2-wire sensing for LED installation

[@shansoley]

Awesome idea. Perfect for existing installations where cabling has been permanently hidden behind plaster! Will be keeping this one in mind. Cheers Mike!

[@avejst]

impressive hardware and elegant feedback to your central processor! thanks for sharing your experience with all of us 👍😀

[@Schwuuuuup]

What I like with this videos is, that they are about overcoming limitations - but with limitations, that I still understand.... sure, the guys and gals working on the next gen Wifi or even Hackers trying to bitwise exfiltrate information out of a victim's PC do similar stuff, but quite often this goes so far over my head... or they don't make youtube videos ;-) But for this one... Powering devices over a LED power supply line below their forward voltage... this goes into the book of tricks, thank you.

[@ElectraFlarefire]

That really is very, very clever. I'll have to try and remember this for future projects. Played a bit with using ground/power lines of some addressable strips as the sense for capastive touch(Using a FET to disconnect it from power when detecting and back to power when running) but it only worked some of the time. Might be worth playing with using PWMed normal LED strips.

[@leosbagoftricks3732]

Very clever idea! Always nice to accomplish things without extra wiring!

[@aktik6000]

If you investigate MBus (Meter Bus) physical layer you see similar thing. Communication goes in one way by voltage modulation (typical serial with voltage offset) and response comes in form of current pulses, quite neat idea and pretty bus-lenght-proof solution👍

[@jaycee1980]

Rather ingenious! I guess stuff like Maxim's 1-wire stuff is employing similar games, but it's neat to see it done discretely!

[@lichtgestalt9540]

amazing idea in such an application! the feedback principle remembers me to one asic chipset, what i was designing around 20 years ago...

[@TheEmbeddedHobbyist]

Nice idea, bit like powering a device over the "one wire" system, but I could be wrong.

[@Darieee]

very nice .. everything. thanks a lot for sharing!

[@squelchstuff]

Very clever solution Mike.

[@Stabby666]

Really nice idea! As the MCU isn't doing too much, maybe a simple zener+resistor could replace the voltage regulator?

[@eliotmansfield]

very clever Mike

[@TheMorpheus017]

nice idea and implementation! thanks for sharing!

[@maxusboostus]

I don't see where the data is coming back to the sender, is it measured via a current sensor or by setting the PWM output to an input during the off time and measuring it that way?

[@BusyElectrons]

The data is returning to the system that controls the PWM via a 4th wire (so you have +24V, 0V, PWM, and the new data line). You can see this (unlabeled) data line in the lower right portion of the completed schematic at kzbin.info/www/bejne/boirZqd5g658mc0.

[@86abaile]

@@BusyElectrons Thanks, I'm pretty new to electronics and that clears things up for me.

[@BusyElectrons]

@@86abaile You're welcome. Glad I could help.

[@jtveg]

The 2 wires that go to the LEDs are the +24V wire and the other wire comes from the junction of the 15V zener and the top of the mosfet as seen in the diagram. The sensing comes from the current draw in the resistor in series with the 15V zener at the PWM supply (source end) of the LEDS. Every time the accelerometer activates it sends a series of current pulses which are measured as voltage across that series resistor. These pulses are smoothed by a cap so you get one long pulse as was shown on the scope. It appears that he then uses that long pulse to increase the pulse width and hence brightness of the LEDs.

[@agentblueuk]

Not sure I fully follow how the sensing feedback works, but really cool idea to get data back over you pwm lines. Idea it had seen elsewhere or your own?

[@miron__]

Funny that a temperature sensor example was mentioned, this is how some soldering irons do it

[@RyanJardina]

Where you been? I missed you

[@srpitu]

nice! Thanks for your work and explanation

[@markmathewson1501]

Absolutely brilliant!

[@RBRM3]

A quick video?

[@Spongman]

i wonder if you could do capacitive touch detection on the trailing edges of the LED power PWM signal for a completely passive solution?

[@PeregrineBF]

Yes. Here's an example: kzbin.info/www/bejne/hIq3aIinibl6ipI That one's posed as a magic trick, the demo of the touch detection starts at about 2:33.

[@bartomiej368]

About accelometer lock up, could that mcu onboard reset it?

[@mikeselectricstuff]

No, i2c interface dies

[@bartomiej368]

@@mikeselectricstuff i meant that you could power it via mcu pin, i guess it don't take much power.

[@andrewsweet43]

@@bartomiej368 I think he has already ran out of GPIO pins in this specific microcontroller in this application. Tiny little 6pin micros....

