HF LED Power Accessories

Unless they’re double layer standard thickness PCBs you might have to wait awhile. Chinese New Year really screws up timelines right now.

That’s awesome! Try and get a video of the field tester frying. It would also be nice to know the resonant peak the device settles on. Unfortunately I don’t have a recent xLED to send you. The electrical characteristics of the field detector will be slightly different because of the components used.

What’s the model number? leumas95 pointed out the IRF6668 to me and I really like the DirectFET package because it has excellent heat characteristics and looks easy to route.

1 Like

The mosfets i found are IPN50R650CEATMA1 which could easily go up to 19MHz at 5w ptot and IPN60R600P7SATMA1 up to 14.5MHz at 7w ptot in theory. Both are in a sot223-3 housing. They will go higher in frequency if you push them. They are on mouser. Not that cheap with 50ct a piece but could be worse. Oh and the circuitboards arrived.

I intensionally made it a bit bigger than it would have to be so i can put tht components on it too. I populated one board with parts but i didn’t tune it yet. Using the sot23 a2shb mosfets it draws 150mA at 3v
And at 3.7v it will draw 400mA at this brightness.
Which is pretty impressive for two mosfets that are smaller than a grain of rice each. They do get quite hot though. As long as you stay below 4v it doesn’t get to hot. It might get more efficient with some tweaking.


So i put some irf 510 on the board and fiddled with values.

at first about this bright and then i tried some more and got this those are the same lighting conditions but the on the second image i got the led alot brighter. At 12v it’s taking up about 160mA. There is no heatsink on the mosfets and they get hot but remain touchable. When it’s the brightest the signal is a hot mess though.

Range is also ok. I still have to experiment more though.

I fiddled around some more and i got it to run pretty brightly at 9V input. At that voltage it draws only 10mA now and is still bright. The range is lower though.

The sinewave is also alot cleaner.
I tweaked it to run at half the frequency of NFC since the bit less coupling is compensated by lower switching losses. With the logic level version of the irf510 (irl510) you could make it run at much lower voltages so i ordered some since my aim is to use round about 3-5v to drive this. https://youtu.be/Wu5QJKM3R98


Do you have a I2C test card or xSIID by any chance? :slight_smile: curious to see how the half speed works with a power harvesting chip.

Looks great!

1 Like

I would recommend you try changing your output antenna shape to one that is wrapped around a pencil to get better coupling.

1 Like

I’ll try that at some point but since i can’t afford them at the time i can’t test those right now.

I’ll also try different coil geometrys at some point. I just have to tune to resonance again when i use a different coil which takes a while.

Oh to the test with the fieldtester: Fieltester 1, Me 0. I have to say though that we couldn’t go up all the way to 500w since we had quite a few losses at different points. We went to about 300W max that was going into the coil. The geometry was quite far from optimal since the friend of mine made it quite a bit bigger than i would’ve

This is the setup and this is the brightest the led got

We also tried a normal iso card to see what would happen there and… well.

The chip is completely gone. Nowhere to be found. So i’d keep every nfc implant that doesn’t use a ferrite far away from a high power field at 13.56 MHz.

I also ordered some more smd mosfets to fiddle around with and try out on the Circuitboards i designed.
By the way i didn’t give up on killing that fieldtester. I have a few things where i’m sure it won’t survive. Also i want to try again at some point with that amp and a smaller coil on the lc tank.

If you feel like DM’ing me your address I’ll ship you an I2C Plus test card to play around with and some more HF field testers for your trouble. Nice vid :laughing:


Today the mosfets should arrice from mouser if the chaos with frost didn’t delay the postage. Also something interesting:

When i was making coffee this morning i checked if the induction plate would make one of the field testers glow and fair enough the 125kHz one lit up brightly on the plate and as expected the 13,56MHz one didn’t light up there.

1 Like

Have you also seen this thread?


Not yet. But i’ll read a little there later

1 Like

Even a couple of videos.


1 Like

So update time!

I decided to change direction and use a Crystal driver IC and a MOSFET driver IC and made a little PCB to test it out.

Here they are:

I had them shipped to @Satur9 as I do not have the skills or tools to solder SMD parts, especially not the package that the MOSFET driver uses.

Satur9 is a fucking legend and just populated one of the boards to test, IT WORKS. There was not enough time to wind a coil and tune the LC tank today (:crossed_fingers: for tomorrow) so we could only look at the gate of the MOSFET, Got a ~13.6MHz sine wave (Not super clean, but decent wave form) with the peak at 8V (Exactly as expected, and this voltage is tunable from ~5-12v) so it’s looking good.

Once the coil is hooked up and there are some photos of the full thing, I or @Satur9 will update here!


So - what is the end goal? Are you talking open source, finished product sales or kit sales? Combinations of the few?

