I was thinking, a while back, about some sort of oscillator to make an LED blink visibly from very low average current. My application was to minimally load a high voltage supply but warn of voltage present. I didn't come up with anything profound.
It's an interesting problem. One possibility is to have a ground-based clock tease an equivalent schmitt once in a while, so it doesn't use the current it would otherwise take to fire. But that needs a sustained low voltage supply!
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
un
I don't know about the high voltage, but the idea of being current/energy e fficient and providing a low duty cycle would be ideal for a unijunction tr ansistor. The capacitor charges up at a low current, then when reaching th e trigger level the cap and additional current is dumped into the LED. Not hing is wasted.
I believe an SCR can be used to create a similar circuit. The only real is sue is the mismatch between the low LED voltage and the power supply voltag e. Adding a dropping resistor loses a lot of energy. Maybe a circuit to c hop the voltage and use a capacitive dropper is the way to go.
-- Rick C. - Get 1,000 miles of free Supercharging - Tesla referral code - https://ts.la/richard11209
Haven't tried this, but might be useful for that application: Use a STN0214 in a modified joule thief circuit; add a capacitor between the base and emitter to extend the off time.
The issue becomes, how much base current does it need to fire? Possibly nanoamps.
And how much average power does a good LED need to be a visible blinking warning? 100 uW? That would, in an ideal world, need 100 nA from a kilovolt power supply.
-- John Larkin Highland Technology, Inc The cork popped merrily, and Lord Peter rose to his feet. "Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"
my suggestion is an avalanche transistor))
The CREE C503B-RBS-CY0Z0AA1 (standard LED) is definitely visible (in relatively dark room) at 0.16uA, supply 4.95V, 22 Meg resistor in series. That calculates to 1.43V across the LED!
One wonders what one of their "better" high intensity LEDs would perform...
Define "good"; the CREE LEDs seem to be able to approach that ideal (see my info previously posted). Also, what would one like for "visible"? The eye sees light pulses apparently as well as if LED was on constantly; rather helpful in this app. So, maybe a pulse of 250uA every second could keep the power drain down and still show existence of HV; rather bright for he CREE LED.
Thanks
I'm not sure why you say that. Many car tail lights blink the LEDs rather than dim the current between tail light and brake light modes. I know this for a fact because I can see it when I turn my eyes. I find it very distracting and annoying.
-- Rick C. + Get 1,000 miles of free Supercharging + Tesla referral code - https://ts.la/richard11209
r than dim the current between tail light and brake light modes. I know th is for a fact because I can see it when I turn my eyes. I find it very dis tracting and annoying.
I'm working on a rather large project right now where I will have to blink the LED's in order to meet the power budget. If that were all I had to do, then OK.
But Nooooooo. (That would be too easy!)
I'm also going to have to tri-state most of the port pins when not actively using them, constantly switch to the low-power internal oscillator (with c lock dividers active) during time-insensitive subroutines, use idle mode, u se power-down mode, use external FET's to control things like resistive por t pull-ups, and tweak the hell out of the code to do as much as possible wi th as few clock cycles as possible. Every FET is very low RdS, and regulat ors are high-speed switchers and LDO's, with on/off control. And of course , all the relays are dual-coil latching types, so it messes with your head a little trying to keep track of their states. (Not difficult, just can be confusing.)
In short: The "Full Monty" of tricks.
If I could have about 10-15 mW more in the budget, I could probably ditch a bout 2/3rd of these complications! OK, maybe 20. :)
Unfortunately, the lead-acid battery is the determining factor (both capaci ty and physical size, and I guess re-charge time - though that is not spec' d), so I'm stuck.
When my post actually loads... We're using those same CREE LED's in the project (about 16 of them), all driven at around 710 uA.
(...and I might have to reduce them a bit more) But, plenty bright at 710 uA. (Both red and green types.)
I wonder how low a supply current could be achieved with the classic
2-transistor (or mosfet) astable multivibrator. Nanoamps? I guess that could be Spiced, given a proper mosfet model.There may be an SCR/PUT type configuration with mosfets, or even bipolars, that would fire at very low currents.
