Choke input vs. Capacitor input filter

[Austin]

Hi, I am trying to figure out the benefits and drawbacks of each, so far it looks like the choke input is better according to what I have been reading but what do you guys think? Like is it better sometimes to put the choke after the first cap?

[loudthud]

The big disadvantage of the choke input is that you need a minimum current draw or the voltage rises to the cap input value. Another is that spikes can occur that will kill solid state rectifiers. The advantage is the you can get more power (~10%) from a given transformer although at a lower voltage. Ripple can be lower and more like a sine wave than a sawtooth.

The (Greek letter) pi filter is the best at getting rid of ripple. You don't need a 5 or 8 Henry choke unless you want to use 10uF or 22uF caps. I built a 150W bass amp with pi filtered B+. First cap was 100uF, choke was 300mH, final cap was 235uF. Ripple at full power less than 1 volt peak to peak. For a guitar amp, low ripple doesn't sound quite right. On the mellow side, not as agressive or in your face. An example is the 70's Sunn Model T.

[TransLucid]

There is a good discussion of filters in Merlin's Power Supply book. One other disadvantage of choke input filters not mentioned by loudthud is that the full supply current flows through the choke, which requires a big one...

[Austin]

My choke measures about 200 ohms dcr so maybe this isn't the best choke to use for choke input. Is it true that the required current rating on a choke is less if you use a capacitor first?

[R.G.]

Choke input filters offer good voltage regulation on their output and make life easy on rectifiers and transformers.

They do this at the costs of (1) chokes are expensive (2) big chokes - that is, storing a lot of energy for filtering - are even more expensive (3) big chokes are hard to find (4) "good" values of choke inductance and resistance are not necessarily easy to find or design (5) output voltage varies *a lot* at low loads, so you're forced into wasting power in bleeders to keep the output voltage from rising precipitously at low loads (6) oh, yeah: big chokes are heavy and space-eating.

Did I mention that they're big, expensive, hard to find and design, and lose a lot of your output DC voltage for you?

[Tom Phillips]

...Is it true that the required current rating on a choke is less if you use a capacitor first?

No. The choke must still be rated to handle whatever current the circuitry is drawing through it. A common guitar amp topology, however, draws the main power tube plate current from the first stage capacitor of the power supply. If a choke is used next in line then the current flowing through the choke is only that required to run the circuits down line. These are the power tube screen supply, the PI and the preamp circuits which, in total, are small compared to the power tube plate supply current. Maybe that is the origin of "required current rating on a choke is less if you use a capacitor first" phrase. In the example I described it's a result of the design of the whole amp that there is a capacitor before the choke and the choke current rating can be less than would be needed for a choke input filter. The important point is the current you draw and which node of the power supply you draw it from. If all the circuitry is connected after the choke then the choke must be rated for higher current. If the high current plate supply is drawn off before the choke then the current that passes through the choke is relatively small.
Cheers,
Tom

[trobbins]

Sometimes you may have a particular power transformer, but with capacitor input rectification the HT voltage would be too high for your application. Adding or using a choke input rectifier can provide a more acceptable HT voltage. For example you may be putting together a one-off design of hotch-potch parts. Or a restoration job shows that the original design operated the valves at or above max rating, or present-day replacement valves can't reliably cope - I've added a choke to a mid-1950's 70W amp for this scenario.

With a choke-input rectifier, the requirement for enough idle loading does require some thought. It may be no problem at all if your filter capacitor voltage ratings can accept the peak level - electrolytics nowadays are much cheaper and smaller, and so this may be no issue for some - I've even used a zener diode on the preamp stage to protect 'lower voltage' capacitors for any short-term start-up situation.

[Austin]

Thank you all very much.These are the organ amplifiers I am working on to make hifi or stereo pa from, have two chokes and I assumed choke input however one choke is smaller and has three taps, but I wonder if I could use it as a plate load for a pentode driver tube and use that tap for the screen? Or is it to be used like two chokes, with a cap on that middle tap?

[Austin]

No. The choke must still be rated to handle whatever current the circuitry is drawing through it. A common guitar amp topology, however, draws the main power tube plate current from the first stage capacitor of the power supply. If a choke is used next in line then the current flowing through the choke is only that required to run the circuits down line. These are the power tube screen supply, the PI and the preamp circuits which, in total, are small compared to the power tube plate supply current. Maybe that is the origin of "required current rating on a choke is less if you use a capacitor first" phrase. In the example I described it's a result of the design of the whole amp that there is a capacitor before the choke and the choke current rating can be less than would be needed for a choke input filter. The important point is the current you draw and which node of the power supply you draw it from. If all the circuitry is connected after the choke then the choke must be rated for higher current. If the high current plate supply is drawn off before the choke then the current that passes through the choke is relatively small.


