diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Tubes / Valves (http://www.diyaudio.com/forums/tubes-valves/)
-   -   How to chose the right choke (http://www.diyaudio.com/forums/tubes-valves/84073-how-chose-right-choke.html)

Mikael Abdellah 1st August 2006 06:03 PM

How to chose the right choke
 
How do you chose the right choke for your power supply??? I have tried to search, because I suspect the question has already been answered before but I cant find a good answer. Is more inductance always better?

I have tried a resonance calculator but Im not sure thats the way to go or how to use the numbers. When I model the power supply with duncans psud2 almost everything I put together looks fine. I need to make a good power supply for my quickly evolving tube preamp. I have measures the circuit to 3.04mA at 250 VDC. When I read about the hammond chokes it says that the inductance value increases if the load is lower than rated. How much does it go up??? How do I chose the right cap? If I use a VERY big cap after the choke it seem like I get a free soft start for my amp. Is this good? When I model it with very large cap values I never seem to reach full voltage. After roughly 2 minutes the voltage is still climbing in the simulation. Can someone help me or at least point me in the right direction.

Jeb-D. 1st August 2006 06:35 PM

The more inductance is better, but there can come a point where it's overkill. And the choke will be bigger, heavier and more expensive than it needs to be. Make sure the DC resistance isn't high enough to hog you supply voltage. Generally, the larger the cap you have after the inductor, the smaller the inductor you can use. Likewise, the bigger the inductor the smaller cap can be used.

Are you using a CLC or LC filter setup?

Mikael Abdellah 1st August 2006 06:51 PM

I have tried to model LC, CLC and RCLC. All with a 20H hammond choke with a dc resistance of 1666 ohm. When I put a 2200uF electrolytic cap behind it, it toke a very long time for the power supply to reach full voltage. Since the chokes are so cheap, 20H is no problem economically. So far the RCLC seem to be the best because I could change the value of the series resistor to fine tune the voltage. I saw somewhere that the wrong values in the wrong place could cause ringing but that has not been the case in any of my simulation.

cerrem 1st August 2006 08:56 PM

The first step in filtering a power supply is knowing what spec you want or need for voltage ripple and current ripple..
For example, the desired voltage ripple reduction is a ratio.... You express this in terms of -dB...20*LOG(ratio)......
A LC filter is considered a DOUBLE POLE..this POLE occurs at 1/(2*pi*LC)^1/2 ....... This LC POLE roll-off, attenuation, at -40dB per Decade....The main objective is to reduce the ripple voltage most comonly at 120Hz....or 100Hz if your mains are at 50Hz..
SO for example, in a 60Hz mains application with a full wave rectifier...set your LC pole at 1.2Hz and then you will have a -80dB attenuation at 120 Hz.....
The LC combinations are endless for the same frequency POLE...typically you want to use common value for economy.... I choose the inductor for the maximum current ripple I am interested in getting..Maximum current ripple I would go to 30% of peak... If your first input cap is too big then your diodes will conduct for a very short conduction angle, which would make big current transients, so watch that...
Make sure the inductor can handle maximum DC current demand from circuit..this way it won't saturate, although there are some cases were you can operate in sat but thats another story...
Each time you add additional RC Poles afterwards, these are -20dB per decade attenuation, you further reduce the ripple..

Chris

hagtech 2nd August 2006 05:39 AM

Mike, try my choke calculator (when used as rectification) at:

www.hagtech.com/theory.html#choke

It will give you nominal values for a critically designed and tuned supply filter. The first thing you have to achieve is a minimum inductance. Anything lower than this value will cause a discontinuous current flow out of the rectifier (bad). Too much inductance doesn't really help either as you end up burning a lot of power with the voltage drops. Better to have two LC sections in series. The second L is not critical and can be any value.

The RLC values interact with each other and are dependent upon load.

jh:)

Mikael Abdellah 2nd August 2006 06:08 AM

wow! thats a really neat calculator hagtech. thank you very much. and thank all of you for helping me.

EC8010 2nd August 2006 07:06 AM

If you're only drawing 3.04mA a choke is not your best choice. Use RC filtering or a regulator. Chokes make sense once the current exceeds about 50mA.

Jeb-D. 2nd August 2006 08:47 AM

Quote:

If you're only drawing 3.04mA a choke is not your best choice. Use RC filtering or a regulator. Chokes make sense once the current exceeds about 50mA.
Ditto, for that little current, you should just go with a CRC or RC filter. You could probably even get away with just a capacitor.

Vripple= DC current/(frequency x capacitor value)

Bridge rectifying will allow a capacitors of half the value, to get the same results.


Use soft recovory rectifiers, either tubes or Hexfred diodes(if going with a solid state power supply). I've found cheap diodes will have a harsh recovery, causing a spike that is audible, even with extensive filtering.


All times are GMT. The time now is 06:01 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2