HT Smoothing Chokes

[M0FYA Andy]

Perhaps a silly question, but is there any difference in a smoothing choke intended to smooth an HT supply sourced from a 50Hz mains transformer (so with a FW rectifier the ripple is at 100Hz), and where the source is a 1600Hz transformer, so the ripple is at 3200Hz?

Or is a 5 Henry or 10 Henry choke simply that, so I could use a standard choke in a 'high-cycle' supply?

They certainly don't look any different from the outside.

Andy

[GrimJosef]

I seem to recall that of the three passives (L, C and R) it's the wound components which are furthest from 'ideal'. So we can make essentially perfect resistors and pretty decent capacitors (even up to quite large capacitances) but wound components, including inductors, are always compromised, partly by the nonlinearities and frequency dependences of the magnetic materials and partly by the inherent resistance of the winding wire and the stray capacitance between turns and layers of the coil. If I were to worry about anything in your particular application it would be this stray capacitance. At low frequencies it wouldn't pass much AC current. But at 3200Hz and its harmonics that might be much less true.

Cheers,

GJ

[G8HQP Dave]

Until you reach the first parallel resonance any stray capacitance works in your favour, as it serves to increase the effective inductance.

On the other hand, the thick laminations of a 50/100Hz choke might cause more loss at 3200Hz so the effective inductance might be a bit smaller. However, I would expect it still to work as a choke so try it and see.

[M0FYA Andy]

That's why I wonder if there is any difference in the way chokes are built.
Certainly a transformer designed to work at 1600Hz is not the same as one designed to work at 50Hz.

Andy

[kalee20]

If the source is a 1600Hz transformer, I'd expect the inductance to be a great deal less.

However, if it has the inductance, and the rated current, marked on it, and you want to use it in a 50Hz application, then if these two numbers are consistent with what you want, use with confidence! G8HQP Dave's point is a good one, going up in frequency the effective parallel losses might be a bit high - but going down in frequency you'll just have a somewhat over-engineered, unusually low-loss, choke.

[M0FYA Andy]

I'm thinking more of going the other way - using a choke intended for 50Hz in a 1600Hz application.
The original in the 1600Hz circuit was 5H, and the replacement would be 10H (because that is available).

Andy

[trobbins]

Chokes in the 10H range for valve amps are likely to have a first self resonance of only a few kHz(from the ones i have measured).

Is 1600hz the fundamental, and hence 3.2kHz is the dominant ripple that you would be aiming to suppress? Note also that it won't take much capacitance for the filter cap to be operating effectively with just its ESR at that frequency.

[M0FYA Andy]

Yes, the AC frequency is 1600Hz, so 3200Hz ripple after full-wave rectification.

Smoothed by this choke followed by a 1uF capacitor to give a 400v HT rail.

I'm not sure what you mean by 'it won't take much capacitance for the filter cap to be operating effectively with just its ESR at that frequency', can you clarify please?

Andy

[trobbins]

1uF will have a 50 ohm impedance at 3.2kHz, whereas a 50uF cap will be down at 1 ohm impedance, which is down near the ESR of the capacitor and so going higher than 50uF is not likely to give you much better attenuation of the ripple voltage.

The typical concerns would likely be if you don't have the choke already, or don't really have room for it, or don't have a problem with voltage drop (such as may occur with a CRC configuration filter).

An easy way to check on the likely ripple is typically to use PSUD2, but unfortunately it only has preset AC frequencies of 50, 60 and 400Hz.