Plate supply - possible filter

[Val_r]

Hi,

I had in mind a twin-T band-reject filter set to 100Hz (or 120Hz) frequency.
Has anyone tried that?
Comments?

:roll:

 

[clintrubber]

This thread might be just the ticket:

http://www.groupdiy.com/index.php?topic=29954

And buy (or google for) the mentioned book. The relevant pages are online.

Bye,

Peter

 

[VacuumVoodoo]

[quote author="Val_r"]Hi,

I had in mind a twin-T band-reject filter set to 100Hz (or 120Hz) frequency.
Has anyone tried that?
Comments?

:roll:[/quote]

Do you want to reject ripple directly on the plate DC supply or in the signal chain? Different methods are needed.

 

[bcarso]

[quote author="Val_r"]Hi,

I had in mind a twin-T band-reject filter set to 100Hz (or 120Hz) frequency.
Has anyone tried that?
Comments?

:roll:[/quote]

EDIT: assuming you are talking about filtering d.c. for the plate circuit, as Alex inquires of...

If you are trying to get rid of ripple at twice the mains frequency, you need to keep in mind that it typically has a fair number of upper frequency harmonics. So sure, you could devise a (passive) notch filter at that 100 or 120 Hz, and it will help---but you will be left with the higher harmonics. These are easier to filter at least thereafter.

And if the supply is not nicely symmetrical (especially the same effective winding resistance in the transformer for each half cycle) you will also be left with some mains fundamental

One of the nice things about the twin-T is that loading doesn't affect the depth of the notch to speak of. So you can chain them, to a point. You will want one final big cap to common though, to go to your plate circuit, in most cases.

 

[Val_r]

Sorry if I have not been clear.

The use I had in mind is for further filtering of supply to plate voltage, so it should be used in aid (and after) the first Pi filter.
In my FFT scans of the psu schematic I'm actually using, I always end up with the frequency from the diode bridge and the second harmonic (so 100 and 200Hz, or 120 and 240).

Other upper harmonics obviously are not so prominent.

Another one of my main goal is to use this arrangement to
filter the heater supply.

What I'm actually using is:
Diode bridge, Pi filter, LM350 as a regulator -> to heaters.

When current demand is high, I always get the ripple from filament supply coupled in the audio through the ground.

I solved the problem by inserting between the Pi and the regulator ic, an electronic filtering stage.
Actually is a medium NPN (such as BD243) mounted in common collector configuration,
and filtering the (relatively) smaller base current becomes much easier.

I was thinking of substitute the electronic filter with the passive notch (Twin-T).

What do you think?

Respect,

Val

:?: