If each LED has a 100 ohm resistor and is fed via 5 volts then the individual LED current is probably around 35mA. (Based on 1.5 volts LED forward voltage leaving 3.5 volts to be dropped across the 100 ohm resistor) So 28 LEDs will demand say 980mA.
To reduce the LED current using a single pot' (wired as a 2-terminal rheostat) you are anticipating that the individual currents will balance out (which they may or they may not - probably not)
Assuming they do and you want to reduce individual current to say 5 mA then total load current is 140mA. (28 x 5)
Individual LED series resistance to produce 5mA would be around 700 ohms. (3.5 / 0.005)
So your pot' will have to simulate 28 times 600 ohms in parallel, in other words 22ohms (600 = 700 - 100) (22 = 600 / 28)
So, you need a 22 ohm pot, rated for 140mA ie 0.43 watts. Let us say 0.5 watts
However, the wiper of A 22 ohm 0.5 watt pot is only rated to carry 140mA, and you want it to carry the maximum current of 980mA so in fact the total pot rating must be 3.5 x 0.980 = 3.5 watts (as you originally suggested)
However, note that the pot value for 28 LEDs is only 22 ohms. If you use more or less LEDs you need a different pot to achieve the same level of dimming.
So, your best bet is to use a 10k pot to produce a 0-5 volts into an arduino (or similar) analogue port, then use a PWM output to drive a FET rated at say 5 amps to drive as many of your LEDS as you wish. Irrespective of whether you are driving 1 or all 75, this will give you fully balanced brightness control from a single pot.
Thanks I'll need to read it a few times to fully understand but does it make much of a different that the LEDs im using have a 3.2 forward voltage (white)?
If using a rheostat, yes it does. If using the arduino, no (within reason). However if you supply the exact model number of LED you intend using you will get a better answer.
I understand that by using the pot im controlling the overall current available to all of the LED's so when the number of lit LED's changes the current available will increase/decrease and affect the brightness of the remaining lit LED's
Will using the FET not be essentially the same as varying the voltage to the FET controls the current through it?
EDIT: I probably should have said that the project is one of those word clocks, so the number of lit LED's will vary depending on the time being displayed. I'm looking for a way to control the brightness which will stay constant no matter how many LED's are lit.
Specified forward voltage is 3.0 to 3.2 and rated current is 20mA
Using the lower voltage and assuming a 5volt supply, we need to drop 2 volts across the current control resistor (R)
By using the lower forward voltage we ensure the current control resistor is sized for maximum current control dropped voltage, hence LED is never over-driven.
R = 2/.02 = 100 ohms (so you got that part correct - well done)
My suggestion of using the PWM driven FET still stands since the FET is not used as a voltage dropping device; rather it is a switch which turns the LEDs on or off at the PWM frequency. If the LED is switched on for say 1mS and off for 1 mS continuously it will appear dimmer than it would if it was permanently energised. Let's say half as bright, though I'm sure I'll soon get told that eye response is non-linear - but it will appear dimmer. Maybe someone will advise on what the perceived dimness tabulation is for PWM ranging from 0% to 100%
Because the FET is switching circuits, and has a very low ON resistance (typically <0.05 ohms), a single FET can be used to control all your LEDs.
jackrae:
since the FET is not used as a voltage dropping device; rather it is a switch which turns the LEDs on or off at the PWM frequency. If the LED is switched on for say 1mS and off for 1 mS continuously it will appear dimmer than it would if it was permanently energised.
Thanks a lot, that cleared it up. I mis-understood the PWM and was thinking that the FET would simply be 'on' all the time but controlling the current with a voltage. Using PWM the FET will allow full current through but the rapid on/off gives the apperance of brightness.
It was this "This on-off pattern can simulate voltages in between full on (5 Volts) and off (0 Volts)" from the Arduino - PWM page that confused me. I was thinking that PWM would give between 0-5v but what its really doing is turning 5v on and off really fast.
Wazzled:
It was this "This on-off pattern can simulate voltages in between full on (5 Volts) and off (0 Volts)" from the Arduino - PWM page that confused me. I was thinking that PWM would give between 0-5v but what its really doing is turning 5v on and off really fast.
On average, though - that is what is happening; a voltage between 0-5v is produced - you need a small extra circuit to really "see" the voltage (a filter), as explained here:
Essentially, though - think of it this way: The PWM output is varying between 0 volts and 5 volts. The amount of time "on" (HIGH or 5 volts) versus "off" (LOW or 0 volts) determines the "average voltage". If it is always "off"/LOW - you have 0 volts; if it is always "on"/HIGH, you have 5 volts. If it is on and off equal amounts of time (regardless of the frequency mind you; although for most purposes, the frequency will be few/several ten's of kilohertz or better), you get an "average" between 0 and 5 volts, or 2.5 volts. If it is more "on" than "off", it skews towards 5 volts; more "off" than "on", it skews toward 0 volts.
I think that FET should do - at least the price is right. Turn-on resistance is a little high but no problem at only 1 amp load.
By the way, you don't put cr0sh's suggested filter into the feed to the FET driver circuit. This must only see the true PWM signal.
The filter is only used if you want to "convert" the PWM signal to an analogue value for display on a voltmeter.
You'd typically connect it into circuit by either tapping off a parallel feed from the arduino output pin or better still the FET output pin.
jack your probably getting fed up with me but the help is appreciated.
I have scavenged a CEP10N4 but not sure if its wired properly. I had the Source going to the positive side of my LED's, Drain to the positive of power supply but it wouldnt turn on either with high or low Gate. If I switch Drain and Source it stayed on and wont go off.
It does work tho if I put the negative of my transistor arrays (ULN2803A, controlling LED negative) to the drain and the source to the negative of the power supply. High on the gate turns them on, low turns them off.
I thought I could use it on the positive or would that need P channel FET?