# Filter Choke for power supply

#### Y2KEDDIE

Sep 23, 2012
259
How does one calculate the value of a filter choke used in a dc supply.

I want to build a dc 600vdc supply , pi filter, full wave bridge rectifier, 60 Hz mains supply .

I have 47Ufd 800 VDC capacitors. How do I calculate inductance for the choke.

#### Arouse1973

Dec 18, 2013
5,177
Please show us an application circuit diagram.
Thanks

#### Colin Mitchell

Aug 31, 2014
1,416
At 60Hz the choke will be enormous. It is much better to use a resistor and larger capacitors.

#### Y2KEDDIE

Sep 23, 2012
259
I remember the days when one would use a type 80 rectifier, feeding a pie, low pass filter (an 8ufd capacitor,10 Hy choke, and another 8 ufd capacitor. Sometimes the field winding of an electromagnetic speaker served as the choke. I was wondering how these values were calculated.

#### Colin Mitchell

Aug 31, 2014
1,416
"I was wondering how these values were calculated."
They weren't calculated.
They only had 8 mickey 16 and 24 mickey to deal with. They just put the values in and listened to the hum.
The speaker magnet was already set by the speaker manufacturer and you just tried the others.

#### (*steve*)

##### ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,505
But they *can* be calculated. If you're actually interested in a theoretical analysis, go look up pi filters and play with the math.

As has been mentioned, low frequency LC filters are rarely used because power supplies these days tend to be low voltage and high current (cf high voltage/low current for valve circuits). The need for larger diameter wire and larger cores at these low frequencies is both costly and inefficient. Switchmode power supplies employ a related technique, but due to the much higher frequency can get away with far smaller components.

#### Colin Mitchell

Aug 31, 2014
1,416
"But they *can* be calculated. If you're actually interested in a theoretical analysis, go look up pi filters and play with the math."

Why bother working anything out when you can simply "turn on the radio."

#### (*steve*)

##### ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,505
"But they *can* be calculated. If you're actually interested in a theoretical analysis, go look up pi filters and play with the math."

Why bother working anything out when you can simply "turn on the radio."

Ummm... because the question was "how do you calculate..." Maybe Y2KEddie is interested in the calculations.

#### hevans1944

##### Hop - AC8NS
Jun 21, 2012
4,723
It has been many years, circa 1966, since I constructed a plate power supply using a 5R4-GY vacuum tube rectifier, high-value electrolytic capacitors, and the biggest choke I could get my hands on. I think it was about 10 Hy and good for at least 1 A. This was the power supply for my 75 W homebrew Novice CW transmitter based on the venerable RCA 6146B beam power-pentode vacuum tube.

For amateur radio purposes, brute force was the name of the game. The ripple frequency of a full-wave rectifier connected to the 60 Hz mains is 120 Hz. A 10 Hy choke has a reactance of 2 (pi) (120) (10) ohms or slightly more than 7500 ohms at the ripple frequency. It is important that the DC resistance of the choke be substantially less than the reactance for good filter efficiency. I don't remember what my choke resistance was, but it did not cause any appreciable voltage drop when I loaded up the transmitter, nor did I ever get any signal reports complaining of "hum" on my transmitter carrier.

As @Y2KEDDIE mentioned, early loudspeakers used a field coil instead of a permanent magnet and this coil was usually part of a pi-network filter for the power supply. This was my first introduction to large inductors used as power supply filters. I was somewhat amazed that I was able to draw a fat, hissing, spark between the choke and chassis ground. I got away with that because vacuum tube rectifiers are inherently current-limited. Wouldn't recommend trying this "experiment" with solid-state rectifiers.

I presume you realize this type of power supply is virtually obsolete, but it sounds like a good nostalgia project. The 47 μF, 800 WVDC, capacitors (very hard to come by with that voltage rating!) will make a fine pi filter with a 10 Hy choke for a 600 VDC power supply. Depending on the load current and maximum ripple you desire, you may want additional capacitance. I used a series vacuum tube regulator in my power supply to provide a variable constant-voltage output, and the negative feedback probably went a long way toward suppressing ripple in the output.

Be careful with your power supply. It can be lethal.

#### Y2KEDDIE

Sep 23, 2012
259
Thanks guys, GPG, give me a few day s to absorb the material.

