An app note on a LED-driver circuit:
This LED-driver circuit includes a hysteretic controller U1 (MAX16820), related power components, and a control circuit based on the quad op amp U2 (LMX324). U1 drives five HB LEDs from a 24V source, using only inductor L1, MOSFET Q1, and catch diode D1.
Thanks rsdio!
I just find the PWM generation circuit around the LMX324 superfluous. The 16820 provides 5V Vcc, perfectly suitable for a microcontroller.
I’d use a small AVR instead, for more flexibility and much smaller component count.
I have not studied the MAX16820ATT at all, but other Maxim LED driver chips have a DIM input which accepts either a DC voltage to control brightness, or a PWM input to control brightness. In other words, I have a hunch that the LMX324 is only there to create the PWM control signal for the 16820, and I also assume that the 16820 doesn’t have the handy DC voltage control feature. I bet Maxim has moved almost that entire circuit into some of their other LED driver chips so that you don’t need a PWM generator.
I am confused when you say that the 16820 ‘provides’ 5 V Vcc, suitable for a µC. For one small matter, it’s actually 5.3 V, which might actually be too much for some chips. It seems more important to focus on what is provided to the 16820. But I do agree that a digital processor can generate PWM as well or better than an analog PWM generator. Perhaps Maxim is just trying to show another way to do things without the standard µC that everyone is using. Note that having a µC requires that you write a program and load it into every board via ICSP or equivalent. An analog PWM generator is ready to go as soon as the chips are soldered on the board – no code writing, no programming of Flash, just solder and go.
On a related note, I’ve looked at several manufacturers for a PWM generation chip that simply take a DC control voltage as input. I don’t want to wire up a quad op-amp with 11 resistors, 2 capacitors, and a diode. I’d rather just provide power and a DC voltage from a single variable resistor, and have an output pin with PWM. That’s a 4-pin chip, maybe 6 with an external capacitor. Since I haven’t found such a pre-fab solution, I’m less surprised that Maxim is showing an example of how to build one the ‘hard’ way.
By the way, I send these App Notes for Ian to sift because I think it would be educational to tear apart these designs to see how they work (or don’t work). Maybe the forum would be better, so drop a link here if you want to discuss this circuit some more (and maybe I’ll read the data sheets this time).
The datasheet speaks about a 5V, 10mA on/board regulator, where do your 5.3V come from ?
But your point about a microprocessor having to be programmed is valid. I suspect you could abuse a NE555 for the purpose.
@Markus, look at the graphic at the top of this page. Vcc ~ 5.3V
I see. Then either the graphic or the Datasheet is wrong. I was looking at the data sheet.