Cascading Cheap Ceramic IF Filters?
 

I found some old postings that did not go into the how-to of cascading cheap
ceramic filters in order to improve selectivity. I have a few 4-element and
6-element 455kHz filters with which I am experimenting in an IF
downconverter for sound card DSP, specifically for DRM decoding which
requires a nice flat 10kHz bandwidth. I'm wondering if it is not a good
idea to DC couple them directly back-to-back. Should they be AC coupled?
Should there be a terminating resistor to ground or a T-pad at the junction
to provide a more uniform load/source impedance? Or is it best to make up
for the losses and provide the proper terms by putting an active stage
between them than having the gain makeup before or after?

TIA,
Tom

See my article at: http://www.karlquist.com/FCS96.pdf
Toward the end I show some cascaded ceramic filters.
You can't necessarily just cascade them directly. Of
course it would be ideal to have buffer amplifiers in
between stages, complete with impedance transforming
stages, since the filters usually don't work with 50 ohm
terminations. However, as you can see in the article,
there are shortcuts.

Rick N6RK
www.karlquist.com
www.n6rk.com

"Tom Holden" wrote in message
. ..
I found some old postings that did not go into the how-to of cascading
cheap
ceramic filters in order to improve selectivity. I have a few 4-element
and
6-element 455kHz filters with which I am experimenting in an IF
downconverter for sound card DSP, specifically for DRM decoding which
requires a nice flat 10kHz bandwidth. I'm wondering if it is not a good
idea to DC couple them directly back-to-back. Should they be AC coupled?
Should there be a terminating resistor to ground or a T-pad at the
junction
to provide a more uniform load/source impedance? Or is it best to make up
for the losses and provide the proper terms by putting an active stage
between them than having the gain makeup before or after?

TIA,
Tom

"Rick Karlquist N6RK" wrote in message news:gdnJc.93979$Oq2.67677@attbi_s52...
See my article at: http://www.karlquist.com/FCS96.pdf
Toward the end I show some cascaded ceramic filters.
You can't necessarily just cascade them directly. Of
course it would be ideal to have buffer amplifiers in
between stages, complete with impedance transforming
stages, since the filters usually don't work with 50 ohm
terminations. However, as you can see in the article,
there are shortcuts.

Rick N6RK
www.karlquist.com
www.n6rk.com


How old are these filters - there was a design here in 1971 (The
Deltahet, using the Wadley Loop priciple) based around Murata 455KHz
filters - quite impressive specs, variable bandwidth, etc. Trouble was
the filters went hydroscopic fairly quickly, and I note that there are
very few (at least here) ceramic IF filters now available.
73 de VK3BFA Andrew

The filters in my article were purchased in 1995.
Note that they are 10.7 MHz, not 455 kHz. Ceramic
filters are alive and well and still available at Digikey and
Mouser.

Rick N6RK


"Andrew VK3BFA" wrote in message
om...
"Rick Karlquist N6RK" wrote in message
news:gdnJc.93979$Oq2.67677@attbi_s52...
See my article at: http://www.karlquist.com/FCS96.pdf
Toward the end I show some cascaded ceramic filters.
You can't necessarily just cascade them directly. Of
course it would be ideal to have buffer amplifiers in
between stages, complete with impedance transforming
stages, since the filters usually don't work with 50 ohm
terminations. However, as you can see in the article,
there are shortcuts.

Rick N6RK
www.karlquist.com
www.n6rk.com


How old are these filters - there was a design here in 1971 (The
Deltahet, using the Wadley Loop priciple) based around Murata 455KHz
filters - quite impressive specs, variable bandwidth, etc. Trouble was
the filters went hydroscopic fairly quickly, and I note that there are
very few (at least here) ceramic IF filters now available.
73 de VK3BFA Andrew

On Thu, 15 Jul 2004 15:26:54 GMT, "Rick Karlquist N6RK"
wrote:

The filters in my article were purchased in 1995.
Note that they are 10.7 MHz, not 455 kHz. Ceramic
filters are alive and well and still available at Digikey and
Mouser.

Rick N6RK

Believe he mentioned a problem that they could go "hydroscopic" in
Australia?



"Andrew VK3BFA" wrote in message
. com...
"Rick Karlquist N6RK" wrote in message
news:gdnJc.93979$Oq2.67677@attbi_s52...
See my article at: http://www.karlquist.com/FCS96.pdf
Toward the end I show some cascaded ceramic filters.
You can't necessarily just cascade them directly. Of
course it would be ideal to have buffer amplifiers in
between stages, complete with impedance transforming
stages, since the filters usually don't work with 50 ohm
terminations. However, as you can see in the article,
there are shortcuts.

