Inverter to make 110v LEDs NOT flicker?

I was pretty happy when I found my first LED Rope Light. I.E. reasonable price, found on the shelf at a store, etc.

Then, as someone who lives with someone who says "no fluorescents - too much flicker', I was "horrified" to find that this rope LED light flickered like crazy. Every time you moved your eyes, you'd see it flicker.

Then again, "of course"...when you saw your first LED Caddy tail light, it was so "crisp" - turning on and off so QUICKLY.

OK, so

How about a "box" with a 110 (or, 117? 120?)V plug, and a receptacle, into which you plug the rope light.

The "box" contains a power supply producing DC, then an oscillator to turn it into -- what, 400Hz? 30,000Hz? (would an "audible" Hz possibly be heard - like the opposite of "microphonics", .... well probably not if there is nothing magnetic associated with the output circuit... )

Some fundamental electrical questions: Not the safest, but could it run without a transformer? I mean, the rope light is plugged into a "live" circuit anyway...

But I don't think I'd want to put a rectifier bridge on the "mains" directly - well, maybe through a fuse?

_________
Starting over: (haha). I'm thinking of a circuit:
- Fuse
- isolation transformer (1:1) - this is what I'm asking if I need
- Bridge
- Filter cap
- Feed to a pair of ?? (advice sought) MosFets, producing an output square wave
- Feed the (400Hz?) square wave to an outlet, clearly labeled "for LED Rope use only".

-------
The MosFets of course would be part of an oscillator (I believe I could find a circuit) - and no, I don't know a colpitts from a multivibrator.

_____
I can start by either checking the specs of the rope(s), or putting my wonderful "Kill-a-watt" (meter) on it, to see what it draws, and "select components" (*) based upon that.

(*) I am a "4000 foot guy". I understand concepts of a lot of things - this is one of those. I have never used a MOSFET, tho I know I'd have to spec its voltages (at least 200V maybe more) and current (probably < 10W thus like 90MA) etc. So I could use some pointers to "beginner" (but who understands the fundamentals) help.

Thank you all so much.

P.S. by background, I have taught myself digital electronics back in the '70s, then designed my own microcomputer circuit cards, and got a selectric and disk drive hooked to my Altair; and now I do PIC programming, but I have had *no* experience with anything but digital circuits.

Ward Christensen
("Xmodem & BBSs")

Wow Ward, I remember you from my old BBS days and Crosstalk.

How many LEDs are in a typical chain? Are they 110VAC rated LED ropes (like Christmans lights)?

Jesus, Ward Christensen? It's me, Duffy.

They may have half the LED's connected one way and half the other, so each set is doing a half-cycle. If that's the case, then going to a higher frequency may be your best bet - unless you can get into the rope, separate the strings, and add a rectifier and capacitor to each chain.

I bought an LED rope light recently, and it flickered a lot.

Here we have mains at 240 V 50 Hz and the LED rope light came with a transformer, with about 24V AC output. The rope contains a bridge rectifier, but no smoothing capacitor. There are 10 or so LEDs in each string, and the LEDs were only on about half the time, but they were flashing at 100 Hz.

I couldn't get the LED rope open without destroying it, so I added a bridge rectifier and a smoothing capacitor inside the transformer. The flickering stopped.

It isn't necessary to get the ripple very low, if all the LED strings are the same length. As the capacitor discharges, the voltage across the series resistors falls, so there is less current, and the discharge slows.

As long as the minimum current is more than about 25% of the peak, the flicker won't be noticeable.

It doesn't make any difference whether there is a transformer supplying it. You can simply rectify and smooth the mains. You might want a small resistor to limit the inrush as the capacitor charges. Alternatively, you could just take power from the capacitor of a switch mode power supply of the right power rating.

I would guess that the lights aren't supplied with unrectified mains, or they would flicker really badly. However, if the string is wired with some LEDs only working when the current flows one way, and others only working when the current flows the other way, you will have to use some sort of inverter to switch the current forwards and backwards very quickly.

I hope you all realise that 100 Hz flickering isn't visible. If it were, cars like the latest VWs wouldn't be legal..........

The Road Vehicles Lighting Regulations 1989

The photo shows a VW tail light, a street light and a conventional sidelight, all taken with deliberate camera movement. The street light is powered at 50 Hz so it is modulated at 100 Hz. The LED is also flashing at that speed, but much more obviously because it is off for more time.

I can detect LEDs flashing at up to about 600 Hz, depending on conditions. The flashing LED tail lights a real pain.

Thanks. Bill, Duffy(!!)..!

