Constant current driver

[SirJMD]

Anyone know a good constant current driver, that delivers around 300mA?

Got a few requirements:
- Dont want to buy a driver
- Prefer leaded instead of SMD
- Input voltage @ 21VDC or 15VAC
- 300mA constant
- Good effeciency (80+)


Ive made a schematic with LM3407, that should be 90+ efficient. Downside is, that the LM3407 is a SMD - a quite small one actually (recommended footprint for the pins is only ~1x0.4mm!).

So what im looking for, is like "Have a look at the IC named xxx", or something like that


Looking forward to see what you guys can come up with

 

[zzonbi]

34063

 

[SirJMD]

34063

Looks interesting, but from what i can see, you need to use its reference voltage to set the current. Seems to me, like it isnt that good for constant current source.

 

[zzonbi]

The current may be set by Rsc, with a lower drop (I guess they all use some voltage as reference).

AMC7150 is even simpler, with only 5 terminals.

 

[SirJMD]

The current may be set by Rsc, with a lower drop (I guess they all use some voltage as reference).

AMC7150 is even simpler, with only 5 terminals.

Altho its an SMD, the footprint is actually quite okay. Gonna give it a look aswell.

 

[SirJMD]

Updated the requirements a bit. Im gonna do some calculations - i might just go with a capacitor and an inductor. Should give an efficiency close to 99%.

 

[Ekke]

Should give an efficiency close to 99%.

With what and how? I don't think so..

 

[SirJMD]

With what and how? I don't think so..

Well, the capacitor doesnt dissipate power and neither does the inductor - that just leaves the LEDs.

After a few ms, the inductor will act as a short - which just leaves the capacitor and the LEDs. I plan to use 7x1W. If we believe 99% is true, that would mean 70 mW dissipated in the capacitor - sounds a bit high to me.

 

[Ekke]

Well, the capacitor doesnt dissipate power and neither does the inductor - that just leaves the LEDs.

After a few ms, the inductor will act as a short - which just leaves the capacitor and the LEDs. I plan to use 7x1W. If we believe 99% is true, that would mean 70 mW dissipated in the capacitor - sounds a bit high to me.

Constant current driver? Huh? :thinking:

 

[SirJMD]

Constant current driver? Huh? :thinking:

The capacitor will limit the current thru the circuit. However, if the applied voltage changed, so does the current. But if it does change, it will only be a tiny bit - it wont matter.

At a given frequency, the capacitor will have a reactance equal to 1/(2*pi*f*C), with f being the frequency and C being the capacitance in farad.

 

[Th232]

Could you post up a circuit diagram?

Right now the only similar circuit I can think of is where you put the cap in series between the power source and the LED. We use that trick in neurostimulators for controlling charge injection, but that's essentially an AC signal and has no inductor, so I can't see how it would work here.

 

[zzonbi]

If the inductor acts as a short, the capacitor acts as an open circuit. So you wouldn't even need those...

 

[SirJMD]

If the inductor acts as a short, the capacitor acts as an open circuit. So you wouldn't even need those...

Do you know what a damped oscillation is?


Th232 --> Sure, here it is:

Normally you would use a resistor instead of the inductor. The reason why its even needed, is because if the surge-current you risk having thru the LEDs, if you turn on the circuit at the top of the sine wave.
The idea behind it all, is that the capacitor doesnt dissipate power - eventho it limits the current, just like a normal resistor. We like that But we dont like the surge-resistor, since it will dissipate power.

 

[r0g3r]

Im using Lt1676 to drive 3 LEDS in series.
Voltage range is 9 to 50 volt DC with around 87% efficiency.
To limit the current im using a 1.5 Ohm resistor and a pnp transistor.
When the current is 350mA the voltage drop over the 1.5 Ohm resistor
starts the pnp to conduct and the collector (connected to the feedback pin)regulates the voltage.
lovecpf

 

[Th232]

Thanks for the diagram, now it makes sense, especially the inductor bit. Would you believe that after mentioning an AC signal in my first post I then thought "But nah... why would he use AC?", completely missing the part of your OP where you mentioned 15 VAC? Very interesting way of doing things.

Small question though (mainly since I haven't played around with this), what'll the current ripple be like? My initial impression is that you'll have the LEDs changing brightness by some amount whatever freq the AC signal is at,but by how much I'm not really sure.

 

[SirJMD]

Im using Lt1676 to drive 3 LEDS in series.
Voltage range is 9 to 50 volt DC with around 87% efficiency.
To limit the current im using a 1.5 Ohm resistor and a pnp transistor.
When the current is 350mA the voltage drop over the 1.5 Ohm resistor
starts the pnp to conduct and the collector (connected to the feedback pin)regulates the voltage.
lovecpf

Throw in a schematic always fun to see how others have made their own stuff

Thanks for the diagram, now it makes sense, especially the inductor bit. Would you believe that after mentioning an AC signal in my first post I then thought "But nah... why would he use AC?", completely missing the part of your OP where you mentioned 15 VAC? Very interesting way of doing things.

Small question though (mainly since I haven't played around with this), what'll the current ripple be like? My initial impression is that you'll have the LEDs changing brightness by some amount whatever freq the AC signal is at,but by how much I'm not really sure.

Yea i can imagine This forum's focus is more on DC, since most people in here only deal with flashlights.

Youre correct - the LEDs will flash with ~50 Hz. Ive done some testing, and not really noticeable. Its a small sacrifice to make, in order to get a higher efficiency - which is my primary focus, since im going for highest lumen/watt as possible.

Ill update the thread when i get rest of the calculations done

 

[zzonbi]

"Do you know what a damped oscillation is?"

An oscillation consumed through an element able to dissipate energy, such as a resistor or those leds?

So you count on the coil too to limit the inrush current. So in exchange for using more leds and flicker you save efficiency and a constant current driver.

 

[SirJMD]

"Do you know what a damped oscillation is?"

An oscillation consumed through an element able to dissipate energy, such as a resistor or those leds?

So you count on the coil too to limit the inrush current. So in exchange for using more leds and flicker you save efficiency and a constant current driver.

A damped oscillation could look like this:

What i want here, is the amplitude not getting too high - i dont want a big surge-current.


An inductor will try to prevent any sudden change in current, where a capacitor will try to prevent any sudden change in voltage. The idea here, is to use the inductor as a surge-resistor. As time goes (were talking ms here), the inductor will conduct more and more - at the same time, the capacitor will limit the current more and more, going towards the designed value.

 

[r0g3r]

Throw in a schematic always fun to see how others have made their own stuff

Ok here it is.
The red line shows the led current.

 

[space]

This is for SirJMD:
I think you should start to simulate this circut. I don't belive this can be called constant current. LED's will go to 0 current on each cycle. What frequency is the supply? The inductor could probably be dropped.

space