[0:00] In the last video, I showed off these wireless LEDs and I said that it would be really easy
[0:05] to make your own.
[0:06] So let’s give that a go.
[0:08] We’ll reverse engineer the LEDs and create our own versions from discrete components.
[0:13] We’ll then reverse engineer the driver circuit and build our own version of that
[0:16] and then we’ll see how well it works
[0:19] Finally, we’ll do a cost comparison to see if it was even worth it.
[0:22] We should end up with a complete working DIY equivalent that you can easily build at home.
[0:28] Let’s take a look at the LEDs.
[0:31] The actual operation is very simple, they are just a resonant Inductor and Capacitor
[0:35] circuit with an LED across it.
[0:37] The circuit is tuned to the frequency that is being output by the driver coil.
[0:42] In the previous video I wired up a couple of wires on each side of the LED and hooked
[0:47] it up my oscilloscope, and we saw that the circuit is picking up a signal of 217kH.
[0:53] The inductor is marked on the bottom 222 that corresponds to a 2.2mH inductor.
[1:00] Using this inductor value and the frequency the circuit is resonating at we can work out
[1:04] what the capacitor value should be.
[1:06] Plugging the values of the frequency and inductor into the calculator we get a capacitor value
[1:11] of 244pf - 220pf is the closest standard value.
[1:17] This will give us a resonant frequency of 228kHz which is not a perfect match to the
[1:24] 217kHz but is close enough.
[1:26] For the inductor, I’ve got a couple of options - I have these bare inductors and I have these
[1:31] heat shrink covered wire-wound inductors.
[1:34] If you peel off the heat shrink of a spare inductor you can see that they really are
[1:38] very similar to the bare inductors - the only difference is the heat shrink.
[1:42] It does look like there is less wire, but it’s hard to really tell without taking it
[1:46] completely apart.
[1:47] It’s important that you get unshielded wire-wound inductors for this project - it’s the coils
[1:53] of wire that will be picking up the transmitted power and if you don’t have coils or your
[1:58] inductor is shielded, nothing will be picked up!
[2:01] Soldering the capacitor across the inductor wires is pretty simple, I found it easiest
[2:05] to wrap the thinner capacitor leads around the quite thick inductor legs and then solder
[2:10] the joint.
[2:11] I’ve done this for both the inductor options and if we put them inside the coil we can
[2:15] measure the induced voltage.
[2:17] The yellow trace is the bare inductor and the blue trace is the heat shrink wrapped
[2:22] inductor.
[2:23] There’s not a lot of difference and any difference we are seeing could just be due to component
[2:27] tolerances.
[2:29] With both inductors, we are getting a peak to peak induced voltage of around 160 volts.
[2:33] Obviously, this falls off quite quickly as we move the inductors away from the driver coil.
[2:39] Now, of course, you can use other inductor values - just make sure you calculate the
[2:43] appropriate capacitor value - I suspect that larger inductors may have more turns which
[2:48] may give you longer ranger due to the higher induced voltage - but they may also just use
[2:52] different core materials and there may be other tradeoffs in terms of parasitic resistance
[2:57] which will use up some of the energy.
[3:00] Talking of energy - how much current are we getting?
[3:03] I’ve placed a 1k resistor across the circuit and measured the voltage across it.
[3:07] The peak positive voltage is just over 2 volts and the negative voltage is around 3volts
[3:13] - positive and negative is just determined by which way up the inductor is.
[3:17] The asymmetry of the signal is interesting and is probably due to how the power coil
[3:21] is being driven - more on that later.
[3:25] Given we have a 1k load we have a maximum of 2-3mA - which is more than enough to drive
[3:30] an LED.
[3:31] The version that I bought from Aliexpress only has one LED, so it’s only really using
[3:37] half of the energy that is being induced in the circuit.
[3:40] So I’m going to solder two LEDs in opposite directions across the circuit which in theory
[3:44] should give me double the brightness.
[3:47] I’ve made a bunch of different colours and they all work great - the different LEDs will
[3:52] all have different ranges due to the varying forward voltage requirements of each colour.
[3:58] So that’s the LED side of things - what about the driver circuit and coil?
[4:04] Do your projects look like this?
[4:06] Are you tired of dodgy connections causing gremlins and bugs in your circuits?
[4:11] Maybe you should get a PCB from PCBWay!
[4:14] Check out the link in the description.
[4:17] I’ve got a photo of the PCB and I’ve marked up the traces - in the previous video I wasn’t
[4:22] really sure what this chip was - turns out that I was being a bit slow, it’s the XKT001
[4:27] - as marked on the PCB.
[4:30] The large chip is simply a MOSFET.
[4:33] Here’s the reverse-engineered schematic - it’s pretty simple.
[4:37] The XKT001 is outputting 217kHz on pin 2 and switching the MOSFET on and off.
[4:44] The mark to space ratio is about 40%.
[4:47] The coil and capacitor of the transmitter are connected between the power rail and the
[4:51] drain of this MOSFET and is simply being switched to ground.
[4:55] Here’s a trace of the voltage at the drain of the MOSFET along with the signal being
[4:59] applied to the gate.
[5:01] Recreating this circuit is pretty simple.
[5:03] We need an oscillator, a power MOSFET, a gate driver, a coil of wire and a suitable capacitor.
[5:10] One thing to make sure of when you build this circuit is that you use a logic level MOSFET
[5:15] as the circuit will be running at 5v from a USB supply.
[5:18] You want your MOSFET to be fully turned on when 5v is applied to the gate.
