Help with this LED setup (recommend digi rheostat)

Thread Starter

lwj5

Joined May 17, 2010
13
This is my circult diagram for my high intensity LED torch. (attached)

My idea is to use PWM for the control of the brightness to achieve high efficiency.

However, I encounted a problem...

This digital rheostat only has a 64 steps, hence leading to a increase from 66% to 96.9% duty cycle for the last step. This is not favourable to me, is there anyway to solve this, or can some1 recommend me a digital rheostat with >100 steps.

anws value for B:
if using 10Ω-10kΩ - B is 316.2Ω for max duty cycle variation
if using 100Ω-100kΩ - B is 3162Ω for max duty cycle variation
if using 50Ω-50kΩ - B is 1581Ω for max duty cycle variation

max variation is from 3.07% to 96.9% , if possible pls recommend me a better digital rheostat with more steps. Feel free to comment on my setup

Datasheet for digi rheostat
http://datasheets.maxim-ic.com/en/ds/DS1869.pdf
 

Attachments

Last edited:

SgtWookie

Joined Jul 17, 2007
22,230
A standard 555 timer cannot operate from less than about 4.5v.

A CMOS 555 timer has very poor source characteristics at low voltage. You would be better off to use an N-ch logic level MOSFET to sink current from the LED/driver.

You don't show the diagram of the driver for the LED, nor any characteristics of the LED.

High-efficiency LED drivers use an inductor to keep the current through the LED relatively constant.

Before we can go any further, you really need to give more details about the LED, and how you are intending to power it. 3.7v seems like you might be considering batteries.
 

Wendy

Joined Mar 24, 2008
23,415
You could try a variation of this...



Drop R9 to 100Ω, remove the R10 leg, and eliminate the R11 resistor and it should work just fine off 3.7V. The 7555 is a CMOS type 555 chip.

It is loosely based on an idea Wookie referred me to quite a while ago, and I ended up writing an article about it.

CMOS 555 Long Duration LED Flyback Flasher

If you are interested I'll redraw the schematic into something closer to your application.
 

Thread Starter

lwj5

Joined May 17, 2010
13
A standard 555 timer cannot operate from less than about 4.5v.

A CMOS 555 timer has very poor source characteristics at low voltage. You would be better off to use an N-ch logic level MOSFET to sink current from the LED/driver.

You don't show the diagram of the driver for the LED, nor any characteristics of the LED.

High-efficiency LED drivers use an inductor to keep the current through the LED relatively constant.

Before we can go any further, you really need to give more details about the LED, and how you are intending to power it. 3.7v seems like you might be considering batteries.
Thanks for ur reply, i will look up more on that mofset u said
Sorry i left out some details i thought was not needed.

From 555 wiki page i found that some 555 timer are able to run all the way down to 2-3V

Here's the LED website, its max 12w. However this is a different bin. The led I am getting is 3.7-4.5v

www.dealextreme.com/details.dx/sku.12721

Here's the driver the specs can be found at the bottom of the page

www.dealextreme.com/details.dx/sku.1885

It will be powered by a 3.7v 18650 battery
 

Thread Starter

lwj5

Joined May 17, 2010
13
You could try a variation of this...



Drop R9 to 100Ω, remove the R10 leg, and eliminate the R11 resistor and it should work just fine off 3.7V. The 7555 is a CMOS type 555 chip.

It is loosely based on an idea Wookie referred me to quite a while ago, and I ended up writing an article about it.

CMOS 555 Long Duration LED Flyback Flasher

If you are interested I'll redraw the schematic into something closer to your application.
thanks for ur reply

I am currently outside right now. i will take a look at it once i reach home.

thanks
 

Wendy

Joined Mar 24, 2008
23,415
There are 2 types of 555 times, the older standard has a minimum voltage of 4.5VDC to make it work, and for technical reasons it just doesn't work as well in that voltage range. The newer type, the CMOS 555 can go down to 2V, but has exceptionally weak drive.

I cover a lot of this in my article LEDs, 555s, Flashers, and Light Chasers.
 

SgtWookie

Joined Jul 17, 2007
22,230
Your battery will not provide enough "headroom" for the LED plus the regulator plus a switch to provide PWM.

You need closer to 4v for a supply.
 

Wendy

Joined Mar 24, 2008
23,415
I was talking about 30ma parts, which probably doesn't apply to this project.

The schematic I showed at a minimum doubles the LED supply voltage, probably even more. The experiments I ran at 7.2 V suggest a resistor is required to limit the current on the 3 LEDs whose total Vf added up to 9.6V, with plenty of head room for more. I may run some more experiments at 3.6V since that is a popular battery voltage. It can put out 60ma if coxed. Wookie, if the coil were increased would this increase the current?

The oscillator driving the flyback coil runs at 140Khz. You can PWM this circuit using the reset pin on the 555. The biggest disadvantage is the size of several 555s (or one 556), but since these parts can be had in surface mount format they can be shrunk to really tiny sizes.

