using PWM to save power?

[stocky6409]

I havent got any gear with me at the moment and just brainstorming ideas

I have an app (using a 14M) where i need to have a number of LED indicators but wish to save power as well

Is it logical to say drive the anodes of the LEDS with a PWM signal and then pull the cathodes LOW when i want them to be "on" - pick a duty cycle to make the LED look lit but use less power to do so?

Or have i had too much coffee today?

 

[goom]

I think that, due to persistence of vision, a PWM driven led with a relatively low duty cycle will appear alnost as bright as one which is constantly on. As long as the frequency is above about 60Hz (I think), it will not appear to flicker.
Having said that, people do use a variable duty cycle to dim/fade LED's.
If visible flashing is acceptable, what about a 100ms pulse every second or so?

 

[moxhamj]

Absolutely - you can save a lot of power this way, and the picaxe PWM is so fast there will be no flicker either.

 

[stocky6409]

Thanks guys - something to work on tonite when i can get to my breadboard!

Stocky

 

[hippy]

I'm not entirely convinced.

If a LED + R is run on a static signal it will have a certain brightness. Change the static signal to a 50% duty PWM and its brightness is reduced. PWM reduces average current but so too would have increasing R.

Due to the eye's response, above a certain point, LED brightness often does not appear to increase or decrease as PWM changes. That ideal brightness could I believe also be achieved by using a suitable R.

The advantage of PWM then is perhaps in being able to control the average current used when the LED is on and tweak it for a desired brightness level, but I am sure exactly the same could be achieved by using appropriate resistors to start with.

 

[Dippy]

Trouble is that maths doesn't equal perception.
It's well worth a try trying various R values and dutys and your eyeballs.
Let us know the results of comparing 'DC' currents versus average duty currents.

 

[stocky6409]

I will let you know i find.
The application is fairly sensitive in regards of keeping the overall drain as low as possible.
I might be able to dynamically control the current drain depending on battery state by using the PWM....we will see

 

[Dippy]

If your budget can stand a few extra pennies then consider high brightness LEDs, maybe the clear types.
You'll be able to run them lower and have an immediate power saving.
Obv there will be other power-saving tips for the rest of the app too.

 

[hippy]

For maximal power saving; don't forget that turning on PWM may slightly increase current

I agree with using high brightness LED's. The bag full of unbranded 1980's LED's I have aren't particularly high efficiency but give a satisfactory output with 5V and 1K8, so about 2mA. They can be run on less and I'm sure there are more efficient ones around.

 

[BeanieBots]

Is it logical to say drive the anodes of the LEDS with a PWM signal and then pull the cathodes LOW when i want them to be "on"

I don't think there's a great deal more to say on the power/brightness issue (except maybe buck inductors, but that's getting excessive) but I would like to point out that the proposed method does result in the LED being reverse biased when off. Many LEDs are quite happy with that, but I have a met a few that die instantly when treated like that. Double check with the data sheet to see what the max reverse voltage is first.

 

[fernando_g]

If you still have leftover input pins (with ADC capability) in your circuit, use one of them to sense a photocell, and then adjust the PWM according to ambient light conditions.

 

[stocky6409]

Its a solar regulator so i sort of have a LARGE phototcell attached already
Once i get the cct tidied up a bit more i'll post it for comment!

 

[stocky6409]

Currently (!) the circuit has a self consumption figure of about 5.8mA with debug running and 5.3mA with debug disabled.

Playing with the LED dropping resistors and PWM I think i have a happy medium of brightness vs current draw - I estimate i have saved about 0.25mA over straight dropping resistor option but have yet to log a full battery charge/discharge cycle to calculate that over time.

I will be able to drop the draw a little more as the low voltage disconnect pin is only driving a LED at the moment instead of the actual low-side switching MOSFET.

All in all i'm not far off the commercial regulators approx 4mA self consumption figure - pretty happy really!