Understanding this power supply

[summilux]

Everyone,

I am building Thorsten's Toccatta preamp and I am trying to understand the power supply (attached below). The HT supply has 39ohm resistors between the rectifiers and the large caps on both the + and OV rails. From my basic understanding of electrics, the resistor on the + rail is there to limit the startup current into the caps (correct me if i am wrong please). What is the purpose of the resistor on the OV rail?

I am trying also to understand the MOSFET part of the supply, so if there are any informative sites you can send me to about this type of voltage regulation, I'd appreciate that.

Thanks

Jonathan

[kathodyne]

slow rise of hv?

[sagarverma]

the capacitor between earth and neutral is banned in many countries.plz check properly b4 using it lest regulatory officials come knocking on your door.

[amperex]

It is an RC network that limits high peak currents into the 47uF cap for a quieter power supply. I would use the design.

[poobah]

The 39 Ohm resistors could have been a single 78 Ohm resistor in either leg of the power supply. By splitting the value into 2 39 Ohm resistors it is possible to spread the heat dissipation over 2 devices. Also there is some capacitance between the primary and secondary of the transformer. By splitting the resistance across both legs of the power supply, better isolation is achieved. The power line is always a nasty source of noise, so this a cheap way of getting a little improvement in isolation.

The MOSFET regulator is not really a regulator in the strict sense. It is basically a source follower driven by a voltage divider. To understand how it works:

1. Forget about the zener doide, it is there to protect the FET.

2. Replace the 10K resistor with a short, it is there to protect the zener.

3. Forget about all the caps... for the moment.

Assign some names to the components and voltages:

* Call the sum of the resistances from the 1M pot and the 1M resistor - R1

* Call the 10M resistor - R2

* Call the 330K - RG

Now, the source follower works just like a cathode follower... more or less. The difference being is that the gate voltage must be higher than source voltage for the FET to conduct. This voltage is about 2 - 4 Volts.

* Let's call the voltage across RG - VG (gate minus source)

* Let's call the voltage in (at the drain of the FET) - VI

* Let's call the voltage out (at the source of hte FET) - VO

The voltage, VG, stays more or less constant, so the output voltage basically equals the voltage of the divider formed by R1 and R2 minus VG. The R1/R2 divider has hefty current load on it, so we have to take that into consideration to determine the output voltage.

The equation for the output voltage is:

VO = [((VI/R1) + (VG/RG)) * ((R1*R2)/(R1+R2))] - VG

Plug this into a spread sheet... use a value of 3 Volts for VG. Try different values for VI. You will see that VO is always a constant fraction of VI (minus the VG). Of course if you change R1, the fraction of VO versus VI will change with it. Smaller R1... more VO.

So it is really not a regulator, but a buffered voltage divider.

Now add the caps; the 47uF stabilizes the gate voltage and therefore the source voltage. This circuit will work to reject any ripple in the audio band present at the drain of the FET.

Later,

[poobah]

Boy did I miss the point! Not the HOW but the WHY...

Amperex & Kathodyne bring up excellent points The 47uF cap on the divider makes the VO come up nice and slow... about one minute.

Cool!

[amperex]

You're no idiot, for sure.

[dhaen]

Using a resistor in each leg is sometimes done to reduce common-mode noise. I can't second guess if Thorsten did it for this reason.

[Frank Berry]

In many power supplies, a vacuum tube rectifier will provide the slow start of the power supply. In this full wave bridge design, the MOSFET provides much the same thing. It's a good design.

I wouldn't replace the 10k resistor with a short. It IS designed to protect the zener diode. Removing the resistor won't improve the performance of the circuit to any great extent.

[poobah]

Hey guys,

I wasn't suggesting to actually omit the zener or the 10K resistor in the real circuit... by all means they belong there.

I was just omitting them for the sake of simplifying the math and the understanding of the circuit thereof.

Cheers,