LED Dimming Control

NT

The PSU is fine. LEDs use a constant current PSU. Using a 50 watt PSU with a 100 watt LED just means half the light. In fact, that's what I want to do, control the current to control the brightness.

So it does what you want.

So it doesn't do what you want.

If you add a separate current limit, the psu probably will max its V_out & stay there, and control will be in the hands of the secondary current limit. whether that's a problem depends on what you want, but it doesn't need to be.

NT

I'd pop the commerical power supply open and check out the circuit. My guess is there's a resistor in there that fixes the current setting. By replacing the fixed resistor with a pot, you can have variable current. Pretty much all of the commercial lighting power supplies have a variable version and a fixed version, and the circuit is the same except for wires to an externally-mounted pot.

Jon

This particular unit is an open frame low end unit... *lowest* end unit. Maybe I'll try to trace out the schematic. I want to control it automatically, so I suppose a FET may replace the resistor. Also, there will be 100 watt units while this one is 50 watts. So I'll have to do it twice, at least.

I found the controller chip and the current sense seems to be rather low value current sense resistors in the source leg of the FET. They use four 0.68 ohm resistors in parallel. I think it would be hard to add anything to that to make it programmable. I guess I'll have to start over and either find a dimmable unit or make one from scratch.

I'm not familiar with this sort of circuit. The typ application in the data sheet for the BP3319 shows the secondary totally isolated from the primary. This board has one component feeding back from the negative of the secondary after the diodes to the positive of the primary after the diode bridge and an inductor. It looks vaguely like a capacitor, but it is hard to tell. The marking is HCY222M and a couple of voltages. Any idea what this component is? Why would it be there?

There's no such thing as a '100watt LED', there has to be a bunch in series for 32V, and in parallel if they take 1.5A.

So, just short across some of the series elements (it'll still be 1.5A, but at 15V that's a dimmer string). That's how a 'constant current' gives you dimmer lights.

Hmmm... how exactly do I cut the case open and tap into all the individual LEDs?

sorta related,

no, you're over-thinking it.

adjust the duty sysle of the boost switch.

ignore the voltages, read the current off the label.

mosfet | FB ----+ | 0R16 _|_ ////

so do this

in ----. | 10K mosfet | | FB--+--[1k]---+ | 0R16 _|_ ////

maybe needs a capacitor parallel with the 1K, value depends on the operating frequency

Why would a fet with Rdson = 0.17 ohms not work?

NT

Hmm. Tiny little connections behind a silicone barrier. Not gonna work, I guess.

So, it has to be a modification to the power supply. There must be a reference voltage somewhere, that it compares to the series resistor drop, and you want to attenuate that reference voltage (so the difference amplifier stabilizes the output current at a lower value). Rather than a variable power resistor, just vary the reference voltage... if it isn't internally generated and inaccessible.

Shunting current around the LEDs with a resistor would use the same power (just waste the excess as heat), but would dim the light.

Or, get a different power supply.

Y capacitor ?

switch in 1-4 of the resistors with low Rds fets? change resistors for more options

or choose resistor for minium current and add a variable divider from the resistor to the CS pin

Similar problem to the LED. Reference voltage is inside the chip. The series resistors are four parallel 0.68 ohm resistors, so not a range that is easy to emulate with a FET. However, to lower the current this would would need to be increased. So maybe that is possible. This is not my field of expertise though.

I thought of that, but it's a *lot* of heat, 100 watts.

Hard to find actually. At least in a form factor I can use. I'm trying to mod an existing torcherie lamp. Remember those 500 watt death traps? If they fall or a curtain blows over top they set things on fire.

At 100 watts the are much less likely to do that. To get rid of the 100 watts I am using an old frying pan as the light directing bowl and heat sink. I'm including a quiet fan to blow a bit of air over it all. The LED needs airflow too, but I haven't figured out a way to do that without obstructing the light, I'll worry about that a bit later. Some air deflectors might work here.

Ok, I'll bite, what's a Y capacitor?

For the boost circuit to work the input voltage has to be lower than the output. That is, something has to make the output of the constant current supply drop. That would be drawing current into the boost regulator. If the boost regulator draws more current it will be driving the output voltage/current up. This is done by adjusting the duty cycle, but it doesn't change the basic instability. Fortunately the max voltage of the existing supply isn't all that high, around 50 volts I believe. So maybe a buck with a current control would be the appropriate second stage.