Warning: Creating default object from empty value in /homepages/11/d132647312/htdocs/Amfone/mkportal/include/SMF/smf_out.php on line 47
Ripple Reduction for Choke Filter DC Power Supplies




 
The AM Forum
September 25, 2017, 09:32:51 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: [1]   Go Down
  Print  
Author Topic: Ripple Reduction for Choke Filter DC Power Supplies  (Read 350 times)
0 Members and 1 Guest are viewing this topic.
WA4WAX
Member

Offline Offline

Posts: 158


« on: September 09, 2017, 12:09:52 AM »

Try this one.

Resonate the choke, and put 40 or 50 uF on either side.    Let me know what happens.

BTW, resonate at 120 Hz, not 100 Hz.


Page 16 on your pdf reader.


http://americanradiohistory.com/Archive-Practical/Wireless/50s/PW-1957-09.pdf
Logged
WBear2GCR
Member

Offline Offline

Posts: 3382


Brrrr- it's cold in the shack! Fire up the BIG RIG


WWW
« Reply #1 on: September 11, 2017, 06:12:18 PM »

Of course this is not "new'.

Thought I'd mention that one can simulate these filters in LTSPICE, which is a free download.
A bit of a learning curve, but you can tweak the values and see the resulting graph!

These are a class of filters now known as "elliptic" or "Cauer" filters, iirc.

Basically they work by creating a notch filter to "suck out" energy at some frequency.
The notch is usually very sharp, and can be very deep (as in >60dB) compared to the usual
filter of a cap-choke-cap with is rolling at a slope of 18dB/octave. So for practical values,
one might have maybe -36dB of reduction at 120Hz, perhaps more - but the addition of
the paralleled L+C creating the notch drops some frequency (chosen by the values)
down into a deep crevasse. The response higher up in frequency lifts and more or less
returns to the original slope of the basic filter.

The attached file is NOT exactly the same, since it uses two elements, but it shows more or
less how it works - look carefully at the dB scale!

I had an LTSPICE screen capture of an actual LF filter I had designed and used, but can't find
that jpeg at the moment. Sad  In the example below, the response pops back up too far, in my
example it followed the same black line after the deep dip...



* RESONATED-vs.-STD-Filter.jpg (370.25 KB, 1400x1050 - viewed 23 times.)
Logged

_-_- bear WB2GCR                   http://www.bearlabs.com
WBear2GCR
Member

Offline Offline

Posts: 3382


Brrrr- it's cold in the shack! Fire up the BIG RIG


WWW
« Reply #2 on: September 11, 2017, 06:26:59 PM »

Here's another example.

This was a simulation for my Globe Champeeon 350.
The stock speech amp features a splatter filter (iirc) after a clipper. Ick.

So I foolishly decided to see IF I could use the stock inductor and make a more or less
brickwall LP (low pass) for the audio passband. You can see the results. The bright GREEN
trace is the frequency response.

The other trace is the phase. THAT is a problem... or could be, since the thing swings
the phase enough to make the positive peaks become the negative peaks! depending on
frequency. Otoh, the amplitude response would be peachy keen! Cheesy

But you can see the effects of the elliptic filter, the FAST drop in the amplitude. The values
around the main components are extremely sensitive to Q, which also means the impedances
and given the relatively poor choice of inductance value (aka don't use the stock inductor!!) not
going to be easy to duplicate the response in a practical circuit. Although, maybe...


* ELLIPTIC FILTER 4H+SERIES-1.pdf (922.38 KB - downloaded 22 times.)
Logged

_-_- bear WB2GCR                   http://www.bearlabs.com
WA4WAX
Member

Offline Offline

Posts: 158


« Reply #3 on: September 11, 2017, 07:40:47 PM »

I know of Cauer filters..........and more!

I thought it was a nice way to clean up power supply hum.  Cheap and easy, but use adequate voltage rating.  No comments on using it in your audio circuit.

In practical work, one could get a bag of caps, and select the one that gives best rejection at 120 Hz.

I love notch filters.......use 'em in MRI work.

Matt :-)

Logged
Opcom
Patrick J. / KD5OEI
Contributing
Member
*
Offline Offline

Posts: 6693



WWW
« Reply #4 on: September 11, 2017, 10:40:44 PM »

15H can be a large value to resonate for the purpose depending on the DC resistance and the current through the choke.
Logged

Radio Candelstein - Flagship Station of the NRK Radio Network.
WBear2GCR
Member

Offline Offline

Posts: 3382


Brrrr- it's cold in the shack! Fire up the BIG RIG


WWW
« Reply #5 on: September 12, 2017, 08:13:44 AM »

I know of Cauer filters..........and more!

I thought it was a nice way to clean up power supply hum.  Cheap and easy, but use adequate voltage rating.  No comments on using it in your audio circuit.

In practical work, one could get a bag of caps, and select the one that gives best rejection at 120 Hz.

I love notch filters.......use 'em in MRI work.

Matt :-)



Send me your unused/obsolete MRI amplifier chassis!!

Sure in practice a bag of caps is a viable way to go... but unless you have THE IDEA of the effect of the
cap vs. frequency one might NOT find the actual good value to use! A little bit of trimming values
could be worth 20+dB or more of filtering AT 120Hz. The big advantage in using with PS it that the frequency
of interest is EXACT and KNOWN!! And, it doesn't move!!! Yay! Cheesy

The filters work precisely the same way for audio as it does for a PS, so the examples are to show graphs
that illustrate how the notch looks and works. The image with the orange border shows the name of a neat
freeware simulation package that does this stuff fairly easily.

                               _-_-
Logged

_-_- bear WB2GCR                   http://www.bearlabs.com
Opcom
Patrick J. / KD5OEI
Contributing
Member
*
Offline Offline

Posts: 6693



WWW
« Reply #6 on: September 14, 2017, 01:12:37 AM »

the result from LTspice was <1V peak ripple. The low frequency noise seen there over a period of seconds is an artifact of the simulation. The cap and choke are tuned to 125Hz.

I think the 0.25uF value of C for 100Hz with the 15H choke in that article might need checked. 0.25 resonates at 100Hz more likely with 10H. But I could be wrong on that. Anyway the resonant choke is a very good solution and a smaller H choke can also be used with larger C - just keep in mind it all has to pass the 120Hz ripple frequency with as little loss as possible or the effect of the scheme will be diminished.


120Hz
5H/0.35uF
3H/0.586uF
the numbers are from
http://www.1728.org/resfreq.htm



* resonant-150V-50mA.jpg (321.22 KB, 1170x992 - viewed 23 times.)
Logged

Radio Candelstein - Flagship Station of the NRK Radio Network.
Pages: [1]   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.049 seconds with 18 queries.