-- John Larkin Highland Technology, Inc The cork popped merrily, and Lord Peter rose to his feet. "Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"
How does blinking the LED reduce power consumption compared to reducing the current? Seems to me it is a constant power/energy situation when taking into account the power/energy in the current limiting.
-- Rick C. -- Get 1,000 miles of free Supercharging -- Tesla referral code - https://ts.la/richard11209
A good LED puts out more visible light per mA than a bad one.
-- John Larkin Highland Technology, Inc The cork popped merrily, and Lord Peter rose to his feet. "Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"
Since you are talking about latching relays, it sounds like sometimes some high power will flow through your device. Why not steal some power from that circuit to power the ON indicators ?
Regarding knowing the state of a latching relay, decades ago I made an optoisolator consisting of a neon lamp on the mains side and a photo Darlington feeding some CMOS logic.
You could use two neon lamps on the high power side (possibly both in series for 230 Vac and above) and use one for the feedback and the other as an ON indicator. With a very low current LED, you could even put it in series with the neon lamp. If neon lamps are out of the question and the high power is AC, you could use a capacitive voltage divider to feed an ordinary LED optoisolator and indicator LED.
Blinking an LED fast ( > 100 Hz) but not too fast can give an impression of higher brightness to the human eye than you'd get with regulating the current. I haven't confirmed this in detail, but I have heard somewhere that a 20% duty cycle at, say, 200 Hz gives the equivalent apparent brightness of 100% duty cycle with half the current.
So no reference???
I screwed up a board design using resistors that were too small and the LED s were quite bright. I dimmed them by duty cycle modulating at about 20%. I think it was faster than 200 Hz, but still... I would find this a hard concept to believe unless there is a reference somewhere. I can't think of a mechanism in the eye that would provide for perceived brightness being t he same on a duty cycled light but only if the flash rate were in some spec ific, small range.
In fact, as contrary information, the tail lights of autos are often dimmed exactly this way, but pulsing them. I find it immensely irritating when I see a dozen tail lights as I move swing my view across the road in front o f me. I expect this flash rate is somewhere around 100-1000 Hz and is bein g done specifically to dim the lights.
So the more I think about it, the less I am inclined to believe there is an y such effect making a blinking LED appear relatively brighter than a solid ly on LED at the same power levels. In fact, there should be some effect, even if small, reducing the relative brightness at high current levels.
Do you have any references other than "I have heard somewhere"?
-- Rick C. -+ Get 2,000 miles of free Supercharging -+ Tesla referral code - https://ts.la/richard11209
I think that 100nA for bipolars in a multivibrator would give a very stable and repeatable operation; besides, 100nA will nicely light up a CREE LED.
I might try to find one if I get the chance. But it is certainly a common technique.
The retina cells react stronger to a change in light intensity. So the initial part of the flash causes a stronger reaction than later parts. Thus the blinking looks brighter. But if the blinking is too fast, the cells don't have the time to "reset" between on-times, reducing the effect. The effect is biggest with slower blinking (though the higher level processing in the brain, which emphasises changes and movement, has a bigger contribution for very slow blinking - that's why bike lights often blink, as I'm sure you know). But below about 100 Hz and you can see the flickering if you move your eyes, and that gets annoying.
Of course you dim them that way - it is the simplest and most efficient method. You just have to reduce the duty cycle more than you would otherwise expect from a simple linear calculation.
Real life practice.
I can't claim to have done any measurements or scientific comparisons, however.
My HV warning problem would blink the LED pretty bright, at 1 Hz maybe. All sorts of wild LED pulsers are possible, but I wanted to make the basic 1 Hz oscillator use very little current.
We could experiment. I'm guessing a 50 millisecond light pulse, maybe a bit less, would be a good investment in current.
-- John Larkin Highland Technology, Inc The cork popped merrily, and Lord Peter rose to his feet. "Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.