Yah, LED Rope Lights are made of strings of -|>--|>--|>-- until they work across 110 v and do their OWN half wave rectification, and a parallel set going --<|--<|--<| how ever many it takes. Let's see, white or blue LEDs go about 3v "let's say", so do we do the RMS? (120v) and assume (test?) their breakdown for the 1.414* = as much as 180V.

But wattage like I say is pretty small. I guess I'd like to make a gizmo that goes 50W so I could do several (or putz with without trashing it)

Take Care!

"AC" vs "DC" rope LEDs

Diver: Very interesting, that you found a bridge in the rope light!!!

I will have to investigate mine further.

Mine going to the 120v, could be it has a rectifier also, I guess I'll have to poke about some, thanks for your thoughts!

I couldn't actually open up the rope as it was all potted. Also it is supplied with a transformer which makes fault finding safer.

I found it was supplied with 20 V ac, and it flickered, but the lights seemed to be on about half the time. The supply has to overcome the forward voltage of the LEDs before they light, and the ac is a sine wave, so the LEDs would only light about 1/4 of the time if it was half-wave rectified.

If the peak voltage is Vp and the LED voltage is Vd, the duty cycle is arccos(Vd/Vp) / 90 if it is full-wave rectified.

I estimate that the diode voltage was about 70% of the Vp as the rope wouldn't light at all with a 12V supply. That's consistent with the LEDs being on half the time.

Then I supplied the rope with about 25V dc and found that it lit either way round. That means that the rope must contain a bridge rectifier.

To add a capacitor, I had to add it after a bridge rectifier. Not being able to open up the rope meant that I was stuck with the second one as well, but four diode drops instead of two isn't significant at 20 V.

Two bridges to cross

Diver300 you're very clever, I hadn't thought of just ignoring "perfection" (of only having one bridge).

And I had to read a 2nd time to see that you of course don't get illumination for the full half wave:


I wonder if LEDs have hysteresis like so many other things do²

Thus my diagram - on at red, on for green, off at 2nd red -- or PROBABLY sooner, as I expect LEDs have no -- or little -- hysteresis.
____________
(1) This Footnote intentionally left blank

(2) lower extinguishing voltage than igniting, e.g. Neons, or dimmers³

(3) I have to crank my bathroom dimmer up rapidly then back down, to obtain a middle-of-the-night dim light when it has been off

Thanks!

I don't think the LEDs have hysteresis.

I think duty cycle is important, the higher the forward voltage, the lower the duty cycle which will make flicker more noticeable.

Duty Cycle

<soh>
Hero999: Thanks!

I wish I'd write this stuff down - Actually I do pretty well - using Freemind from Sourceforge.... Specifically: That I once saw a nice graph/table of brightness - vs duty cycle AND amperage. (family of curves, not 3D)
(There's already a thread on this: https://www.electro-tech-online.com/threads/led-brightness-vs-freqeuncy.14152/ so maybe no need to get into details here)

In the case of LEDs it seems you can "zap" 'em with an incredibly high voltage, as long as the duty cycle is short enough, and thus obtain the most light for the energy expended.

THAT is why I'd still rather make a little AC to DC to AC power supply, which I could just "plug the rope light into" (or any other things like LED xmas tree lights, etc).

I am very familiar with programming PICs, and thus could -- with a couple pushbuttons, create a way of changing the duty cycle up/down. I'm not sure how I would control voltage with a PIC (meaning I could search / figure it out)...but for openers, stick with 110v, and just change the duty cycle.

I was going to say I first need to determine if MY rope light is "DC" (all LEDs one way, possibly a bridge) or if it was what I assumed namely that half the LEDS light on one cycle and half on the other cycle (though this would I think make more obvious flicker?)

I say "was" because the best bet would be to just have a programmable duty cycle for "either or both" sides of the output.

HAH! I've just defined a "dimmer" for my rope light - as the duty cycle approaches optimum, that's brightest (by definition) and as it approaches 0, that dark. haha.

I think when I have time, I'll put my rope light on an extension cord, find someplace where I can swing it around and -- when dark -- see just what the pattern is.

\/\/ <eot>

P.S. I have a nice little power supply with a voltage and a current knob.

I can thus stick a non-resistor-limited LED across it, crank the current down, bring the voltage up, until it lights, if no light, then V back down, I up a bit, try increasing V again, etc.

SO I find that if something (say a white or blue LED) will illuminate at 2.8V (not measured, just what if), WHAT HAPPENS when you stick it across a 3V lithium cell? There is no current limiting resistor, so I presume the current is some kind of formula relating to the ... what? Internal resistance of the battery? Another way to ask: Suppose an LED supports 3-5V, and I put it on 5, with no resistor; I assume it will be very bright then self destruct?

