Help designing a cheap LED dimmer

Thread Starter

-live wire-

Joined Dec 22, 2017
901
I am going to be getting some LED strips for illuminating my house. They will be replacing florescent bulbs. There will be about 10-15 4 ft sections, in different locations. They are 12V, and each section will consume about 4-5W. They will likely produce over 500 lumens. Each 4 ft section will be completely separate from the other ones. They will, unfortunately, each need their own power supply. There is no way around that. That means an additional $3 per light, doubling the cost. With the number of lights, it is already pretty expensive. And there are additional costs, like acrylic or some other material to diffuse the light and make it look better. I want to minimize costs here.

I would also like the ability to adjust the brightness of the strips with PWM. They may be too bright initially. I have looked at PWM boards and the cheapest ones are $2.50. For 10-15 lights, that's and additional $30-40. I would like to design one that is under $1. I could use some help with that.

So the main component here is the switch. I already have 10s of these IRFZ44N mosfets from international rectifier (datasheet). They look like the best option here. I will use a simple pulldown resistor here to avoid certain issues. I need to generate some sort of PWM to drive them. It needs to be between 0 and 100% (or very close). It should be at a reasonable frequency, .5-50kHz. I do not care if it is sort of audible. I do not really know what to do here. What is the most economical solution? Is there some IC that requires very little other passives? Is 555 the best option? I am not too familiar with 555 or other oscillators, and have never built any kind of oscillator circuit.
 

crutschow

Joined Mar 14, 2008
23,758
Here's the description of a simple PWM circuit that requires only 5 components, not counting the MOSFET.
It will go to within a percent of so of full off (red trace below) to full on (blue trace below).
The yellow trace is 50%.

upload_2018-6-20_9-35-35.png
 

crutschow

Joined Mar 14, 2008
23,758
Why do you have U3-U6? It seems excessive.
Not when you know what it does. ;)
There are six inverters in one CD40106 package, so I'm using the five spares in parallel as a low output-impedance driver to rapidly charge/discharge the high gate capacitance of the MOSFET.
That allows it to switch rapidly with a minimum of switching dissipation.
The slower a MOSFET switches, the more power it will dissipate during the switching time.
And how does that circuit work?
It's a Schmitt trigger relaxation oscillator.
 
Last edited:

Thread Starter

-live wire-

Joined Dec 22, 2017
901
Oh, I did not recognize the Schmidt trigger. I thought it was just an inverter or something else. So this diagram does not show the power connections for the IC, correct?
There are six inverters in one CD40106 package, so I'm using the five spares in parallel as a low output-impedance driver to rapidly charge/discharge the high gate capacitance of the MOSFET.
That allows it switches rapidly with a minimum of switching dissipation.
The slower a MOSFET switches, the more power it will dissipate during the switching time.
So it would be properly connected to ground when it is off, and would not need anything else to prevent false triggers?
 

crutschow

Joined Mar 14, 2008
23,758
So this diagram does not show the power connections for the IC, correct?
True. The power and ground connections are not shown.
So it would be properly connected to ground when it is off, and would not need anything else to prevent false triggers?
CMOS gates have complementary push-pull outputs (below) and will actively drive the MOSFET gate between Vdd (Vcc) and ground, never floating.
upload_2018-6-20_11-33-25.png
 

Thread Starter

-live wire-

Joined Dec 22, 2017
901
True. The power and ground connections are not shown.
CMOS gates have complementary push-pull outputs (below) and will actively drive the MOSFET gate between Vdd (Vcc) and ground, never floating.
View attachment 154792
Ok. Also, I am working on this DIY arc lighter project. This is the thread for it. I will probably go with a simple H-bridge and oscillator to drive it. I will have this issue where if there is a set impedance, there will be far more power consumed at higher voltages than at lower voltages (lithium cells ranging from 3V-4.2V). So I have to compensate for that somehow, to acheive fairly constant power. A DC-DC converter does not seem practical. One way to do this would be to drive it with a VCO. If there were lower frequencies at lower voltages, there would be less inductive reactance, more current, and close to the same amount of power. I would need to be able to adjust the VCO, and get frequencies between 100kHz and a few MHz. I really have no idea what the inductance of the coil is, and it may need to be changed to acheive a larger turn ratio.

But I do not really know how to go about making such a VCO. It seems like that kind of Schmidt trigger circuit would be partially dependent on voltage, with how a capacitor charges faster to a set voltage with a larger voltage applied. But would that Schmidt trigger circuit allow up to a few MHz, or is a 555 circuit or other better for this application? And could it consume less than 200mA, even at higher frequencies. What do you recommend here?
 

Hymie

Joined Mar 30, 2018
721

Thread Starter

-live wire-

Joined Dec 22, 2017
901
Well DIY is only $1, as opposed to $1.50, and it would be a good learning experience. It may only be $5-$10 saved, but there are things I would much rather spend that money on.
 

olphart

Joined Sep 22, 2012
78
Amazon has Several vendors of PWM 12V units, IR & RF as remote.
Just be Very Careful on current - stated ratings are Usually WAY too high.
Also, IR is interesting in a multiple unit situ.
RF is great, but most you'll see is all same channel.

I buy units rated at 6 amps, but the MOSFET is Peak at 6, 2 is constant and That's with a heat sink.
As bought, you can't push past 1A.
Gotta love the Ch...., steal the IP then lie about the specs on selling it back to us...

Reason I buy is the parts cost 2+ times the unit cost, I can't compete (as is).
Also, if you know MOSFET tech, they're a Great base to graft Whatever you want as output capacity.
 

ebeowulf17

Joined Aug 12, 2014
2,958
The DIY approach comes to less than a dollar per item, and it would also be a good learning experience.
I'm all for diy and learning, but I'm skeptical on costs here... if you don't know how to build your own diy version, how do you know you can do it for a dollar? Hard to price something you haven't got plans for yet...
 

Thread Starter

-live wire-

Joined Dec 22, 2017
901
I will go with the schematic crutshow posted. I went on mouser, found suitable parts, and they totaled at 96 cents per board. That's how I know.
 
Top