Using a CLC for a LM4780

[Mr Teal]

I've tried searching, but haven't seen much information posted on using inductors in the power supply of a gainclone. Most info seems to focus on the negatives of parallel inductance. I've got a few larger, low resistance and probably fairly low inductance toroidal inductors and I'm thinking of adding them between my two main PS caps.

My system will be small, two channels using two bridged 4780s running on 39V rails or so from a 150VA 28.4V-0-28.4V toroidal xformer. PS caps will be two 10,000uF 63V Nichicons per rail, with bypasses of 100uF, 1 uF and 10nF. Ground common will be a 1/8" x 1 1/2" aluminium flat bar.

A little overkill I know, but I have the bar for another project, and at $6 for all 4 of the 10,000uF caps, I couldn't say no. Would the L make a difference, or will a difference not be noticable?

[hacknet]

IMHO, inductors are superb hum reducers but the gain clone has a very good PSRR so the cut down in hum might not be extremely essential. on top of that, the inductor introduces R into the PS, increacing voltage sag when transients come...

it may do magic for the vocals and highs thou....

[Franz G]

increacing voltage sag when transients come...

... exactly, and you get a "soft" psu.

High inductance chokes (>5 Henry) are interesting for class A tube amps psu's.

But: this amps dont "sucks" peaks of several amps, like a poweropamp.

An choke with low DC resistance, high inductance and high powerhandling would be very very big!

Franz

[SheldonD]

You may also need to conksider that:

if you use choke input power supply the voltage DC out will be less, about 90 % of the voltage of capacitor input power supply.

This is of course dependent on sufficient size of inductance to store the energy required for peak demand.

If this condition is not met it then the output will be somewhere higher as capacitor will now supply some of the stored energy.

[owen]

Ahhh chokes.

Those torroids are interesting beasts. I would use the secondary winding (lower DCR), but the primary will have a huge inductance.

Now chokes do add DCR to a PSR, but, they block RF very effectively. They also can store a large amount of energy in the core that is 'called on' during PSU sag (the collapsing magnetic field induces the voltage). Just make sure the NC winding is exactly that - not connected, otherwise you have a real shock hazzard.

The best place for these is on the buffer, but there is nothing in principle that would preclude thier use in the clone psu itself.

As you will be CL, then the actual line voltage will be >0.9 of the RMS input, but < 1.4 . Because of the way chokes 'work', they significantly reduce the instantaneous load on the diodes (good).

Personally, I'd clip lead one in, and see if I like it. I like CLC supplies elsewhere, and the GC should be no exception (in many ways I prefer them to regulation)..

So have fun!

Owen

[Mr Teal]

A little bit of rail drop isn't a problem. I have 4 ohm speakers, but they really don't need anywhere near the power that the amp can put out at 35V into 4 ohms. I doubt I've ever listened to them at more than 10W. I'm more concerned with the quality and stability of the PSU. If the rails drop to +/- 28VDC, or even 25V, I wouldn't cry.

I'd hope it would help out a bit with RFI. I live in an apartment, and my audio gear unfortunetly has to by plugged into the same breaker as my TV, PS2, and computer. Very electrically noisy. :P

[hacknet]

you might want to try doing something like CLC-regulator.

and a CLC will still give you 1.414 times AC as long as your C can support the current draw.

[Mr Teal]

I just did a little modeling, I'm not sure how accurate it is. With the 500uH inductor in between the two main filter caps, the ripple voltage stays about the same, but there's a dramatic difference in it's shape. With the inductor in place it's much more sinusoidal, while the straight C is a fairly hard ramp wave.

Here's the paralleled C supply, two 10uF (ESR modeled as 10mohms each). Secondary resistance 0.5 ohms, 28.4V, load of 2.5A

[Mr Teal]

And this is the CLC. Same transformer and load setup, but with a 0.2 ohm 500uH inductor in series.

[Mr Teal]

Suppose I forgot the file.