One of the things that bothers me is my lack of ability to predict this circuit reliably. I've gotten around this lack by hands on experimentation.

We had another project for another gentleman that was similar in some ways, in that it used a 3.6V battery and a high power LED. I didn't see a current rating for this LED, I assume it is either 350ma or 700ma. The other project wound up using 3X batteries. The advantage was the batteries would provide maximum power until they were dead (and it would die all at once).

We need a current rating for this LED.

If you draw 700ma for the LED the battery life will be measured in minutes, especially since there will have to be a conversion which decreases the efficiency even more.

http://forum.allaboutcircuits.com/showthread.php?t=32364
 
Last edited:

SgtWookie

Joined Jul 17, 2007
22,230
Bill,
The regulator our OP is considering is linear (actually, three 350mA regulators on a wafer), and needs a very small amount of headroom. If it has too much headroom, it will dissipate lots of power as heat.

The Vf of the LED is somewhere around 3.7v, but that would be as much as a freshly-charged battery. The LED would start going dim as soon as they turned it on.

The way our OP wired his driver transistor is as a voltage follower; so they'd get at least a 0.7v drop across it and plenty of heat - if they had that much headroom to spare.

The use of a power MOSFET is indicated, but it would have to have a very low gate threshold voltage and a low gate charge, as well as a low Rds(on).

Using straight PWM would result in power being dissipated in the battery. An inductor in series with the LED and a recirculation diode like an 1N5817-1N5819 would help keep the current flow more consistent. A capacitor across the battery would help to decrease power loss in the battery during PWM.

As far as the digital rheostat, it would be better to use a digital pot. The wiper of the digital pot goes to pin 7 of the 555; the upper "end" connects to the bottom of R1, and the lower end connects to C1. The diode goes from pin 7 to C1.
 

Thread Starter

lwj5

Joined May 17, 2010
13
It is loosely based on an idea Wookie referred me to quite a while ago, and I ended up writing an article about it.

CMOS 555 Long Duration LED Flyback Flasher

If you are interested I'll redraw the schematic into something closer to your application.

Thanks Bill for your recommendation of the 7555, but but I don't think a inductor will be necessary in my setup as the AMC7135 drivers are designed to maintain a constant relatively constant current and to prevent overloading.

BTW, that is a very helpful guide, it helped me understand the working of that transistor in the circuit.


Your battery will not provide enough "headroom" for the LED plus the regulator plus a switch to provide PWM.

You need closer to 4v for a supply.
I think I can neglect the source's voltage for now. I checked the website, I will be getting the 3.6-3.7V P7 LED instead. There are people who commented that they were able to power up the LED with a 3.0V so I am relatively certain that this should work.

For that 1st post, when you mean, place a N-channel mosfet, do you mean like this(attached)?
and another few silly questions... sorry I am not very good with electronics
1) I am not sure whether the output of the 7555 would be of high enough voltage for the logic to work, can anyone enlighten me.
2) I am not sure whether I wired the wires leading up to the mosfet correctly up in the diagram correctly.
3) If it is correct, do I need to place a resistor R2? If so how much of a value to place
 

Attachments

Thread Starter

lwj5

Joined May 17, 2010
13
Bill,
The regulator our OP is considering is linear (actually, three 350mA regulators on a wafer), and needs a very small amount of headroom. If it has too much headroom, it will dissipate lots of power as heat.

The Vf of the LED is somewhere around 3.7v, but that would be as much as a freshly-charged battery. The LED would start going dim as soon as they turned it on.

The way our OP wired his driver transistor is as a voltage follower; so they'd get at least a 0.7v drop across it and plenty of heat - if they had that much headroom to spare.

The use of a power MOSFET is indicated, but it would have to have a very low gate threshold voltage and a low gate charge, as well as a low Rds(on).

Using straight PWM would result in power being dissipated in the battery. An inductor in series with the LED and a recirculation diode like an 1N5817-1N5819 would help keep the current flow more consistent. A capacitor across the battery would help to decrease power loss in the battery during PWM.

As far as the digital rheostat, it would be better to use a digital pot. The wiper of the digital pot goes to pin 7 of the 555; the upper "end" connects to the bottom of R1, and the lower end connects to C1. The diode goes from pin 7 to C1.
sorry I was typing that post and didnt see your post earlier, I will go and change the circuit right now
 

Thread Starter

lwj5

Joined May 17, 2010
13
I was talking about 30ma parts, which probably doesn't apply to this project.

The schematic I showed at a minimum doubles the LED supply voltage, probably even more. The experiments I ran at 7.2 V suggest a resistor is required to limit the current on the 3 LEDs whose total Vf added up to 9.6V, with plenty of head room for more. I may run some more experiments at 3.6V since that is a popular battery voltage. It can put out 60ma if coxed. Wookie, if the coil were increased would this increase the current?

The oscillator driving the flyback coil runs at 140Khz. You can PWM this circuit using the reset pin on the 555. The biggest disadvantage is the size of several 555s (or one 556), but since these parts can be had in surface mount format they can be shrunk to really tiny sizes.

