Confusing Problem With Regulator

[10100011]

Hi Folks..

I'm in need of some suggestions as to what's going wrong with my circuit. I've so far constructed a 12v and 5v linear power supply, and wish to add a 3.3v adjustable supply using one of these:

Adjustable Voltage Regulator

I'm trying to turn the 20v DC from the rectifier into 3.3v DC - and I'm using the positive buck converter topology. In the datasheet (the datasheet) it gives an example circuit to achieve 5v DC - what I've done is use the exact values except for:

1) The inductor I have is two smaller ones in parallel, giving 40μH instead of the 100μH they have

2) I've changed the resistor value using a 1k resistor in series with a potentiometer, with the ground pin (pin 3) not connected. When I was assembling the circuit, I adjusted the pot to give an overall resistance of 1.24k

When I tested the circuit earlier on, the voltage on the dummy load was 16-17v and climbing, and soon after magic smoke started coming from the resistor (a several watt resistor, 50 ohms). When I switched the power off, I went to measure the resistances once again but no matter how I measured the resistance, it only showed 30 ohms - either accross the 1k resistor, the pot, from the 1k resistor to the regulators 'leg' - but all it would show was 30 ohms!

Basically, I have no idea why it's not working as I've done everything I realistically could to replicate the circuit they've drawn.

[Sch3mat1c]

Guessing the excessive voltage killed it... not sure how though. Excess current could do it too, but I'd think it would be current limited nicely.

How's your layout? If it's loose wires on a bench, you get what you deserve, but if it's nice and tight, on a ground plane on a PCB, it may be a mystery.

Better is to use a buck chip -- that's a boost chip and it's doing all sorts of crazy floating stuff to even run. Regulation will be terrible because it's only getting feedback in pulses through a 1N914.

Tim

[tekwiz]

Parallel inductors? At less than half of the correct value?

[colin55]

To get 40uH, you must have two 80uH inductors in parallel. You could use this inductor by itself.

The simplest is to add 3 diodes to the ouput of the 5v to get 3v3.

[Ice-Tea]

Doubt the inductor would be the issue... Unless that controller does not have pulse skip/discontinuous mode? Would expect additional ripple, maybe, especially at lower output current (which is kinda the case with that 50 ohm resistor?

My guess would be that there's an issue in your feedback loop. Badly connected, bad contact in the pot? If the feedback loop ' sees' 0V it will increasse duty cycle all the way up...