In general you don't want to leave unused inputs floating in either analog or digital circuits. It's inviting unintended things to happen. In the case of an opamp, the floating inputs will be very susceptible to noise and the output may flail around and even oscillate. This will not only put noise into the other circuits, but will consume power. For the same reason, don't tie the inputs together. Tie one high through a large resistor and the other low through a large resistor. That will put the opamp into a known and static state. Depending on the details of the opamp, you might get slightly lower power by switching which input you take where. Measure and see.

 

FYI I didn't get Tracecom's breadboard photo and info until tonight. Even with the lag in the forum I really appreciate everyone's feedback and it is a big help and makes it more fun.

 

I believe Chazmon3 is finished with this thread, so I am going to post another LM324 LDR PWM controller circuit that I have found, modified, and breadboarded. It runs on 12V (which I prefer) and uses all four of the op-amps. It seems to be capable of finer control resolution than the previous circuit.

Comments are welcome. Thanks.

 

That is good to know about the unused pins on the IC and will pay heed to it. Should there be more precautions in the unused pins in the circuit below?

This circuit is from David Cook's site http://www.robotroom.com/PWM.html and is the one that is 93% efficient. The only difference is that I used a 74HC14 because that is what I had on hand (should get the 74AC14 in the mail today). From pin 4 powered a IRF510 MOSFET through a 100k resistor to the gate.

Well the 74AC14 did arrive and when I put it in place of the 74HC14 I found that it was way less efficient by almost 50% plus the out put was also about half. Was lucky that I used the IC had here rather then the one that is recommended at least for my application. Experimentation is worth while!

I used a half watt 10Ω resistor for the LED's so replaced with a 1 watt and raised the efficiency to 96%. So on that note I will be experimenting with different resistors plus any other way to make it more efficient.

 

Nice circuit Tracecom! I am actually going to be powering several led's with a Li ion battery so the voltage will be from about 3-4 volts or if I use two batteries then can double that voltage range, so 12v is outside my limit. It is great that you use all four op-amps and get a better resolution for adjustment.
Chazmon3

 

If you drive multiple LEDs, you should have a resistor in series with each one. The reason for this is you want the currents to be matched as closely as possible. With multiple LEDs, it is likely that not all will have the same forward voltage at the desired current. When you put LEDs in parallel, the one with the lowest voltage will tend to hog more current. If they are dissipating significant power, it will generally be the hottest, and since Vfwd has a negative temperature coefficient, it will tend to draw even more current as it heats up. In the extreme case, this can lead to thermal runaway. In your case, it may just result in unequal brightness.
You can get away with only one resistor if you select matched LEDs, but this may not be possible unless you have a lot of them.

 

Yes that is an great point on the resistors. It would be nice to be able to power the LED's in series, but a boost circuit is way to watt expensive.
I am just now doing a test with an antique rheostat (top rh side of photo) that is made for 5 amps, and is adjustable from 0-18 ohms. When I make the final circuit will use one resistor per LED plus the low wattage for the resistors are low cost and easy to get.
Got to love this old rheostat! Will post results of my tests later today.

 

Here is a graph (on grid paper) I made and with a diagram of the way I hooked up the meters. I am surprised that it goes nearly to 100%. Maybe its not the right way to measure efficiency or just the nature of Harbor Freight meters. But all and all think this is the best circuit to use for my purpose and it should be over 90% efficient with the power levels that will be used. Of course will follow Ron H's advice to use one resistor for each LED, so a lot of work to be done on it yet.

 

Hey I am back. No I am not finished with this thread, but if no one adds to it I guess its dead ha ha !
Liked Tracecom's breadboarded circuit and found another configured just like it with the exception of some different resistors and not with the light sensitive resistor LDR, plus its powering a motor instead of a LED. This circuit runs well on one Li Ion battery http://www.pcsilencioso.com/cpemma/pwm_erg.html
This is the type of circuit that I want to use and would like to control it with two LDR that would give an operating range of light that they would be exposed to. It would be a regulator of sorts. Anyone have any suggestions?
Charles