Can anyone tell me what this filter is doing?

Help Support GroupDIY:

Lee_M

Well-known member
Joined
Nov 23, 2013
Messages
232
Location
Dorset, UK.
Hey guys,

I just picked up an old BBC attenuator module and there was a strange LC filter board fitted inside the module, but not actually electronically connected to anything. The board is labelled with the same BBC code number as the module it was fitted into (AT2/7A), so it seems fair to assume that it's an original part and was designed to be connected to the 600Ω balanced attenuators that make up the rest of the module, but I haven't managed to find any information about it online.

I've drawn up a schematic (see attached image), L2 and L5 are currently unidentified values (the other inductors have hand-written labels on them) so I'd need to desolder them to take measurements. They look identical to the other inductors, so I'd expect them to be in the same ballpark as 1.14mH.

Can anyone tell me what exactly it's doing? It seems to be some sort of balanced notch filter, but C7 and C8 look like they're doing something else.

Cheers!
 

Attachments

  • Mystery filter.png
    Mystery filter.png
    9.9 KB · Views: 62

shabtek

Well-known member
Joined
Jun 8, 2005
Messages
1,999
Location
midwest usa
band pass filter

(could not find a calculator with parallel LC tee network in quick search)

paralleled LC elements are band pass and shunt C is low pass
 

Lee_M

Well-known member
Joined
Nov 23, 2013
Messages
232
Location
Dorset, UK.
Isn't band pass filtering done with series LC components?

The only time I've seen parallel LC used in BPF circuits is in the shunt part of the filter, as seen in the image attached to this comment.

The only "shunt" components in this circuit seem to be C7 and C8.
 

Attachments

  • filter-lc-bandpass-typical-circuit-03.svg
    10.7 KB · Views: 7

Bo Deadly

Well-known member
Joined
Dec 22, 2015
Messages
2,442
Location
New Jersey, USA
Actually I believe that's a balanced and damped band stop plus low pass filter but it's complicated enough that I would have to model it in LTSpice to be sure of the specifics. Caps are high impedance at low frequencies where inductors are high impedance at high frequencies. So high frequencies pass straight through the caps and low frequencies pass straight through the inductors. Middle frequencies do not pass and thus it's a band stop. And the shunt caps clip the highs. Definitely higher order but not exactly sure what degree without a sim.
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,347
Location
Marcelland
It's a Cauer filter. It produces a very deep notch at about 19kHz (>100db) but is only -20dB at 65kHz.
Maybe it's designed to filter out the 19kHz subcarrier in stero FM...?
 

jacomart

Well-known member
Joined
Sep 13, 2017
Messages
161
Location
Tuscany, Italy
There's a combo box...

abbey road d enfer said:
This calc does not cater for elliptical (Cauer) filters. These filters have parallel LC elements in the series branchs, which result in notches.
 

Attachments

  • Schermata 2021-02-07 alle 00.25.55.png
    Schermata 2021-02-07 alle 00.25.55.png
    973.7 KB · Views: 43

jacomart

Well-known member
Joined
Sep 13, 2017
Messages
161
Location
Tuscany, Italy
Clearly this type of filter has the disadvantage of having a very high group delay near the corner frequency, so that in the filtering of video signals (to avoid ghost images) and also in very selective crossovers (to eliminate phase distortion) it is necessary to correct the GD with the use of suitable all-pass filters.

https://www.researchgate.net/publication/289049366_Correction_of_crossover_phase_distortion_using_reversed_time_all-pass_IIR_filter

https://en.wikipedia.org/wiki/All-pass_filter

cheers
Jaco
 

Lee_M

Well-known member
Joined
Nov 23, 2013
Messages
232
Location
Dorset, UK.
Thanks for the help, guys. It's much appreciated.

I guess I can safely pillage it for the inductors, as it doesn't seem like the filter will be useful to me as it is.
 

Studio Mollan

Well-known member
Joined
Oct 26, 2011
Messages
297
I know BBC made a lookahead limiter, am6/17 I think, that used a bunch of inductors and capacitors to make a really short delay. They put the delay on the audio signal and NOT the sidechain. Resulting in a limiter with a few milliseconds of lookahead.
I’m not at all sure if this is the case here. I’ve never seen a schematic of one.
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,347
Location
Marcelland
Studio Mollan said:
I know BBC made a lookahead limiter, am6/17 I think, that used a bunch of inductors and capacitors to make a really short delay.
That is an All Pass Filter. Typically take 10-20 inductors and capacitors to generate the required delay. This is not the case here. Group delay is far from being constant.

I’m not at all sure if this is the case here. I’ve never seen a schematic of one.
Google BBC AM6-17. take also the opportunity to search for the EMT266, where an inductorless (active) version is used.
 

Studio Mollan

Well-known member
Joined
Oct 26, 2011
Messages
297
I just found the Schemo, here on groupdiy.
Thanks!
/
Emil
abbey road d enfer said:
That is an All Pass Filter. Typically take 10-20 inductors and capacitors to generate the required delay. This is not the case here. Group delay is far from being constant.
Google BBC AM6-17. take also the opportunity to search for the EMT266, where an inductorless (active) version is used.
 

Latest posts

Top