Linear power supply chokes

[John B]

I was doing some simulations for a linear power supply and was thinking about the peak currents and conduction angle through the bridge rectifier. In this case it's just a basic 50Hz transformer, full wave rectified and fed into about 24mF of filter caps. Naturally this makes the current pulses through the transformer short and sharp, so I experimented adding an inductor of around 10-25mH between the positive of the bridge rectifier and the main filter bank. Although the current waveform is distorted, as the inductance of the choke is increased, the peak current flowing through the transformer approaches the DC current at the output of the filter caps.

While I can find a few references, it doesn't seemed to be used that much in a linear context, that is, the choke being used as an energy storage device along with the caps.

So when and where is it worth bothering with? I've wound a 20mH inductor with 1mm wire for use.

Pros:
Peak current through diodes and transformer is reduced.

Cons:
Adds impedance and so the maximum DC current is reduced compared to capacitors alone.

Anything else to add?

[ledtester]

If I understand it correctly, this information from Hammond Manufacturing suggest that adding an inductor will increase the DC current capacity over just using a capacitor:

http://www.hammondmfg.com/pdf/5c007.pdf

[John B]

Yep, what I'm talking about is like the last example "Full Wave Bridge Input Load", but for a given resistive load, an increase in choke inductance results in less ripple, but a lower DC voltage compared to caps alone.

[IanB]

It seems the choke + capacitor arrangement was (commonly?) used in HV DC supplies for valve/vacuum tube equipment, but was apparently not adopted or considered necessary for low voltage supplies. Maybe because the LV supplies usually have a voltage regulator for ripple rejection, while this was absent in the HV supplies.

Although a filter choke may may allow a smaller transformer to be used, that must be traded off against the size and weight of the choke. It may be that the trade-off isn't worth it.

[SeanB]

Choke on valve HT was there for higher anode current applications, as rectifier tubes really do not have a very good overload characteristic, and have very poor pulse rating compared to a silicon rectifier diode. Thus the choke, to reduce peak pulse currents, and thus losses across the rectifier, and as a bonus lower ripple on the supply lines, so the poor power supply ripple rejection of the amplifier was not going to introduce hum with high ripple voltage. Remember on valves the other grids are also connected nearly directly to the power rail with almost no decoupling. Valve rectifier anode drop is current dependent, and even a small rise in current leads to a very much hotter running anode, and as they typically ran at 500C at full rated current, just a small increase could get them to 800C where they glowed visibly red, and then they would start to have secondary emission, go into thermal runaway and kill both the rectifier tube and the power transformer.

On lower voltage higher current supplies the choke however gets very big very fast, in fact the best way to have a choke power supply is to design it into the power transformer, making it larger, thicker wire and have some leakage inductance in it. Thus the ferroresonant transformer, which can have a pretty low ripple supply voltage at a very high current, and as a bonus the conduction angle of the diodes is very large, as the high harmonic distortion of the output makes it a more trapezoidal waveform than a sine wave. Added bonus it is self limiting current wise as it will tend to be a constant power over a certain current, but under that is is relatively constant voltage.

They made great wet cell battery chargers, as they were able to charge a fully discharged bank without any attention, and would float the cells once charged. Only disadvantage was they ran hot at high power, and had a fixed no load dissipation that could be a good part of the full load rating.

[gbaddeley]

While I can find a few references, it doesn't seemed to be used that much in a linear context, that is, the choke being used as an energy storage device along with the caps.
So when and where is it worth bothering with? I've wound a 20mH inductor with 1mm wire for use.
Pros:
Peak current through diodes and transformer is reduced.
Cons:
Adds impedance and so the maximum DC current is reduced compared to capacitors alone.
Anything else to add?

Does your simulation include losses in the power transformer, diodes, capacitors and interconnections? If you actually build the circuit and measure the current pulses (eg. across a resistor in series with the diodes, say less than 0.1 Ohm), you will probably find the current peaks are less than the simulation and are stretched out in time. Chokes are rarely used in low frequency low voltage rectifiers because they are not worth the cost and size. Power supplies generally work fine without them and can easily handle the peak currents.

[David Hess]

SeanB covered why choke input power supplies were common for tube power supplies.

However they were also found in low voltage high current linear power supplies which used thryristor based preregulators as shown below.

[John B]

While I can find a few references, it doesn't seemed to be used that much in a linear context, that is, the choke being used as an energy storage device along with the caps.
So when and where is it worth bothering with? I've wound a 20mH inductor with 1mm wire for use.
Pros:
Peak current through diodes and transformer is reduced.
Cons:
Adds impedance and so the maximum DC current is reduced compared to capacitors alone.
Anything else to add?

Does your simulation include losses in the power transformer, diodes, capacitors and interconnections? If you actually build the circuit and measure the current pulses (eg. across a resistor in series with the diodes, say less than 0.1 Ohm), you will probably find the current peaks are less than the simulation and are stretched out in time. Chokes are rarely used in low frequency low voltage rectifiers because they are not worth the cost and size. Power supplies generally work fine without them and can easily handle the peak currents.

Playing around with the series resistance is what made me not worry about it in the end. I've actually built up the unregulated power board now. There's enough parasitic resistance between all the components you mentioned (and relays).

Making efficient use of the transformers and minimising losses is probably an unnecessary concern on this reused and recycled shenzen market special power supply.

[Kleinstein]

In the tube radio days the electrolytic caps where relatively expensive compared to a choke. In addition the choke saved on the rectifier and transformer.  The cost balance tends to be even better with smaller voltage as the caps get more expensive per energy stored. With chokes the costs go up slower than linear to higher power - so it is more attractive on higher power. With higher power the parasitic resistance also goes down, making the current peaks even worse.

Sometime the resistance of a fuse before the rectifier also has a positive effect.

For a time those chokes were standard in PC power supplies to get an acceptable PFC. That was the time be active PFC was required.