P.P.S. I am a big fan of the "Joule Thief", and have one in the bedroom where I keep bumping a corner. It runs "weeks" on a DEAD LR41 cell, i.e. one that no longer works as one of 4 in a PIC project, and I suspect MONTHS on a "dead" AAA battery (I see no reason to give it a "good" or "new" battery since it runs so well down to perhaps .8v or a little under. THEN -- haha - I might put two "dead" LR41's in series etc.

The lithium battery has a high internal resistance and will limit the power.

When an LED dies from over current, it's wavelength stretches first. Sorry to say I've watched this happen a few times - a green LED turns yellow, a yellow LED turns red, I assume a red led turns infra-red.

Who are you living with that's flicker-sensitive?

> Who are you living with that's flicker-sensitive?

Apparently some of my S.O.'s girlfriends can NOT go 'shopping' due to the fluorescent lights.

I think HER main concern in our house is the lack of "warm white" as well as the flicker.

I went away one time and came back and all the CFs had been changed back out to regular bulbs.

I still use CFs in any area exclusively mine.

What will she do when they phase out regular bulbs in the U.S.?

Ward, did I ever tell you about that time I tried to make the big strobe light out of the fluorescent tube on the frisbee project? It didn't blink at 60hz. Didn't even flicker.

If you want to prove this to yourself (or your friends) take a mirror, look at the reflection of a fluorescent light in it, and tilt it back and forth quickly. No variation in the light. Now compare it to a neon - bunch of short orange dashes.

There's no flicker on fluorescent tubes, not even CFC's. They use very high-persistance phosphors.

1. I have a string of LED lights from xmas out on the deck; if I move my eyes back and forth rapidly, from 20' away the flashes look about 3" apart.

2. Duffy, moving a mirror - isn't that cheating? Aren't you supposed to move your head fast until you get silly? Oh, THAT'S my problem! haha

But seriously: * placed a rotating object (a paint stirrer was what I could find) in my variable speed electric drill. Observations:

1 - torchier (Halogen light) - just see a blur,

2 - CF: Same as torchier. "perhaps" the slightest bit of "strobing", I should construct something more obvious than just what was at hand.

3 - Fluorescent lights - I can see little slightly "brown" bands as it spins - these must be the "Dark places" and the F dimms between cycles. To prove my point, I estimated the RPM (say, 6-7 per second) and the # around the circle (oh, 20?) so 7 x 20 = 140, 6 x 20 = 120, "proof" that I was seeing an effect of 60 Cycles (dimming observably, 120 times/sec).

The phosphor IS long persistence, but a simple visual test can show it is NOT uniform, I believe it would take a light sensor approximating the human eye, and an oscilloscope, to see the actual wave form - estimating 80% valleys in the 100% tops.

I will have to have a discussion with her regarding CFs. Cheap CFs are the wrong °K, so I need to find some "warm white" ones, or even daylight - if possible. I think I like / am used to / the yellowishness (of the sun, of incandescent bulbs) etc.

Thanks for these great chats.

Just like old times.. (sniff)

Hey, Ward and Duffy, messing with LEDs like in the Olden Days.

Duff, I don't remember *any* frisbee project, much less one with lights...

Ward, are you looking for a "painless" solution for the LED rope lights (i.e. one that does not involve cutting into the string anywhere)? Then you may have to go with the higher-freq inverter plan. However adding appropriate filter caps should also work. Requires surgery though.

"Newbie" indeed....

The "Luminator" project? You don't remember THAT? The one that exploded from high RPM's on the pancake motor? The one that cracked in half on the sidewalk at the feet of the guy from Toys Я Us? This is Ted, right?

It is Ted(!) No.. uh-uh nuttin' about no Luminator. Wish I'd seen it though, always good to watch demos that explode. Toys Я Us was suitably impressed? About when did this take place?

Hey.. Ward? We should turn Duffy loose on some of the Blinkie code. In.. you know.. his Spare Time of course.

Late 80's, I think. Used the pancake motor out of one of those Pioneer laser disc players to spin it in the display for the trade show - got it to 100,000 RPM and it blew itself to hell, threw the batteries right through the cardboard.

What's the Blinkie code?

duffy said:

got it to 100,000 RPM and it blew itself to hell, threw the batteries right through the cardboard.

Click to expand...

Let's focus on what is important: any deaths?

Ah... those old LD pancake motors. Starting to remember.. a little. Did you have to build a custom drive to get that kind of RPM out of the motor?

Blinkie code is here 2DKits.com, 2DKits.com Home of the Blinkie I thought I just posted this, but now it seems to not be here. You'd find them infernally boring probably. But they DO speak RC-5. They are mostly to have a cheap board and some through-hole parts for beginners to learn how to solder on.