Rick N6RK
www.karlquist.com
www.n6rk.com


How old are these filters - there was a design here in 1971 (The
Deltahet, using the Wadley Loop priciple) based around Murata 455KHz
filters - quite impressive specs, variable bandwidth, etc. Trouble was
the filters went hydroscopic fairly quickly, and I note that there are
very few (at least here) ceramic IF filters now available.
73 de VK3BFA Andrew



---
J. M. Noeding, LA8AK, N-4623 Kristiansand
http://home.online.no/~la8ak/c.htm

Behold, J M Noeding signalled from keyed 4-1000A filament:

On Thu, 15 Jul 2004 15:26:54 GMT, "Rick Karlquist N6RK"
wrote:

The filters in my article were purchased in 1995. Note that they are
10.7 MHz, not 455 kHz. Ceramic filters are alive and well and still
available at Digikey and Mouser.

Rick N6RK

Believe he mentioned a problem that they could go "hydroscopic" in
Australia?

Then he lives in Adelaide - same weather as Vancouver, RAIN!

--
Gregg
*It's probably useful, even if it can't be SPICE'd*
http://geek.scorpiorising.ca

Thanks, Rick. Looks like a fascinating article - I've just read the part on
the filters. As a starting point, I'll try two 6-element 455kHz filters with
1.5k desired source and load resistance and couple them with a cap scaled
from your 22pF value. At 10.7MHz, I calc its X=680 or 2.06x the desired
source/load of 330 Ohms. So 2.06 x 1.5k is 3090, leading to a coupling cap
of ~100pF for 455kHz.

I did find this in an old posting by VE2BVW in the thread 'Cheap 455kHz SSB
filters?':
"In an old article in 73 (December 1977) about the Yaesu
FRG-7, reference is made to an "by Ron Risher VK3OM in the March 1977
issue of "Amateur Radio". He describes... an alternate filter, consisting
of four cascaded SFD-455-B solid state filters linked by small coupling
capacitors".

In the same thread, from W7ZFB:
"Using the simple 5 pin 2 element SFD455 in series
makes a pretty good filter if the element coupling caps are
around 10pF. Two in series is better than some of the multi
element filters and four in series gets quite good.
More gets proportionally better."

His coupling capacitor seems awfully small for this frequency with an X of
~35k versus typical 1k5 to 2k specified source/load impedance. I would
expect the attenuation through the filter to be very high.

Regards,
Tom

"Rick Karlquist N6RK" wrote in message
news:gdnJc.93979$Oq2.67677@attbi_s52...
See my article at: http://www.karlquist.com/FCS96.pdf
Toward the end I show some cascaded ceramic filters.
You can't necessarily just cascade them directly. Of
course it would be ideal to have buffer amplifiers in
between stages, complete with impedance transforming
stages, since the filters usually don't work with 50 ohm
terminations. However, as you can see in the article,
there are shortcuts.

Rick N6RK
www.karlquist.com
www.n6rk.com

"Tom Holden" wrote in message
. ..
I found some old postings that did not go into the how-to of cascading
cheap
ceramic filters in order to improve selectivity. I have a few 4-element
and
6-element 455kHz filters with which I am experimenting in an IF
downconverter for sound card DSP, specifically for DRM decoding which
requires a nice flat 10kHz bandwidth. I'm wondering if it is not a good
idea to DC couple them directly back-to-back. Should they be AC coupled?
Should there be a terminating resistor to ground or a T-pad at the
junction
to provide a more uniform load/source impedance? Or is it best to make
up
for the losses and provide the proper terms by putting an active stage
between them than having the gain makeup before or after?

TIA,
Tom

Gregg wrote in message news:gjBJc.58285$eO.20454@edtnps89...
Behold, J M Noeding signalled from keyed 4-1000A filament:

On Thu, 15 Jul 2004 15:26:54 GMT, "Rick Karlquist N6RK"
wrote:

The filters in my article were purchased in 1995. Note that they are
10.7 MHz, not 455 kHz. Ceramic filters are alive and well and still
available at Digikey and Mouser.

Rick N6RK

Believe he mentioned a problem that they could go "hydroscopic" in
Australia?

Then he lives in Adelaide - same weather as Vancouver, RAIN!

Nope, live in Melbourne - Adelaide is very dry! - Melbourne is
miserably damp in winter - but we have had a drought for 10 years so
its not!.

Rick amended my statement, my experience was with 455Khz ones -
hopefully the 10.7Mhz ones are better and no doubt manufacturing
methods have improved in the last 30 years. Well worth a try -
although there has been so much published on using computer xtals to
do the same job is it worth the effort - depends on whats available I
guess!

73 de VK3BFA Andrew

On 15 Jul 2004 20:04:54 -0700, (Andrew
VK3BFA) wrote:



Rick amended my statement, my experience was with 455Khz ones -
hopefully the 10.7Mhz ones are better and no doubt manufacturing
methods have improved in the last 30 years. Well worth a try -
although there has been so much published on using computer xtals to
do the same job is it worth the effort - depends on whats available I
guess!

73 de VK3BFA Andrew

and you are sure you din't mean mechanical filter?

---
J. M. Noeding, LA8AK, N-4623 Kristiansand
http://home.online.no/~la8ak/c.htm

Am I just crying wolf, or would cheap ceramic filters have terrible phase
linearity vs frequency?

Don't most AM broadcast radios contain ceramic filter(s).