So i sat down to repair this desulphator and the design bowled me ...

My understanding of the operation is:

1. Mosfet turns on current ramps in both inductors
2. Mosfet turns off the inductors reverse polarity and their voltage increases to L*di/dt This causes the diode to forward bias and dump the voltage / energy back to the input

Control is through a 3845, voltage at the drain is rectified and made dc via 4934 / 47nf and 1uF then put to a TL431 operating an opto connected to the FB pin.

My question:
Couldnt the designer have removed the 220uH inductor and 470uFcap? Or is there something more to the design that i dont understand?

cant give out the entire schematic but can answer all questions ....

Thanks!

 

What's a 3845? Do you mean an SG3845?

What's a 4934?

You're missing the point of L2, the 1mH inductor and C10, the 470uF cap; L2 keeps the current flow into C10 relatively constantly. The 220uH inductor L1 is charged via the cap and the MOSFET drain. When the MOSFET turns off, the voltage across L1 swaps polarity, causing a spike through D1. C10 keeps that side of L1 at a relatively fixed voltage.

 

sorry UC3845 and 1N4934 .. both not on schematic

so L2/C10 sort of 'isolate' the input of 220uH from HV pulses of itself .... like a filter at the input of the 220uH .... cool!!

L2/C10 simply block all the HF pulses and if we know the freq of the power stage we can find out the component values by making fcutoff to be 10 times less than fsw and using

w = 1/sqrt(L*C)
w = 2*pi*fcutoff

arbitrarily setting C and then winding for L .... right?

Thanks!

 

Well, you're over-complicating things again. Basically, the portion of the circuit you posted is almost the same as Alistair Couper's original design. Alistair used a 555 timer and a P-ch MOSFET and no current sense resistor though:

Alistair used a P-ch MOSFET because it required fewer components to set the duty cycle on the 555.

By the way, if you wish, you can connect a battery trickle charger across your 470uF capacitor to keep the battery maintained, as long as you don't saturate the 1mH inductor.

 

Thanks! It is! and I thought it was designed in our Lab ....!

I read Alistair Couper's design and he hasnt expalined much of the design ...

My question is: Am I right on the purpose of the LC filter?

The frequency of the power stage is 1kHz, and the cutoff freq of LC is 232Hz about 4 times less ... to make it fit the 10x criterion of a good filter the components would pose a problem ... right?

Thanks alot!!!