and it kills me every time.
I'd prefer NOT to spend $15+ a bulb for crazy high end but would be more fine with say, $5+ a bulb
Indeed, to the housefly, an incandescent-bulb-lit house slowly pulses from light to dark, so probably too with cheap line voltage choppers.
This sounds amazing to me, but also confusing. My understanding is that an incandescent bulb works entirely by heating up the filament. There should be no way that the filament is cooling down and heating up in anywhere close to the rate that it would require to get some kind of wagon-wheel effect, which is what it sounds like you're describing.
The varying current through the tungsten will result in a varying power consumption (varying at a 60hz rate). The varying power consumption will result in a small amount of varying temperature on the filament.
Now, the filament does not cool down instantly, so the result is its temperature will vary by a few degrees, but likely not enough to ever be perceptible to human eyes. But with a sensitive enough (and fast reacting enough) temperature probe, one could likely measure the temperature rise/fall of the filament that is synchronized with the line frequency.
I'm not sure you've shown that this would be visible to a housefly as a "slow pulse from light to dark," as the post I was replying to claimed. First of all, you said the power consumption will vary at a 60hz rate, but it seems to me it will vary at a 120hz rate, since the current will be at its max magnitude twice per cycle (as you say in the first paragraph). But more to the point, I don't know whether the temperature/lumens will decrease enough in 1/120th of a second to be visible to houseflies or anything else.
Trying to research the answer, I found this physics lab worksheet [1] from Pasco (the makers of sensors), for a high school lab measuring the output frequency of incandescent bulbs vs fluorescent bulbs, but without doing the lab I don't know the results.
The numbers here show a >25% difference for an example bulb.
Also, even ignoring how the cooling starts off at its fastest, with radiation scaled to the fourth power of temperature... incandescent lights switch off in about a tenth of a second. Even if that was completely linear, it would be very visible. Even a 1% flicker would be visible.
https://www.youtube.com/watch?v=eUprJS9sXYU
https://www.youtube.com/watch?v=jXx5eEzMmYo
It's not an extreme brightness change, but it's possible to do a lot better.
Anecdotally from my experience as a macro photographer, the threshold for flies is lower than for dragonflies; flies often startle in response to movements almost too tiny to perceive having made, while dragonflies are relatively easy, usually requiring only a few minutes to approach from a distance of a meter or so to the ~12cm minimum focus distance of my best macro lens. (It's hard to know; the passage of time isn't of much interest in the focused flow state that's required for this sort of activity.) It helps that disturbed dragonflies tend much more often to return to the same perch, but with some practice it's possible to make the entire approach without disturbing them at all. The only really tricky part about it is that, when you're moving slowly enough, they also tend to land on you, which can be somewhat distracting if you're ticklish.
In any case, the existence of a minimum rate of change for perception of motion suggests that flickering light below a certain frequency might well be perceived as strongly discontinuous.
(It also merits mention that my macro rig includes three very powerful flash heads mounted around the lens front element. I've never observed dragonflies, wasps, bees, flies, or spiders to react to these in any way, even when firing from a distance of six inches; the only reaction I've seen has been from fall webworm caterpillars, which displayed a communal defensive response, and that may have been as much due to the shadow I cast, or to the polistid wasps hunting nearby, as to my flashes firing. The wasps notably did not care at all about me or my flashes, especially after they also found the nest and busied themselves with its rapid depopulation.)
Then compare it to something like a Phillips hue bulb. The difference between the two is pretty big.
On a side note I sometimes wonder if other animals have a quicker “refresh rate” on their eyes and see nothing but very visible flicker on these bad LED bulbs. I also wonder if some animals could see the raster scan in action on old CRTs...