One of the things that bothers me is my lack of ability to predict this circuit reliably. I've gotten around this lack by hands on experimentation.

We had another project for another gentleman that was similar in some ways, in that it used a 3.6V battery and a high power LED. I didn't see a current rating for this LED, I assume it is either 350ma or 700ma. The other project wound up using 3X batteries. The advantage was the batteries would provide maximum power until they were dead (and it would die all at once).

We need a current rating for this LED.

If you draw 700ma for the LED the battery life will be measured in minutes, especially since there will have to be a conversion which decreases the efficiency even more.

http://forum.allaboutcircuits.com/showthread.php?t=32364
this P7 LED has a wattage rating of 12W, so that should convert to around lets say 3A rating?

the 18650 has 2400mAh, normally I wouldn't need to max out this LED as its pretty bright, so around duty cycle of 15% should do quite fine, the eff of the driver is also quite high at

3.7V >95%(est from graph) http://www.candlepowerforums.com/vb/showthread.php?p=2406347

so that battery should last quite long.
 

SgtWookie

Joined Jul 17, 2007
22,230
The 2400mAh rating is at a constant discharge over a 20 hour period, or a load of 120mA.

Your LED/regulator will be nearly 9 times that much of a load, so you will have quite a bit of power dissipation in the battery. Straight PWM will still place the peak load directly across the battery. A cap and inductor will remove the peak load effect, but you will have to have a high PWM frequency.
 

Thread Starter

lwj5

Joined May 17, 2010
13
I found a Digi Pot which suit my need, a 128 steps pot

I never thought of using a Pot, thanks for the advice.

I uploaded the version, with the advice you gave. Pardon me if I made any mistake.

I havent add the caps and inductors, I got to stop here for today, it's getting late.
 

Attachments

SgtWookie

Joined Jul 17, 2007
22,230
You have the MOSFET on the high side of the LED & regulator. That's not going to work, I'm afraid.

The MOSFET needs to be below the LED & regulator. Otherwise, as you raise the gate voltage, the source will rise, and you will heat the MOSFET because it will be in the linear region.

If you want to switch on the high side, you will need a P-channel MOSFET. However, N-channel are more efficient (electrons flow more easily than holes).
 

Wendy

Joined Mar 24, 2008
23,415
I have an revision to my LEDs, 555s, Flashers, and Light Chasers (this link is to the revision under progress). Check out chapter 10, Transistor Drivers, it goes through MOSFETs on the final section. You need what is called a logic level MOSFET, which is basically a standard MOSFET designed for lower voltages.

It is probably obvious I'm following this thread with interest. LEDs and 555s are both specialties for me.
 
Last edited:
If you have any chops with microcontrollers, then this is a perfect application for one... avoid all the analog and read the two pushbuttons and change the duty cycles in the code.
 

SgtWookie

Joined Jul 17, 2007
22,230
Yes, this would be a very good application for a uC.

Battery, cap, uC, inductor, LED, recirc. diode, low-threshold N-ch MOSFET, a couple switches and a few lines of code.
 

Thread Starter

lwj5

Joined May 17, 2010
13
it goes through MOSFETs on the final section. You need what is called a logic level MOSFET, which is basically a standard MOSFET designed for lower voltages.

It is probably obvious I'm following this thread with interest. LEDs and 555s are both specialties for me.
Nice read there, glad you are following.

If you have any chops with microcontrollers, then this is a perfect application for one... avoid all the analog and read the two pushbuttons and change the duty cycles in the code.
Yes, this would be a very good application for a uC.

Battery, cap, uC, inductor, LED, recirc. diode, low-threshold N-ch MOSFET, a couple switches and a few lines of code.
Haha, yes I have thought of using a uc to control the pot, on/off switch and also thought of getting a 3 small 7-LED digit display to display the duty cycle.

However, I have almost no knowledge of this, the one time I used it was on a R/C toy car, that was controlled with a flowchart kind of programming(PICAXE). The other programming I am familiar with are Java and VB. Another constrain I have is that I am fitting this into a small package, I really have no idea how much space this could take. Could any1 start me off in the right direction, a recommendation of a uc?


I have edited the circuit to include the diodes, the inductor and the cap C3, however I have no idea the values for C3, R3 and Inductor.

R1-100Ω
R2-50Ω
C1 - 0.06uf
C2 - 0.01uf
C3 - 270uf (perharps?)
R3 - 1kΩ?

this would result in a ~1000Hz frequency, is this too much or little?
 

Attachments

Last edited:

Wendy

Joined Mar 24, 2008
23,415
The freq sounds about right, basically fast enough you can't see it. How you're using your digital pot concerns me a little, I'm not too familiar with them. The circuit I think you are trying to emulate looks a little like this (minus the motor and extras).



Side note, if you can get your hands on them Shottky diodes will work much better at such low voltages. A normal diode drops around 0.6 V, where a Shottky drops around 0.1-0.2V.

The inductor I can't help you with, but the caps are calculated F=0.7/R3C1
 
Top