Cheers,
Tom

I asked that because of what I read here: "If, on the other hand, you are selecting a choke for a capacitor input supply (such as the typical Marshall or Fender design), then the requirements are relaxed quite a bit." from
Chokes What did he mean by that then? The requirements for the amount of inductance?

[R.G.]

I asked that because of what I read here: "If, on the other hand, you are selecting a choke for a capacitor input supply (such as the typical Marshall or Fender design), then the requirements are relaxed quite a bit." from
What did he mean by that then? The requirements for the amount of inductance?

He meant, as I read it: "If you are not insisting on making the choke the first filtering element after the rectifiers, but instead are making a capacitor the first filtering element after the rectifiers and the choke only after the first filter cap, then the total amount of current flowing through the inductor is much smaller, as is the total power filtered by the choke, so the choke can be smaller than with a choke-input filter, and the design requirements as far as wiring resistance, wire size, core size, inductance value, and minimum load can be much more relaxed. A choke used after a first filter cap only filters the ripple left on the voltage of the first filter cap, not the entire power/current of the whole amplifier."

A choke-input filter is fundamentally different than a capacitor-input filter with a choke after it, and doubly so if the majority of the current comes out of the first filter cap and only a small amount goes to the rest of the circuit through the choke, like it does on almost all tube amps.

[Austin]

He meant, as I read it: "If you are not insisting on making the choke the first filtering element after the rectifiers, but instead are making a capacitor the first filtering element after the rectifiers and the choke only after the first filter cap, then the total amount of current flowing through the inductor is much smaller, as is the total power filtered by the choke, so the choke can be smaller than with a choke-input filter, and the design requirements as far as wiring resistance, wire size, core size, inductance value, and minimum load can be much more relaxed. A choke used after a first filter cap only filters the ripple left on the voltage of the first filter cap, not the entire power/current of the whole amplifier."

A choke-input filter is fundamentally different than a capacitor-input filter with a choke after it, and doubly so if the majority of the current comes out of the first filter cap and only a small amount goes to the rest of the circuit through the choke, like it does on almost all tube amps.

So I should put a cap first and supply the power tubes from that point then and then use the choke I have as a filter for the rest of the amp, bias and preamp tubes, to be on the safe side since I don't know the current rating of this choke? Or rather try the choke first and see if it gets too warm in operation? I want to use 4 6f6g (since I happen to have about 10 of them) tubes in fixed bias coldish class ab for max headroom btw. I think it would be nice to have a stereo P.A. I could also use for my home stereo.

[trobbins]

Maybe a simpler approach to designing is to start with the original product and understand its circuit operation (parts values and operating levels). Then a better estimate can be made as to changing or moving parts - especially if you don't have specs on particular parts.

That said - chokes can be crudely tested for inductance at varying DC levels, and an estimate made (along with DCR and choke wire size) of a likely suitable max operating DC current. But a test jig and effort is needed.

[Austin]

Maybe a simpler approach to designing is to start with the original product and understand its circuit operation (parts values and operating levels). Then a better estimate can be made as to changing or moving parts - especially if you don't have specs on particular parts.

That said - chokes can be crudely tested for inductance at varying DC levels, and an estimate made (along with DCR and choke wire size) of a likely suitable max operating DC current. But a test jig and effort is needed.

I don't have a schematic or even a model number for the two chassis, they are Baldwin organ amps is all I know. I have no idea what tubes they came equipped with even. Of the two chokes, one is 200 ohms dcr and is almost as big as the output transformer. The smaller one has three wires with dcr 2.86k from end to end and a wire in the middle, 870 ohms from one side and 1.98k as measured from the other side. Doesn't that sound like a 20% tap for the screen of a pentode plate load for a driver tube?

[Skywave]

One point that I do see mentioned above is with a pi input filter. At mains switch-on there will be surge current through the rectifier as the reservoir capacitor (this is the first capacitor in a pi filter) charges. This is particularly relevant if a double-diode rectifier valve is in use, because every such valve will have a maximum value of capacitance for that capacitor, so that that surge current is within the rectifier's capacity. If that capacitance value is substantially exceeded, the valve could be damaged. For example, common values of that capacitor are in the range 16 to 32 uF. However, with a 5R4GY, the max. value is 4 uF.
OTOH, with a choke input filter, that problem will not arise - because of the action of that choke.
Al. / Oct. 4, '21