Q: how find specific LED

Maybe "scanning" would have been a better term. One section of the design has (8) LED's. Instead of having them all on at once, I scan through them real fast so that only one is on at a time. Thus, 1/8th the power.

But, I could also blink them. Instead of being constantly "on" (to the eye), I guess I could flash them too, but that's going to be confusing to the user, and probably violates the product specs.

On Wednesday, March 18, 2020 at 11:37:23 AM UTC-4, snipped-for-privacy@downunder.com w rote:

OK, Now you're just getting medieval on it. :)

I didn't mention it, but our LDO that powers that part of the circuit gets close enough to its max I-out limit that we avoid throwing all the relays a t once. Even with latching relays, that keeps peak coil current demands lo w enough to squeak by / justify, etc...

This particular build is a pain in the ass.

Unrelated to this thread, but maybe touching on your comments about automotive LED taillights, have you ever noticed that traffic lights sometimes have multiple LED's out at once?

It got me wondering whether the designers had enough forethought when they designed the circuit board to make sure that if one or more of the LED driving circuits fail, you don't end up with a valid-looking visible traffic light:

For example, a green, round "Go" light, that unintentionally morphs into a right or left turn arrow, if any of the LED drivers fail. Yikes!!

Thankfully, all the ones I've ever seen do this (a population probably less than 20), the "missing" LED's were scattered about.

mpm wrote in news: snipped-for-privacy@googlegroups.com:

Their nature is to 'fully' conduct once 'on'. However, their nature is to also self destruct if the current is too high because there is an internal drop and thus 'working voltage'.

Pulse width modulation is used because that way the diode can be switched on and emit in its ideal operating range, and yet provide the user with a reduced light output by cycling it on and off and using 'duty cycle' to govern the resultant "level" of output on the lumens side. The waveform matters. Turn on, but turn back off before it has a chance to get to full on brightness, AND space out segments of that method. So short on pulses with adjustable wait space. At that point the short legnth on pulse can have its width adjusted wider. And thus one has multiple points in a circuit in which to affect an adjustable output. Most flashlights have fixed modes which have such set points pre set. It looks like "half brightness" until you use it on your bike at night and the street lamp syncs in and out with its flashes.

Trying to do that with flat DC control on a device which like to conduct real good when on? Not nearly as easy and a lot of blown units later and the entire planet moved to PWM modes.

But fully current limited, and you should be able to get good life out of it. Read its spec, and uhhh... follow it.

mpm wrote in news: snipped-for-privacy@googlegroups.com:

1/8 the light as well.

All 8 on is bright.

All 8 sequenced is one on with 8 times the shelf life, but still only one on at any given moment, and they have no afterglow, so...

Ever see a cosmology guy on Youtube talking about the stars?

His show is called "What Da Math"

It is funny he says it so fast you can;t hear it all.

mpm wrote in news: snipped-for-privacy@googlegroups.com:

Use an opto-coupler to have a "monitor switch" on the coil without loading, and run a little LED bus down through the gang of optos and relays and source the tiny amount of power it would need for the LEDs...

Out of thin air! Yeah... that's it. It really can't be that much. Fire 'em directly. Microamps, man... really.

Have the relay firing charge up a cap or supercap diode isolated from the coil that then fires the LED and has persistence that does not require additional power for a given period.

Now THAT sounds pretty cool, IF you do not have too high a repeat rate. You said latching, so... latching LEDs!

I too remember seeing that and found it in the HP Optoelectronics Applications Manual from 1977 ISBN 0-07-028606-X in section 2.3.4

The key fact is that LED intensity is actually slightly supra-linear with forward current. Exponent varies from 1.1 to 1.4 depending on LED - but 40+ years on that may have changed a lot?

piglet

Yes, it is a little bit dimmer, but not objectionably so.

on

he

re

LEDs were quite bright. I dimmed them by duty cycle modulating at about 2

0%. I think it was faster than 200 Hz, but still... I would find this a h ard concept to believe unless there is a reference somewhere. I can't thin k of a mechanism in the eye that would provide for perceived brightness bei ng the same on a duty cycled light but only if the flash rate were in some specific, small range.

mmed exactly this way, but pulsing them. I find it immensely irritating wh en I see a dozen tail lights as I move swing my view across the road in fro nt of me. I expect this flash rate is somewhere around 100-1000 Hz and is being done specifically to dim the lights.

s any such effect making a blinking LED appear relatively brighter than a s olidly on LED at the same power levels. In fact, there should be some effe ct, even if small, reducing the relative brightness at high current levels.

I can't find an accessible copy. Mostly this is in paper form.

Thanks anyway.

--

  Rick C. 

  ++ Get 1,000 miles of free Supercharging 
  ++ Tesla referral code - https://ts.la/richard11209

e

sion

the

here

he LEDs 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 th ink of a mechanism in the eye that would provide for perceived brightness b eing the same on a duty cycled light but only if the flash rate were in som e specific, small range.