I’m really interested in this project, I love my blinks and the stuff I am messing around with


1 Like

Good question lol… Until this evening did not know if this prototype would even be able to contain it’s magic smoke… and given we still need to like light stuff up with it and measure the power draw… basically I do not want to get ahead of myself.

I consider this specific board just a proof of concept, there was no end goal when designing it other than testing if this concept would work. There are parts that are optional. There are smaller parts available. Not much thought was put into the layout. Etc etc…

I see there being maybe a few designs that follow on from this prototype.

  • A focus on portability and convenience. I picture some sort of “Keychain” type device, small as possible, easy to carry, might not even need a FET on the output, runs of a small battery.
  • A focus on wearables. So runs a decent period of time without becoming burning hot. Maybe turning it off and on programmatically over say Bluetooth or in response to sensor data. Would probably end up with per project designs for things like the “Power Bracelet” that the project was originally targeting, along with some other ideas like the “Power Bra” and a “Power messenger bag” concept I have been considering.
  • A focus on output power. I am tempted to try make something stupidly over powered with way more thought put into power and heat management that I can run off the wall and lights my hand when I wave in its general direction :rofl:

Conflicted on this personally, I want it to be as accessible as possible to as many people as possible, but I am not interested in getting into the manufacturing or retail space so unless @amal for example is interested in working something out…

Regardless I plan on sharing any of my efforts will be shared for others to follow including the PCB designs, likely with one of the share alike licences. The only issue is that the skillset required to populate the PCB is not trivial and there are not more useable alternatives so not sure how useful the openness will be to the majority of the community.

1 Like

I’m not great but I’d love to play around with your ideas here too - any hopes for a list of the components and some schematics to mess around with on my end?
Or if you wanna talk kitting it out - I’m game to pay a premium for a diy setup down the line.

1 Like

Let’s wait until tomorrow evening when I get the antenna wired up and test the range with some LED implants. It’s likely there will be one more revision before they’re ready to share. I’ll post pics.


Yeah, once we know this dose in-fact work tomorrow, I will happily write up the BOM and schematic files etc but until it is tested I see little value in writing it up and even if I did, it’s getting late for me.

But 100% want to enable this so, one way or another I will upload the info on this design.

1 Like

@Satur9 wound a coil today and hooked it up. It worked. Using a vector network analyzer we found the LC tank circuit was likely off by a significant margin with a resonance of 2.9MHz so that potentially hurt the performance significantly.

This is the populated board:

We tested it with a xFD non implantable HF field detector and got rough measurements of the max range at different input voltages. We did this by holding the implant parallel to the coil.

At 12V the circuit was pulling 320ma and we could light the implant out to 7mm.
At 20V the circuit was pulling 380ma and we could light the implant out to 13mm.
At 30V the circuit was pulling 300ma and we could light the implant out to 18mm.

We also had a go with a NTAG I2C Plus test card and got a good 3cm range even with the non ideal coupling between the flat card and the cylindrical antenna.

Hoping that a better tuned LC tank pushes this even further but so far I am really excited about the potential of this new design.

One of the concerns I had with this was that the MOSFET (an IRF510PBF) would overheat. Each of the voltages we tested at caused the heatsink to slowly steadily heat up, and running it at 30v it got quite toasty during the minute or so we had it on for testing. It did however stay stable, both the output and the current draw not changing noticeably.
Given how large the heatsink I am using is already we do not think cooling a single MOSFET is feasible for this so I will be investigating running multiple in parallel to spread the load over multiple parts that hopefully can run more efficiently.

I also want to try driving the coil of the output of the circuit without a MOSFET and just see how effective that is.

I will also be thinking a bit more about how to share the schematics and other things in the most effective way possible, maybe uploading all the files to GitHub along with some documentation? Before I do that I want to do some work on the part footprints I made for this PCB to fix some mistakes I made before I share it. I also should likely add some notes to the schematic it’s self to indicate optional parts or how to pick the values of parts that do not list values etc etc.


I have a quick question: what does the signal look like thats driving the mosfet? Guessing from the amount of power dissipated in the mosfet it sounds almost like the mosfet is driven by a pure sine wave. This would at least explain why it’s heating quite a bit. Driving it in linear mode might give a good sine but the mosfet stays to long at high resistance and heats accordingly. More efficient is switch mode. You give the mosfet of steep of a square wave as possible. The signal directly from the mosfet looks horrible but the lc tank should filter it after into a nice sine. I know that you can pull 15w out of that mosfet without it getting hot if you drive it right especially on a heatsink like that.

To my project just an image so far as a little tease:

The mosfets finally arrived so i put one per type on the board but since the smd resistors are not here yet i didn’t go further from here.