A bad design that for most control range LEDs are off, then there is some brightness control and then turn a little bit more and you have your LEDs fried. Trying to control LED brightness without series resistor by adjusting voltage going to then is a bad design. The LM317 should have been configured to adjustable constant source operation mode.

Here are some links to LM317 constant current drivers:
https://www.instructables.com/Super-simple-high-power-LED-driver/
https://www.electronicecircuits.com/electronic-circuits/lm317-constant-current-calculator/
https://www.ti.com/document-viewer/lit/html/SSZTC13
https://www.youtube.com/watch?v=t6nqd8_g5A8

Create a Simple LED Driver document at https://www.ti.com/document-viewer/lit/html/SSZTC13 gives some details

The LM317-N was the first adjustable voltage regulator, introduced way back in 1975. It’s still widely used in the industry as a constant voltage supply. But did you know that this device can do much more than regulate voltage? Its versatile architecture allows you to use it not just as a voltage regulator but also as a cool, constant-current LED driver.

Wait. Any linear regulator could be used for LED lighting with the addition of a current-limit resistor … what’s new with the LM317-N? Well, it’s actually not what’s new but what’s old that makes the LM317-N click as a simple LED driver.

High frequency strobing can be hard to detect with your eyes. This issue relates to the fact we have a 50Hz electricity supply. In this system, electrical current alternates or cycles 50 times per second. Well designed LED lights effectively ‘smooth out’ this constant back and forth of the current flow. Cheap and nasty LEDs cycle on and off completely, 50 or 100 times per second. In countries with 60 Hz power, the flicker can be at 60 or 120 times a second. We know this is bad because old-style fluorescent tubes can suffer from the same problem. High frequency flicker in offices has been linked to increased headaches and visual discomfort.

If you want a flicker free LED bulb, how to find it? The markings on the packet usually don’t tell about how much this specific LED bulb flickers or not. So you need to test yourself. Your phone camera can potentially provide some insight into LED bulb flickering (the pulsing screen is an indicator of this flicker): At hardware stores they have a row of sample bulbs that are on, put your phone camera up to them to see how much flicker. Then buy the best ones.

IKEA LED lamps flickering comparison

The majority of people have seen flickering light bulbs. LED lights, despite their incredible efficiency and supposedly infinite longevity, do flicker on occasion, just like any other lighting system. If you have flickering LED bulbs, that flicker either all the time or when mains voltage varies, homeowners may have to troubleshoot and fix the issue to ensure maximum comfort and functionality of their lighting system.

When using LED lamps, certain appliances may not function properly. LED bulbs have commonly issues when using lighting switches that also incorporate a dimmer controller. Most LED lights are not compatible with traditional dimmer switches, causing flickering in some instances. If you have dimmer switches, start by purchasing products that have market-proven wide compatibility. Just because other lights have flickered, it does not mean that all LEDs will flicker with your dimmer.

The light fixtures in your home will eventually also get old, broken, and rusted. This can nonetheless result in flickering light bulbs. The simplest way to avoid the dangers of flickering LED lights is to ensure that your LED lights are properly fitted (twist the LED bulb to make sure it’s not just a loose). The next test for flickering LED bubs, remove the flickering bulb and replace it with a working one. If the flickering persists, it could be your light fixture. If the flickering persists after you change the fixture, have a professional inspect the wiring. If there is something wrong with the power source, you should call an electrician to have it serviced.

Low-cost LED bulbs may result in flickering, so it’s beneficial to use high-quality ones. High-quality LED bulbs often come with features, that can help prevent flickering and ensure optimal operation. But claimed high quality, well known brand name or high price does not guarantee good results.

If you see any brief flickering, consider if anything else is going on in your home. When huge appliances are turned on, lights may briefly flicker. Led lights mainly flicker either because of voltage changes or a loose wiring connection. Voltage change can happen when turning on a large appliance. Light flicker caused by starting intense appliances like a corded vacuum cleaner or other tools is usually not a cause for concern. Because they suck a lot of power, they can cause a voltage reduction. If your LED lights hook up to the same circuit as the appliances, they may flicker or dim.

Energy utilities control off-peak loads by sending a ‘ripple’ through the electricity grid. Often, this electrical signal can show up in the form of flickering lights or noisy fans. Other disturbances on the grid can be due to nearby electrical loads (like neighbour cranks up their welder or high power grinder).

The LED’s could be tested optionally with following currents: ca. 1 mA, 2,5 mA, 5 mA, 10 mA, 20 mA and 50 mA. I have tape over the 50 mA pins for safety (not to fry small LEDs accidentally with over current for them).

The test contacts are designed to accept brand-new LED’s as well as soldered out LED’s with residual soldering tin at the connecting pins. The connectors work acceptably for this application, but they do not always make the most reliable connection to LEDs.

In order to facilitate selection of LED’s of equal brightness, there have been placed two test sockets with the same currents (10mA) side by side. So you can quickly get comparable brightness LEDs before soldering them in your projects.

Let’s take a look at what is inside the box.

As you can see there are there are only connectors, resistors and battery in it. The 9 V/DC battery provides power and the resistors limit the current to the LEDs. Here is a closer look at the circuit board.

With 9V battery a resistor makes a decent current source only for LEDs with forward biases in a narrow range. The actual current will vary by about 10% for FB in the range of 1.6-2.2 V for most LEDs. The voltage drop over blue and white LED may be in range 3 to 3.5 volts, which causes the current through LED to be 20-25% lower than the shown current. The reason for that is that with traditional LEDs the voltage drop on current limiter resistors is around 6.8-7.4V. With blue/white the voltage drop will be only around 5.5-6V. The current through resistor is directly proportional to the voltage over it.

