How do you calculate choke size in a power supply?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Engineering has many constraints. :) When I looked up what iron costs at Allide, it selected for me reall quick.

OK, when I get a chance I will look it up in Jones book if someone else does not beat me to it. I think he had some guidelines. I was also told the best reference was the old ARRL handbook.

It matters how many poles in the filter. C-L-C, or L-C-L-C etc, if droppers or not and so on. Of course you need sufficient current ratings.
 
Like math? This is not a simple answer or question.

There is a minimum current load that must be met to make the choke work and then there is the output capacitor size, as well as maximum load currents.

The minimum current is aproximately equal to:

Imin (mA) = [Vin(RMS)] / [L(H)]

You can calculate the ripple of the choke-cap system as follows:

[Vrip(RMS)] / [VDC] = 1 / [ 6 * sqrt(2) * 2 * PI * f * L * C]

L = inductance
f = AC line frequency
C = capacitance
VDC = DC voltage at cap
PI = 3.14...

There is more to the process. Maximum peak current through the choke is:

Iac (pos peak) = [0.544 * Vin (RMS)] / [2 * PI * f * L]

You need to consider other things with the design, such as snubber networks.

When I built my amp I bailed on the idea I had of originally adding a choke. This is just my opinion, but chokes were very popular decades ago and for good reason. Banks of large capacitance for filters simply were not practical. Today you can buy huge capacitors which simply did not exist in the 1940s, '50s, and '60s. The solution was a filtering choke (and a slide rule).

Again, my opinion... Today a brute force capacitance approach is easy and it offers another advantage; reserve power.

AC Ripple, the ultimate target we are trying to slay, can be calculated by:

Delta V = I / [2 * f * C]

You can design a multi-stage filter network and work out the AC ripple at each stage. My power supply is well under 1% and the cost to build it is cheaper than using a choke (and uses less real estate).

I am sure others will chime in and perhaps correct me if I have strayed down the wrong path somewhere. :)
 
Disabled Account
Joined 2010
"How do you calculate choke size in a power supply?"

It depends on where the choke is located. If you're designing a choke-input filter, the choke must be the first filter element and must have a certain minimum value called the "critical" value. This value is calculated by:

L (henries) = E (volts) / I (ma)

If the choke is located after the first filter cap, or if more output voltage will be required than is available from a choke-input filter, the choke can be substantially smaller.

Incidentally, there are more criteria necessary to the design of a good supply than simply output voltage and ripple. PSUD can predict those things. A supply can also ring and exhibit high Z resonances at its output when excited by the amplifier's dynamic current demands. Examining these characteristics requires looking into the supply from the output side, an exercise for which I use LT Spice (free from National).
 
Last edited:
For a power supply with a fairly constant load, preamp or class A amp, this formula works great and is easy to remember. Figure out the load resistance, E/I, then divide that by 1200. the result is in Henries. Example; 400VDC with 20ma of current needs a 16.6666666H choke.

Craig
 
Disabled Account
Joined 2010
Figure out the load resistance, E/I, then divide that by 1200. the result is in Henries.

Just want to point out that this is nearly identical to the forumla I posted. It just looks different because the two formulas use different units (Amps vs. mA). In essence, the above formula divides by 1,200, where mine divides by 1,000. Dividing by 1,200 produces a smaller choke and less margin for error due to component tolerances, but it is mathematically correct.
 
"How do you calculate choke size in a power supply?"

It depends on where the choke is located. If you're designing a choke-input filter, the choke must be the first filter element and must have a certain minimum value called the "critical" value. This value is calculated by:

L (henries) = E (volts) / I (ma)

If the choke is located after the first filter cap, or if more output voltage will be required than is available from a choke-input filter, the choke can be substantially smaller.

Incidentally, there are more criteria necessary to the design of a good supply than simply output voltage and ripple. PSUD can predict those things. A supply can also ring and exhibit high Z resonances at its output when excited by the amplifier's dynamic current demands. Examining these characteristics requires looking into the supply from the output side, an exercise for which I use LT Spice (free from National).

A small correction: LTspice is free from Linear Technology Corp at linear.com (not national.com).
 
chokes also have a resistance in their specifications.
I understand that thicker wire will give a lower resistance.
But, given two chokes of same/similar price, would it be better to get a lower resistance with a lower H or a higher resistance with a higher H?
Or is it more complicated than that.
(For use in a simple SE with a auxiliary Cap ((motor run cap)) ).
Thanks,
Paul
 
By your wallet thickness.....Engineering has many constraints. When I looked up what iron costs at Allide, it selected for me reall quick.

That's pretty much how I do it.

If the power supply is choke input, then it makes sense to use a decent choke that won't buzz. Old potted units ripped out of military surplus seem to work well.

If it is CRC there is no choke. Stepping up to CLC is almost always an improvement even if the choke is a $8 Triad from Allied.

You need to consider the DC resistance and its effect on your B+ voltage. PSUD is a good tool here.

The choke must be rated for the voltage is will see. Need a choke for a 1000 volt supply, but they are too expensive, use a lower voltage choke and put in the negative side of the supply so it sees a low voltage to ground.

Beyond this use as much inductance as your wallet and physical constraints will allow. Yes, there are cases where too much inductance can cause low frequency instability, but I haven't found one yet, probably due to budget or size limits.

Need a choke to look pretty? Stuff one of those ugly Triads from Allied into the end bells off an old power transformer.
 
For a power supply with a fairly constant load, preamp or class A amp, this formula works great and is easy to remember. Figure out the load resistance, E/I, then divide that by 1200. the result is in Henries. Example; 400VDC with 20ma of current needs a 16.6666666H choke.

Craig

But this formula is valid for a choke input filter to work as a choke input filter (and not as a capacitor input), isn't it? The 1200 should be for mains of 60Hz, for countries with 50Hz mains one should use 1000!?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.