[5:24] You’ll also need a gate driver to make sure the MOSFET turns on and off quickly - if you
[5:29] try and drive it directly from your MCU then it probably won’t turn fully on in time.
[5:34] Power MOSFETs have quite a high gate capacitance so if you try and drive them directly the
[5:38] switching time can be too slow.
[5:41] You’ll also find that a lot of logic level MOSFETs won’t fully turn on when using 3.3v
[5:47] logic - so a gate driver is essential.
[5:50] Great Scott has an excellent video on MOSFET drivers that is definitely worth a watch.
[5:54] For my coil, I’m making one that has the same number of turns and is approximately the same
[5:59] size as the commercial coil.
[6:02] Measuring the inductance of my coil I get around 5.4uH.
[6:06] The coil from Aliexpress comes in at 5.1uH.
[6:11] With the homemade coils, it’s quite variable and you’ll probably end up with anything between
[6:15] 4 and 6 micro henrys.
[6:18] Plugging my 5.4uH value into the calculator with a 228kHz resonant frequency it wants
[6:25] a 90nF so a 68nF or 100nF would be suitable.
[6:31] I’ve ended up using a 68nF in my circuit after a bit of testing.
[6:36] You could of course try and construct the exact required value - but given I’m building
[6:40] this on breadboard just being in the right ballpark will be good enough.
[6:44] How much current is actually going through our coil and MOSFET - the current is limited
[6:48] by a couple of things - the first is simply the resistance of the wire making up the coil.
[6:53] and any parasitic resistances in the components
[6:56] I measured my homemade coil at between 0.5 and 1 ohms.
[7:00] If the inductor is fully saturated then we could have between 5-10 amps flowing through the
[7:05] circuit as the only limit will be the resistance of the coil wire.
[7:09] This makes the pull-down resistor on the MOSFET gate pretty important - if we leave the gate
[7:13] floating it could easily allow current to flow and we could cook our circuit!
[7:18] Provided we turn the MOSFET off before the inductor is fully saturated we’ll get a lot less current flowing
[7:24] We can use this formula to work out the current flowing in the inductor after it has been turned on.
[7:30] Assuming we have a 5uH coil, a 40% mark to space ratio at 228kHz and our coil and parasitic
[7:40] resistances in the circuit are 1 ohm the maximum current through our coil should be just under 1.5 amps.
[7:46] Bear in mind, this is just the peak amps - it’s not drawing this continuously the current
[7:51] ramps up to this value while the MOSFET is switched on.
[7:54] Also, the MOSFET will be turned off 60% of the time.
[7:57] Regular viewers of the channel will know that I have a lot of ESP32s lying around.
[8:01] I’m going to use one of these as my oscillator - it has a PWM output and you can easily adjust
[8:07] the mark to space ratio on it to tweak the power consumption.
[8:10] It is however a 3.3v part so definitely needs a gate driver.
[8:15] Here’s the circuit we’re going to build, I’ve added a 20uF capacitor across the power supply
[8:21] to match the circuit from Aliexpress, this helps keep the supply stable when the MOSFET
[8:25] is switched on and when the inductor starts to draw a large current.
[8:29] I’ve put this together on a breadboard - I’ve got my ESP32 module and the gate driver chip
[8:34] - I’m using a TC4469 but any suitable gate driver chip will do.
[8:40] This is controlling the gate of my power MOSFET.
[8:42] We’ve got the coil and capacitor across the drain and the power supply.
[8:47] If we turn it on our homemade LEDs all turn on.
[8:50] What’s more, the LEDs from Aliexpress also turn on.
[8:54] The homemade version seems to have a performance that matches the professional version.
[8:57] All in all a pretty good result.
[8:59] At the end of the day, it is a pretty simple circuit so we should really expect it to work
[9:04] well - but it’s nice to confirm it.
[9:06] So, let’s do a price comparison of my DIY LEDs vs the Aliexpress version
[9:13] From Aliexpress, I received 18 wired up LEDs along with two driver circuits and coils.
[9:19] To build our DIY LEDs - we need an inductor a capacitor and an LED.
[9:25] You can get 600 LEDs on Amazon for around $12, so for 18 LEDs, it would cost 36 cents.
[9:33] Inductors are a bit more expensive, I found a pack of 50 for around $10, so for 18, we
[9:38] need to spend $3.60.
[9:41] You can pick up a box of 840 capacitors for under $20 so the bill for our capacitors is
[9:47] about 50 cents.
[9:48] The total cost of the LED portion of our project comes in at about $4.50 with the inductors
[9:55] being the majority of the cost.
[9:57] For the driver we need a power Mosfet, these come in at about $1.20 and we need a gate
[10:03] driver - I’m using a fairly expensive chip the TC4469 which comes in at around $5
[10:10] but there’s alternative driver ICs that cost around $1.
[10:13] I’m going to assume you have a handy MCU to provide the oscillator.
[10:17] So long as it can do 220kHz or so you should be ok
[10:22] If you want to build an oscillator yourself, that’s probably an extra a couple of dollars on top
[10:27] To make the two driver circuits you need to double the components, so we’re looking at about
[10:31] $4.
[10:32] So I think the total cost of this project is around $8-$10 to build it yourself depending
[10:36] on how cheaply you can source components and what you have already lying around.
[10:40] The cost from Aliexpress was $19.38 once taxes and shipping was added on total was $28.83.
[10:48] The Aliexpress does come with quite nice compact LEDs that might be quite hard to recreate in a home workshop.
[10:55] And you get a couple of nice coils and driver circuits.
[10:58] Let me know what you think in the comments.