Ceramic filters

[RolandSWL]

Hi all,
 I want to experiment with those inexpensive ceramic filters as made by Murata and others. Not wanting to re-invent the wheel, I thought I'd ask the experts on this forum for some scholarly advice. Are there any good sources of practical information concerning cobbling these filters into tube and s.s. receivers? I have a vague recollection that this was discussed here at some point in the past, but, I cannot find the thread. any help is much appreciated.

Thanks, Roland

[w4bfs]

the most (imho) important thing is to utilize proper source and termination impedances which are usually around 1500 ohms .... if you don't you will have a lot of passband ripple and degraded stopband performance .... I bought some Murata 6 pole, 7.5 khz (ch455e) for less than 1$ each and was surprized just how good they were...

I like your idea about some building blocks using these .... will need agc most likely

[KD6VXI]

Search the mods section on the Kenwoody TS440.   Bear has done some testing.

There has also been mention of it in the tech section.   I believe it's u Der the same Kenwoody search.

At one time I had a cool pdf file from Murata with the different filters and specs shown.

--Shane
KD6VXI

[WBear2GCR]

Yes, I am using one in the CF4(?) position of my 440s. If I recall mine was the +/-6kHz, aka the "12kHz" one but the larger pack with more sections and better skirts. It's a drop in.

For reasons that ENTIRELY ESCAPE ME, except for the fact that I went to bed and this reality that I woke up in, while being essentially the same as the one when I was working on the 440 filter thing, I realized about two months ago that the TRANSMIT filter is CF2!! Or I think it is. Or it sure looked like it was 2 months ago - it may have reverted back to being something else... like an FM only filter on xmit, etc... Anyhow, I have that on my very long list of things to change out for a wider filter...

As far as I could tell, last time I looked at Murata, these filters were EOL in this package. But there may still be some made in the same form factor. They made a number with slightly different characteristics but similar specs. Gotta figure that out.

What do you want to do??

                    _-_-bear

[WBear2GCR]

In general terms the idea is that if there is no existing bandpass element that ur substituting in for, then you'll need to look at the side that would drive the filter and the side that the filter feeds.

The input side needs to have enough current to be able to be loaded by a resistor that looks like the value that the specs call for. It's in the pdf for these filters. There's a suggested schematic that is the one they use to test them (usually). Murata I think also had a white paper on them. The output side need to be similarly loaded. What's already there and the function those parts serve has to be taken into account. DC usually has to be blocked.

In the event that you have too much level then it needs to be padded down (easy) or not enough level, then you'd want something to amplify. In the event not enough current available, or in trying it the preceeding circuit gets heavily loaded, then you'll want a buffer. that could be just a single transistor or JFET follower. Since most IFs here are 455kHz, opamps are possible too... not all, but many.

For tube amps the situation is similar but the signal swing is likely way too much, and will need to be knocked down with a divider, then after the filter amplified back up to whatever the requisite level was... on the output side probably a transistor with some gain will do well enough. Go cascode if you want better performance. Also, should you want to try to run the amplifying stage from the tube set's B+, a high voltage transistor might fit the bill nicely!

Depending on the receiver, you might be able to build the whole thing into something that substitutes for the next tube and plugs in. That would be the tube after the filter, or in some cases (depending on the function - if it's a mixer, then no...) the tube before. Especially if you've gone solid state with a HV transistor or two, pretty neat solution!

In concept it's a black box module with a buffer/amplifier in, then the filter, then a buffer/amplifier out. In some cases either the input side and/or the output side buffer/amps could be the existing parts of the receiver...

[RolandSWL]

Thanks for the replies.
I did read the Bear TS440 information and have scanned the Murata filter document. What I need now is to figure out how to experiment with them.
Nuggets of wisdom so far:
*block dc from ceramic filters lest they suffer from internal metal migration.
*be aware of impedance matching.
*it is possible to cascade filters to tighten the skirts. This also increases the insertion loss necessitating amplification. KIWA does this in their premium filters.
I have a receiver that I converted to a KIWA premium 3.5Khz filter years ago. It may have improved the selectivity but the audio sounds muddy and unpleasant.
I want to build a pc board with multiple switched filters either through diodes or relays. This is mainly for experimental purposes to move from radio to radio.
I'm kind of amazed at the paucity (word of the day) of information available on the internet. It seems that years ago people were building filter boards for all kinds of radios. Maybe they have become appliances not meant for hardware tinkering anymore. So it goes..........

Roland......

[WBear2GCR]

Many have gone with SDR/DSP solutions these days...

I'd use relays. You could get away with a dual wafer rotary switch, since these things are tiny. Depends on if ur going internal or external, and what radio(s).

There's probably a fair amount of info to be had in back issues of various magazines. The stuff on the mechanical filters generally applies, but these are ceramic. Same with crystal filters, the filter element can more or less be seen as a "black box".

In fact you can build a filter out of passive parts, and also via active circuits, using opamps. The ceramic filter is a simple solution though...

Keep in mind that the effect of two identical filters cascaded may not be precisely what you expect in terms of skirt slope and bandwidth and importantly ripple in the passband... but these may not be significant, it's hard to know without testing.

Also there may be merit to using a wide 1st filter, and a narrow second filter... depends on what result you are looking for, and how it ends up sounding.

I found that the single filter I use in the "wide" position on AM is just about perfect for the job, fwiw. It might be nice to be able to switch in a second one, but that is uncertain.

                        _-_-bear

[WBear2GCR]

The Kiwa filter, is that 3.5kHz the TOTAL bandwidth, or is it + & - 3.5kHz?
If it sounds "muddy" it may be because it is not centered properly in your IF passband.

If that is total bandwidth, then it is too way way too tight for AM. For SSB, that ought to be a bit wide and open sounding... but even 7kHz TOTAL bandwidth is too tight for AM, you'd have to tune off to one side to get any highs...

On the 440 and other radios with an "IF Shift" feature you can hear the effect of moving the passband WRT the IF center frequency...

[RolandSWL]

Thanks Bear.
For your amusement: http://www.kiwa.com/kiwa455.html
It looks like the 3.5khz is the TOTAL bandwidth according to the graph. If so, it does explain the muddy audio. Anybody interested in a slightly used KIWA filter?
What op-amp would be a good choice to overcome the insertion loss of cascaded ceramic filters?
MC1350P. I found this I.F. amp I.C. Looks promising.

Roland...............

[WBear2GCR]

probably fine - but a simple JFET or decent transistor with a couple of resistors may do the job too...

               

[WU2D]

Stuck on in my broad as barn door TCS years ago - nice performance and super easy to terminate. The TCS has excess gain and that helps too.