ceramic filter myth
 

Mike Maghakian wrote:

one myth I believed for about 25 years was that a filter rated at 6KHz
at 6db had a 6 KHz bandwith at 6db. that is not true.

Yes it is true! If one looks at the plot of the response of the filter,
and if it is 6 kHz wide at the -6db point then that's exactly what it is,
6 kHz.... at the -6db point... not 8, not 12, not 4....

the 6KHz is the
BEST you can get and you rarely get that. the usual is 8Khz.

Show me the plot you are looking at to come up with this. How do you come
up with 8 kHz?

the
ceramic filter I sell is RATED at 6KHZ at 6db

How can that be? When clearly it is 8.27 kHz at -6db, (per the plot).
Where are you getting these 'ratings' from? That is, the 6 kHz 'RATED' at
-6db, the plot clearly does not show that.

but in actuality it is
around 8KHZ, this is not unusual.

Not unusual at all, because at the -6db point it IS around 8 kHz.

I have done a lot of research this
year and am surprised at how much deception there is out there.

Where is this 'deception' and 'trickery'?

and
how pathetic are most of the ceramic filters in receivers costing up
to $1000. the Sat800 is the best low cost receiver in terms of
selectivity that I have ever seen and it cost only about $400, truely
amazing.

Nothing amazing at all about it. If a filter is rated at the -6 and -60 db
it's easy to tell what the selectivity and shape factors are.

Some confusion might exist if a different set of figures are used.

If I must use ceramic filters, putting three in series will work if you
have some I.F. gain to spare. Otherwise a scavenged Collins mechanical
filter is a nice option.

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

N8KDV wrote:

Mike Maghakian wrote:


one myth I believed for about 25 years was that a filter rated at 6KHz
at 6db had a 6 KHz bandwith at 6db. that is not true.


Yes it is true! If one looks at the plot of the response of the filter,
and if it is 6 kHz wide at the -6db point then that's exactly what it is,
6 kHz.... at the -6db point... not 8, not 12, not 4....

With some of the cheapeast ceramic filters, I think this
varies from sample to sample -- that is, some samples may be
6kHz wide at the -6db point and other samples not --
nonetheless, all samples of the same model receive the same
nominal rating. However, "performance" filters tend to be
more consistent. In any case, a radio manufacturer can
select which samples go in its radios and which ones get
rejected. For bottom of the line radios (esp. cheap
portables), manufacturers are probably not so discriminating.

--
M2

Mike Maghakian ) on 8/4/03 wrote:

one myth I believed for about 25 years was that a filter rated at 6KHz
at 6db had a 6 KHz bandwith at 6db. that is not true. the 6KHz is the
BEST you can get and you rarely get that. the usual is 8Khz. the
ceramic filter I sell is RATED at 6KHZ at 6db but in actuality it is
around 8KHZ, this is not unusual.

I suggest that you read the spec sheets. Most ceramic filters are specified as
having a minimum bandwidth at -6dB and a maximum bandwidth at -60dB. In the
case of higher quality crystal and mechanical filters, you will often find a
typical values given as well.

Typically the ratio of max. bandwidth at -60dB to min bandwidth at -6db is
2.5:1 to 3:1. For example in the case of the LF-D6 which you sell, the ratio
is very close to 2.5:1. The shape factor of a better quality ceramic filter is
typically 1.5:1. This allows a lot of tolerance in the -6dB bandwidth. Again
using the LF-D6 as an example, and assuming a 1.5:1 shape factor, the -6dB
bandwidth could be anywhere from 6.0+ to 10.0- kHz.

This is the way that it has been with ceramic filters for a long time. I will
spare you the trouble of looking up the Murata ceramic filter specs from 20+
years ago. They are no different from the ones that they published before they
stopped making metal cased filters within the last two years. This is very
nice from a manufacturing point of view - it is much easier to meet
specifications. The Murata CFJ series of filters were made to tighter
tolerances, particularly the 455K12, K13 and K14 SSB filters. It is just that
they were difficult for the to obtain in one-off quantities. It was also
possible to obtain actual bandwidths within any specification you wanted from
the Japanese manufacturers - you just had to pay, and buy enough of them.

BTW why do you think that the reviews in Passport to World Band Radio (PWBR)
have included actual bandwidth measurements vs. nominal specs. for many years?
It is precisely because of the broad specifications for low cost ceramic
filters. See, for example the review of the Lowe HF-225 in the 1990 Passport
p. 171.

I have done a lot of research this year and am surprised at how
much deception there is out there.......

This statement is simply ridiculous. The only deception is in your mind. The
devices which you obtain from Murata, NTKK or others conform to their published
specifications. They never improved their specifications because they did not
have to.

...... and how pathetic are most of the ceramic filters in receivers
costing up to $1000.

If a low cost filter satisfies your market, why put a better one in? As a
manufacturer, you could purchase filters to a tighter spec, and pay for it, or
go to crystal or mechanical filters with a large increase in cost, and a
significant increase in size.

The Sat800 is the best low cost receiver in terms of
selectivity that I have ever seen and it cost only about $400, truly
amazing.

Passport 2003 gives the measured bandwidths as 2.6, 5.8 and 7.1 kHz. Possibly
you would like to do some research and tell us what the nominal figures are -
my guess would be 2.4, 4 and 6kHz. this is not great for AM broadcast band
listening unless you tune to one sideband only.