[@mikeselectricstuff]

@@bartomiej368 Would have done if I'd (a) discovered it before doing the PCB and (b) had enough pins

[@iamdarkyoshi]

Really neat idea!

[@davidf2281]

Is that an STMicro LISxxx accelerometer? I was about to order some of those myself, will have to look out for the lock-up problem if so.

[@mikeselectricstuff]

Nope -MC3413 - literally the only accellerometer I could find in stock at the time

[@dcallan812]

Very clever solution. 2x👍

[@robertw1871]

Elegant

[@mixolydian2010]

Very cool.Cheers

[@ebrahimhosseini1685]

So your are reading the data from Gnd line?

[@robertwilldesign]

Perhaps you could think of the strip as presenting three ‘states’ of load to the control circuit, LED on full load (data ignored), LED off with 10mA load = data high, and LED off with no load = data low.

[@realedna]

The zener diode and the 1K resistor following it to ground are on the control board IN FRONT OF the power outputs (plus and minus) to the LED strips. The 1K resistor then already is a current sensing resistor, where the incoming signal is read-off or processed/filtered as a voltage. So GND would be the at the SOURCE of the N-Ch Power-Mosfet, the LED (-)-output is at it's DRAIN and the signal comes from in between (from bridging the MOSFET with the zener-diode circuit). There are essentially 2 alternative pathways, 1 for the on-state (providing 24V to the LED strip) and 1 for the off-state (providing

[@Rich-on6fe]

Nice.

[@andymouse]

Neat ! ....cheers.

[@Xcite1231]

my brain hurts

[@AndyPevy]

Clever...

[@gotj]

Genial.

[@Tjousk]

Neat

[@tonyweavers4292]

I can't get this video to play? All your others do, how strange.

[@NivagSwerdna]

Very sneaky. I like it!

[@mark314158]

Very interesting. Please put up a vid of the final installation.

[@AdityaMehendale]

Wouldn't be a genuine MES video if the first sentence did not include the word "just" ;)

[@gsuberland]

This is a really useful trick. Inspired me to mock up a similar solution for AC, using a triac to apply a resistive load and an opto to detect the AC phase. Measured the current over a shunt, then ran it through a high-Q notch filter to pick out the specific frequency and reject load current noise. One thing I realised is that this is a perfect use-case for DTMF tone generators and detector ICs. The detectors have really high out-of-band rejection and the overlapping tones shouldn't ever occur under normal operation (unless the load is an analogue phone, I guess!).

[@miksu103]

Such an elegant and simple solution. Thank you for sharing your knowledge!

[@BersekViking]

Smart and simple. Love it.

[@jtveg]

Very clever and efficient trick. Thanks so much for sharing.

[@PatriotOnTour]

Da freut man sich doch schon auf den Ström heute Abend!

[@lrdiscovictim]

neat

[@fredflintstone1]

Fascinating idea and achievement !!

[@wthornton7346]

The art behind your art.

[@hippie-io7225]

Brilliant!

[@proluxelectronics7419]

Not enough blue status leds ;-D

[@aktik6000]

😁

[@TheDefpom]

Very clever, nice idea!

[@snik2pl]

Its similar to 4-20 sensing for powered sensors

[@dasrue6983]

Very cleaver! Reminds me of MBUS signalling

[@mdavidhandler]

an eloquent way to share communication on the power buss

[@ZeroStatic]

Great stuff as always, Thanks for posting. :)

[@nrdesign1991]

very clever!

[@made2glow516]

Neat idea!

[@tekvax]

Very clever indeed!

[@dufflepod]

Nice.

[@SeanBZA]

Elegant idea, how about a resistor is series with the input diode to the sensor board, around 100R, to get those turn off current pulses down. Should allow you to use lower filter values, and thus get a slightly faster response, or at least get better resolution for the sensor output. As you still have one pin left on the microcontroller, simply use it to power the accelerometer module,so there will be less issue with brown outs. Use the micro power on reset to get stable voltage rails, and wait for a half second after POR before enabling the second chip.

[@jaro6985]

He said below the brownout issue was discovered after building the board. Its not really something you expect to see IMO.

[@ab_ab_c]

Can you provide us with the specs & a url to your motion sensor board?

[@kain0m]

Judging by the video, it is a custom design.

[@mcgdb3606]

does the feedback signal. go back into the dimmer to adjust how it's driving the power mosfet switch?

[@daShare]

Yes, it would do. He didn't go into that as the real magic was in the LED interface.

[@mcgdb3606]

@@daShare thanks for the clarification

[@felixcat4346]

Thanks for the post 📯.