Chokes, sizing and ringing?

[woodturner-fran]

hi all,


I have read here numerous times how someone would post a PS and someone else would say, oooh, with those values that choke might ring. Or there isn't enough load to make that choke work etc etc.


Could someone explain exactly how a choke:

a) works
b) rings
c) works correctly/loading


and also if you have time, how to work out what H, DC resistance, mA you need with examples if possible!


I hope thats not too tall and order! There just seems to be loads of good vibes around a choke, I've seen people saying how much better a CLC is instead of a CRC etc.

Enlighten me!!

Fran

[Miles Prower]

Could someone explain exactly how a choke:

a) works -- Depends. If this part of a CLC filter, the choke and the second capacitor form an LC low pass filter. This gives a 12db(v) rolloff above its turn-over frequency. That critical frequency is equal to:

f(c)= (LC)^-0.5 (L in henries; C in farads; f in Hz)

If it's a choke input filter, the choke serves as a constant current source into the load during the time the rectifier diodes are off. This only works with full wave rectification since the choke can't keep sourcing current long enough during the half wave off duration.

b) rings -- Coils are the leat perfect passive component. They include not only DC resistance, but also stray capacitance sinceeach turn is at a different potential from its neighbours. That leads to displacement currents, and displacement is capacitive by nature. Given that stray capacitance and the inductance of the windings, you can expect natural series and/or parallel resonances. When shock excited, the coil will ring just as if it were a tuned circuit (because that's what it is).

c) works correctly/loading -- See: Rectifier Applications Handbook for further details.

[MRupp]

I would recommend that you buy a copy of Morgan Jones' "Valve Amplifiers". Amongst other goodies it has a whole chapter on power supplies and chokes. I don't think anyone would want to post several pages of text to give you a full appreciation of the subject let alone enlighten you.

[G]

Originally posted by Miles Prower
Could someone explain exactly how a choke:

a) works -- Depends. If this part of a CLC filter, the choke and the second capacitor form an LC low pass filter. This gives a 12db(v) rolloff above its turn-over frequency. That critical frequency is equal to:

f(c)= (LC)^-0.5 (L in henries; C in farads; f in Hz)

If it's a choke input filter, the choke serves as a constant current source into the load during the time the rectifier diodes are off. This only works with full wave rectification since the choke can't keep sourcing current long enough during the half wave off duration.

b) rings -- Coils are the leat perfect passive component. They include not only DC resistance, but also stray capacitance sinceeach turn is at a different potential from its neighbours. That leads to displacement currents, and displacement is capacitive by nature. Given that stray capacitance and the inductance of the windings, you can expect natural series and/or parallel resonances. When shock excited, the coil will ring just as if it were a tuned circuit (because that's what it is).

c) works correctly/loading -- See: Rectifier Applications Handbook for further details.

I'll even give up a example for use in the discussion.

[pchw]

In MJ's book, he mentioned that it would need approximately
(RMS input voltage) / (inductance) minimal load current in the choke input filtering. Is this still applicable to CLC? If so, what should be the RMS input voltage in this case?

Thanks!!

[Miles Prower]

Originally posted by pchw
In MJ's book, he mentioned that it would need approximately
(RMS input voltage) / (inductance) minimal load current in the choke input filtering. Is this still applicable to CLC?

No. The choke of a choke input filter tends to act almost as a CCS. Therefore, it requires a certain minimum of inductance to work right. If you're doing a CLC, then the choke and the capacitor behind it makes a second order LPF. You'd then size it based on how much ripple you beed to get rid of, and set the turn-over frequency accordingly:

w(c)= (LC)^-0.5 (L: henries; C: farads; w: rads/sec)

Also, make sure that w(c) occurs well below whatever minimum signal frequency you're expecting so as to avoid "filter bounce" -- a resonant phenomonon that'll absolutely ruin your voltage regulation.

The operating conditions are different as well. CLC work isn't so demanding as LC work.

[pchw]

Thank you, Miles!!

[woodturner-fran]

OK,

thanks to you all for the info and help. I went to my trusted MJ and got the info and things are a lot clearer now!


Its a steep learning curve - and it ain't getting shallower


Fran