Yes, Hop, I built the same transmitter in the early 60"s. I don't remember where I got the schematic. I do remember scrounging parts, and as Colin suggested, I think I just used what ever I could find. I remember the 5R4 rectifier, and a dual 40 ufd can cap. I found a 10 Hy 500mA choke and used it because I wanted good regulation to prevent oscillator chirp when keying. (At least that seemed like a good idea at the time.)

The physical size of the 10 HY 500 mA choke was about that of a pint canning jar.

That 40 meter HB rig is long gone
.
I still am interested in calculating values . I've been looking at published designs of various equipment, and the design values are all over the place.so its hard to determine: did they just use what they had, or was there some thought behind the design.

Naturally, regulation, load, ripple, etc. are considerations, but there should be a starting point..

#### duke37

Jan 9, 2011
5,364
The 'Electronic Designers Handbook' by Landee, Davis and Albrecht (1957) gives graphs of filter attenuation for various types of filter and rectifier toobs.
It even compares an LC filter with an LC, LC, LC filter using components of one third the value.

One of the practical problems is that the inductance of the choke(s) will vary with the DC current. Swinging chokes can be used to get more inductance when the current is low.

It does not deal with transient response which can presumably be tackled with fat capacitors.

#### Colin Mitchell

Aug 31, 2014
1,416
You are much better of to use R instead of L. The only reason they used an inductor:
IT WAS FREE !!!!

#### (*steve*)

##### ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,505
The only reason they used an inductor:
IT WAS FREE !!!!

Are you serious?

Maybe I should use resistors instead of inductors in switch mode power supplies too?

#### Colin Mitchell

Aug 31, 2014
1,416
There is a huge difference between 120Hz and 180kHz and 100mA and 5 amps.

#### duke37

Jan 9, 2011
5,364
You are much better of to use R instead of L. The only reason they used an inductor:
IT WAS FREE !!!!
This is rubbish. Inductors are much more expensive than resistors and the stability of the output is less.
The power loss is also expensive and deleterious to adjacent components which are heated. The advent of very large electrolytic capacitors has shifted circuitry towards resistive smoothing but equipment using 100W or more still seem to use inductors.

Inductors are heavier than resistors but if a choke input filter is used, the mains transformer can be made lighter since the current is not so peaky.

Resistors are also less effective than inductors since they do not store energy.

#### dorke

Jun 20, 2015
2,342
Y2KEDDIE,
Let's start at the beginning ,shall we.

What Current does your PS need to supply?
What is the allowed ripple?

The above is absolutely needed in order to give you a solution.

#### Y2KEDDIE

Sep 23, 2012
259
A particular power supply I've been thinking of building is : 180 VDC for my LM-13 Freq. meter/ signal source.
It needs 180 VDC at 6ma. I have a 150 VAC transformer which gives me around 200 VDC at the output of my bridge.
I was thinking of a pie filter: C1 =47 ufd, R1= 1K, C2=10 ufd., and then a R2 = 5k series resistor to drop down to 180VDC.

What filtering would you use, and why?

My thought is to use RC filter for light constant loads. A high value R is picked so the AC ripple component is developed across the resistor,

My thought is large inductors are used in high voltage and power applications where regulation is important. I read current surge is a consideration when using Thoriated tungsten rectifier tubes ,and this was the main reason for using large chokes. With solid state rectifiers of today, maybe this is why there is a decline in using power inductors.

#### dorke

Jun 20, 2015
2,342
For relatively low currents the usage of resistors instead of inductors in filtering is fine.
6ma is low current.
a resistor of 1k will drop 6v on it ,and dissipate very little power (0.036W).
If that voltage drop is o.k with you than go for it(You hasn't specified the allowed ripple yet).
Note:
with a 150V secondary voltage you get 212V peak voltage.
and you should take into account the mains line voltage variations in your area as well.

Last edited:

#### GPG

Sep 18, 2015
452
I was thinking of a pie filter: C1 =47 ufd, R1= 1K, C2=10 ufd., and then a R2 = 5k series resistor to drop down to 180VDC.
Use 5K as R1 and 10μ as the input, and 47μ as the output C. No output R
Better ripple rejection. At such low currents and a C input filter the input to the filter will be close to Vpk. Placing the larger C after the R will reduce ripple but at such low current it probably doesn't matter much. Voltage regulation with change of load current is the major concern

Last edited:

P
Replies
1
Views
1K
jhill
J
Replies
2
Views
1K
Replies
5
Views
626