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 f ront of me. I expect this flash rate is somewhere around 100-1000 Hz and i s being done specifically to dim the lights.

is any such effect making a blinking LED appear relatively brighter than a solidly on LED at the same power levels. In fact, there should be some ef fect, even if small, reducing the relative brightness at high current level s.

-

You can find it on Scribd.

which is subscription based, but well worth the money.

Link:

The section 2.3.4 essentially claims that the human eye is a time averaging detector, and that the time average luminous intensity is what matters - a nd it may not match the area under the curve specified in the datasheet. ( but will be typically close).

e:

I e

ere

t

ession

ng the

ewhere

t

the LEDs were quite bright. I dimmed them by duty cycle modulating at abo ut 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 the same on a duty cycled light but only if the flash rate were in s ome specific, small range.

n dimmed exactly this way, but pulsing them. I find it immensely irritatin g when I see a dozen tail lights as I move swing my view across the road in front of me. I expect this flash rate is somewhere around 100-1000 Hz and is being done specifically to dim the lights.

re is any such effect making a blinking LED appear relatively brighter than a solidly on LED at the same power levels. In fact, there should be some effect, even if small, reducing the relative brightness at high current lev els.

D -

, but well worth the money.

Can't read that of course.

ng detector, and that the time average luminous intensity is what matters - and it may not match the area under the curve specified in the datasheet. (but will be typically close).

Not sure what you mean by "may not match...", but if "the human eye is a ti me averaging detector" then running the LED at 8 times the brightness for 1 /8 the time will accomplish nothing in terms of saving power and in fact wi ll waste some small power in the control mechanism.

--

  Rick C. 

  --- Get 1,000 miles of free Supercharging 
  --- Tesla referral code - https://ts.la/richard11209

use inductors instead of resistors, that will get you increased brightness at the same supply current, but it's probably cheaper to use high-effiency LEDs in most cases

--
  Jasen.

use latching relays with visible state indication, use no LEDs

--
  Jasen.

Interesting idea. I'm not aware of any inexpensive (or otherwise?) small signal -type relays that fit that bill. The max it would save is ~ 700 uA x 8 LED's. Not enou gh to mess with (yet). :)

I recovered another 6 mA in the power budget by NOT letting the 5VDC-to-3VD C isolated regulator run full time. It's only needed when it's needed (abo ut 5% duty cycle), and it turns out I can live with the start-up time. So, we'll switch it on with a FET under microprocessor control.

With the present numbers, I think we can avoid some of the more complicated software power-saving routines and be OK. But that isolated supply was re ally the last of the low-hanging fruit. Further cuts would become noticeab le to the end user (but might still be acceptable if it really comes to tha t.)

Right now, the power budget (simulated) has us at about 29 hours (need 24). This product will go in for UL approval, so I'm sure they're going to put a stopwatch on it.

Check out a better comparison chart; no nonlinear CDS photo resistor; used a LED for detector for better linearity. See attached Red5mmLEDs.xls for info.

Have on order the highest intensity CREE red LED available (C503B-RAN-CA0B0AA1) and the highest intensity Kingbright red LED available (WP7113SEC/J3) at Mouser.

Thanks

Extended chart:

Item Mfg Part Comment

2 CREE C503B-RBS-CY0Z0AA1 STD 10 ROHM SLI-580UT3F STD 12 VCC UAOL-3GRE4 HIGH 11 TT ELECT OVLLR8C7 STD 4 CREE C503B-RBN-CW0Z0AA1 STD 5 CREE C503B-RBN-CW0Z0AA2 STD 6 CREE C503B-RBS-CW0Z0AA2 STD 3 CREE C503B-RBN-CY0Z0AA2 STD 7 CREE C503B-RBN-CX0Y0AA1 STD 1 LITE-ON LTL-307EE RED lens 9 CREE C503B-RBN-CX0Y0AA1 PINK lens 13 CREE C503B-RBN-CX0Y0AA1 RED lens 101 CREE C503B-RAN-CA0B0AA1 STD 102 KingBright WP711SEC/J3 STD

Driver LED d1 comparisons with LED d2 detector measurement Adjustable 20V supply in series with 750K resistor and d1 DVM used to measure d2

d1 d2 Detector voltage

2 2 1.335 10 2 0.324 12 2 0.396 11 2 0.441 4 2 1.239 5 2 1.287 6 2 1.297 3 2 1.084 7 2 1.237 8 2 1.210 6 6 1.204

New detector setup d1 d2 Detector voltage

2 2 1.326 2 101 1.263 2 102 1.283

So, luminance info is not primary factor for detector sensitivity. Seems I stumbled on one of the better CREE LEDs for the detector.

The CREE C503B-RBS-CY0Z0AA1 (standard LED, #2) 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!