73 John KC0GGH


PS. I suggest that you correct you web site. The LF-D6 is specified as having
a minimum bandwidth at -6dB of 6 kHz, and a maximum bandwidth at -70dB of
15kHz

I have been corresponding with a filter expert this year and here is
what he has to say:


Hi Mike,

The curves I've plotted on the LF-C6 and LF-C2A would
seem to indicate that the number in NTK filters is the bandwidth so
LF-D6 would indeed be 6 kHz filters. The 6 kHz is a minimum, however.
The LF-C6 I have plotted on my website is actually about 9 kHz wide.
The Murata CFK455I's spec'ed at 4 kHz measure about 6 for the several
I tested. So ceramic filters spec'ed at 6 kHz would probably be 8-9
kHz typical, a little too wide for DXing' although now that I've found
the spec's, perhaps useful to replace the "wide" filter in a rx with
one that's still wide, but better shape factor.

See my earlier post for links to NTK specifications.

The LF-D6 is spec'ed at 6 kHz @ -6 db and 15 kHz @ -70 db.
Interpolating, that's probably 12.6 kHz wide at -60 db, about a 2:1
shape factor. The LF-D series seems to be the the best of the
different series that NTK made.


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~





N8KDV wrote in message ...
Mike Maghakian wrote:

one myth I believed for about 25 years was that a filter rated at 6KHz
at 6db had a 6 KHz bandwith at 6db. that is not true.

Yes it is true! If one looks at the plot of the response of the filter,
and if it is 6 kHz wide at the -6db point then that's exactly what it is,
6 kHz.... at the -6db point... not 8, not 12, not 4....

the 6KHz is the
BEST you can get and you rarely get that. the usual is 8Khz.

Show me the plot you are looking at to come up with this. How do you come
up with 8 kHz?

the
ceramic filter I sell is RATED at 6KHZ at 6db

How can that be? When clearly it is 8.27 kHz at -6db, (per the plot).
Where are you getting these 'ratings' from? That is, the 6 kHz 'RATED' at
-6db, the plot clearly does not show that.

but in actuality it is
around 8KHZ, this is not unusual.

Not unusual at all, because at the -6db point it IS around 8 kHz.

I have done a lot of research this
year and am surprised at how much deception there is out there.

Where is this 'deception' and 'trickery'?

and
how pathetic are most of the ceramic filters in receivers costing up
to $1000. the Sat800 is the best low cost receiver in terms of
selectivity that I have ever seen and it cost only about $400, truely
amazing.

Nothing amazing at all about it. If a filter is rated at the -6 and -60 db
it's easy to tell what the selectivity and shape factors are.

Some confusion might exist if a different set of figures are used.

There are other methods to use these steep filters. using them in
series with the existing filter in many radios will really clean up a
sloppy slope.

to prove this I am getting a RF-4900, a radio of questionable quality
which I will soup up using among other things an LF-D6 filter in
series with the existing IF filter.

I will have pics and a story when the project is done.


I will OVERCOME the skeptics !

Mike Maghakian wrote:

I have been corresponding with a filter expert this year and here is
what he has to say:

Hi Mike,

The curves I've plotted on the LF-C6 and LF-C2A would
seem to indicate that the number in NTK filters is the bandwidth so
LF-D6 would indeed be 6 kHz filters. The 6 kHz is a minimum, however.
The LF-C6 I have plotted on my website is actually about 9 kHz wide.
The Murata CFK455I's spec'ed at 4 kHz measure about 6 for the several
I tested. So ceramic filters spec'ed at 6 kHz would probably be 8-9
kHz typical, a little too wide for DXing' although now that I've found
the spec's, perhaps useful to replace the "wide" filter in a rx with
one that's still wide, but better shape factor.

See my earlier post for links to NTK specifications.

The LF-D6 is spec'ed at 6 kHz @ -6 db and 15 kHz @ -70 db.
Interpolating, that's probably 12.6 kHz wide at -60 db, about a 2:1
shape factor. The LF-D series seems to be the the best of the
different series that NTK made.

I have an LFH-4S in my '2010'. Is that filter still available?


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 80,000 Newsgroups - 16 Different Servers! =-----

kiwa still sells them. although "rated" at 4Khz, he sells them as what
they areally are, not 4Khz, but good filters at around 6Khz



I have an LFH-4S in my '2010'. Is that filter still available?


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 80,000 Newsgroups - 16 Different Servers! =-----

if you are so smart so it yourself, it is quite easy to do,

N8KDV wrote in message ...
Mike Maghakian wrote:

kiwa still sells them. although "rated" at 4Khz, he sells them as what
they areally are, not 4Khz, but good filters at around 6Khz

And I'll ask the question once again: Just where does this 'rating' come
from? He lists them as nominally 6.0 to 6.4 kHZ, and he has a plot of that
filter he

http://www.kiwa.com/sony.html

Show me please just where the 4 kHz ever, ever comes into play?

Can you possibly do that?

Steve
Holland, MI

Drake R7, R8 and R8B

I know you know everything but not everyone else does so for everyone else:

1) go to http://www.componex.de/english/produkte/

2) go to LF-H

3) go to LFH4

4) behold, it is a 4Khz filter


N8KDV wrote in message ...
Mike Maghakian wrote:

kiwa still sells them. although "rated" at 4Khz, he sells them as what
they areally are, not 4Khz, but good filters at around 6Khz

And I'll ask the question once again: Just where does this 'rating' come
from? He lists them as nominally 6.0 to 6.4 kHZ, and he has a plot of that
filter he

http://www.kiwa.com/sony.html

Show me please just where the 4 kHz ever, ever comes into play?

Can you possibly do that?

Steve
Holland, MI

Drake R7, R8 and R8B