Here is a teardown of another cheap LED tester:
https://www.epanorama.net/newepa/2018/05/01/cheap-ebay-led-tester-fixed/

It looks like when AC voltage waveform instantaneous voltage lower than some voltage, it takes practucally no current and above it it looks like a constant current load.

This is picture of the light.

This looks like this uses one of those direct mains AC powered LED modules, like those described in detail here: https://hackaday.com/2019/02/15/cob-led-teardown/

Closer look

Circuit diagram

This is very simple design. It just needs two resistors (I used three because I did not had exactly 24.9 kohms resistor in hand so I built it from two resistors in series). This is a way simplified version of my earlier PoE Powered Christmas Lights circuit.

The 24.9 kohms resistor connected between power input wires is the IEEE 802.3af PoE identification resistor that makes the PoE power supply to know that there is class 0 PoE capable device (can take up to around 15W power and does not specify exact power it needs). The IEEE 802.3af power supply probes the resistor at lower than 20V voltage low current pulses, so at those voltage levels the LED string does not pass any current (it has many LEDs in series and is designed to work at 31V DC). When power supply detects proper PoE device, is starts to feed 48V DC to the line. That 48V is feed through 390 ohm resistor to the LEDs, this resistor will limit the current an “burn” the voltage 48-31V=17V.

This circuit worked well with two different PoE power supplies with the 31V LED string. I tested that the circuit worked also with LED string that was designed to work at 24-26V DC voltage. Do not try lower voltage LEDs in the output, because trying that risk of overheating the resistor and messing the IEEE 802.3af device detection.

PS. In full standard compliant design there should also be a bridge rectifier between RJ-45 connector and rest of the circuitry to handle situation that the wiring can sometimes change the order of power carrying wires.

This kind of LED hat is something that makes hacker mind spinning. The hat had wireless connection so, can anynone hack their “Mobile APP Bluetooth connection”? I think sending random nasty words to nearby hats could be funny if those become popular. Or sell a large number of other person’s hats as advertisement space for good money. Those were just few ideas that could be illegal to try, so best not to implement them.

Another ideas are that this is something that some hardware hackers must have made tried to make themselves. And yes they have. With Arduino and with LEDs like WS2812 this kind of ideas can be implemented quite easily. Here are some related DIY ideas for hardware hackers:

https://blog.adafruit.com/2018/06/06/gif-playing-led-hat-how-to-wearablewednesday-wearabletech-diy-arduino-tech/
https://www.instructables.com/id/Fancy-LED-Hat/
https://makezine.com/2013/09/03/the-awesome-led-hat/

Researchers have now shown switching between LED and photodetector modes on new type component quickly enough for real-time use, the new diode offers potential for communications. Switching between LED and photodetector modes quickly enough for real-time use, the new diode offers potential for communications. The new LED at its peak emission (~804 nm) allows an optical signal exchange between two identical diodes. The operation speed for both functions can reach tens of megahertz.

“In order to demonstrate the potential of our diode with double function,” researcher Chunxiong Bao explains, “we have built a monolithic sensor that detects heart beats in real time, and an optical, bidirectional communication system.”

Researchers Create a Perovskite Diode That Can Transmit or Receive Optical Signals on Demand
https://www.hackster.io/news/researchers-create-a-perovskite-diode-that-can-transmit-or-receive-optical-signals-on-demand-8eb41a4c5e23

Bidirectional optical signal transmission between two identical devices using perovskite diodes
https://www.nature.com/articles/s41928-020-0382-3

Use of LEDs as a photodiode light sensor
https://wiki.analog.com/university/courses/electronics/electronics-lab-led-sensor

Here is an interesting  two article series that proposes a new wiring system especially for LED lighting.

https://www.edn.com/design/led/4462200/Proposed-LED-wired-IoT-standard-can-reduce-energy-use–part-1-

https://www.edn.com/design/led/4462201/Proposed-LED-wired-IoT-standard-can-reduce-energy-use–part-2-

The new wiring standard will be for interconnecting smart LED light bulbs, wall light switches (on/off/dimmer), occupancy sensors, thermostatic temperature sensors, fans, duct dampers, thermal blankets that unroll over windows, solar PV arrays, and other IoT devices.

The proposed IoT base standard doesn’t define power wires. Power could be fixed 24 VDC, fixed 48 VDC, fixed 110 VAC, or fixed 220 VAC. In proposed system, ±data wires connect multiple devices in parallel, devices communicate bidirectionally.

Researchers say that this makes making LED lights cheaper, because according to them as much as 60 percent of the cost of an LED bulb comes from the driver electronics. I think this number might be old news from era of complicated isolating low voltage supplies.

Nowadays LED bulb makers from China have made the power non-isolating supplies for LED bulbs very cheap (easiest one cheap chip or capacitor). So many of the ideas used here (series of LEDs matching mains voltage and simple current limiting) are already used already on cheap mains powered COB LED modules built with just few components. Like this:

https://youtu.be/PklaByEQSZA

Especially the compact battery holder that takes three AAA batteries looks useful in other small projects.

Those worked OK when lightly loaded. But when I started taking current for a small motor, I could not get acceptable power out of them.

Some measurements showed that when load current is increased, the output voltage drops considerably. It looks like there would be almost two ohms of series resistace inside the battery holder!

Investication revealed that each spring made of thin wire on the battery pack had around 0.5 ohms resistace. Three springs and some more thin metal makes almost two ohms.

Was this resistance because of extreme material cost saving? Or is it intentional distributed series resistance uses as part of simple LED current limited circuitry?

I can’t say for sure. But it can work as such series resistor. I could wire a 3W UV LED just directly to the battery pack output and get reasonable amout of light. The LED directly wired got slightly less than 1W power with the batteries I had as with somewhat used batteries current was around 200 mA. With brand new batteries current was around 300-350 mA, that is right for 1W LED.

End of test.