dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
1180
share rss forum feed

kruser
Premium
join:2002-06-01
Eastern MO

Interpret bandpass filter splitter/combiner

Click for full size
downloadLC bandpass ···atic.pdf 324,268 bytes
Bandpass Filter Schematic
Hi All,

I have a 75 ohm splitter/combiner (can be used in either direction) that I'd like to modify the allowed bandpass range on two of the filter or tank circuits outputs.
I've attached a PDF scan of a schematic I hand drew.

The output labeled as 470 to 900 MHz works fine and has an actual pass range of 440 to 1000+ MHz which is perfect for my needs.

The other two are the problem.
One is supposed to pass 174 to 230 MHz but using an analyzer, it is passing 162 to 250 MHz. I'd like to lower the passband on this output for scanner use and have it pass from 140 to 230 MHz +/- a few MHz either direction.

The low band VHF output is supposed to pass 47 to 90 MHz. My analyzer shows it is only passing 83 to 115 MHz.
I'd like to also lower the output range of this section to better allow reception of the typical VHF low band range of 25 to 60 MHz +/- a few MHz.

I did not read the capacitor values as that would require opening up circuits for valid readings. I did count the turns on each coil though as well as measure the coil ID diameter and the wire gauge used in inches for each coil. All the coils are of the open air type with no coil forms or tuning slugs.

I have tried spreading the coil turns open some on the two circuits I'd like to alter the passband on but I could not obtain any difference in the passband so I suspect one or more of the caps are dictating the passband.

Can anyone understand this and offer any advice how to achieve what I'm after?
I don't need specific values but rather info such as I need a coil with more or less turns at at L8 for example. Or I need a higher or lower value capacitor at C11 for example.

Or - maybe a link to a page that helps one design L/C passband circuits may help.

I know not having the current capacitor values may make this harder but I would think there is a common law that says a higher capacitance here or a coil with more turns there would lower or raise the pass band frequency. That info is what I seek!

My intentions for this thing is to hookup a scanner that just monitors our states VHF low band using the low band port. The low band bandpass filter would eliminate most of the paging tower crud as that is in VHF High and would be outside the passband that the scanner is attached too.
Then I could have scanners hooked to the mid band output but those scanners would need extra notch filters to reduce the paging transmitters power levels. I already have those notch filters.
Then the UHF output would be used on digital scanners like those made by GRE that have very crappy internal filtering and are often overloaded by nearby strong out of band signals. I have a county police transmitter on 155.565 that is only a block or two away. When that transmitter is running, its signal wipes out the GRE (and radioshack clones) front ends when they are tuned to the 800 MHz band. Having a working bandpass filter would help reduce the 150 MHz signals on the 440 to 1000 MHz output.
Of course I'd be using a wideband antenna on the input that is capable of receiving the entire range. I just want to use bandpass filtering before the scanner or whatever radio I'm using so I can eliminate front end overload or desense from out of band signals.

But... the splitter/combiner I have now does not really pass the frequency ranges I need. It was setup for OTA TV station use and the only output that will work as is, is the one marked as 470 to 900 which really allows 440 to 1000+ MHz through before it starts attenuating the out of band signals. Thanks to crappy Taiwan components, that bandpass filter section actually fell within the range I need and worked out well for typical scanner use!
I wish the other two sections would have worked out that way but I was not that lucky. The low band (47 to 90 MHz) section does not even pass the range it was intended to pass.

All three bandpass sections are wide enough to cover the bands I need, they are just passing too high of frequencies so they attenuate my desired frequency ranges.

Anyway, if anyone could suggest what components to try changing that would alter the passband to a lower frequency range, I'd really appreciate it.

The three items marked JMPx on the drawing are simple soldered jumpers on the board. I kept them in my drawing as they helped me make the drawing by using those as a reference points when I'd need to take my eyes off the board for a bit for eye rest!

I'd also imagine the four caps mounted right at each input or output are standard DC blocking caps and have no effect on the bandpass of each circuit. I think each circuit also has a blocking cap between each bandpass section that should not need to be changed in order to alter the bandpass range of the mid and low range sections.
Let me know if I need to try and determine some of the cap values in the two sections I'd like to alter and I can disconnect some things so I can use my cap sniffer and determine the true value of some of the caps within a section.
I thought about posting this in the ham radio forum as a lot of those guys used to build there own bandpass filters and would likely know what needs to be changed to raise or lower the passband of a given circuit but I'd imagine those guys that still mess with this stuff also read the electronics forum here as well.

Thanks Everyone!

EDIT: Sorry for the cruddy hand drawn shematic, I never was much of an artist but hopefully it is clear enough and can be understood!


aurgathor

join:2002-12-01
Lynnwood, WA
kudos:1
Reviews:
·Frontier Communi..

First, you'd probably want to read off the values of caps -- that's a very important information. Next, L/C filter design is one of the more complicated areas of EE -- there are some *really* ugly computations plus many different filters (Butterworth, Chebyshev, Elliptic, etc.), so reading a page or two may not be enough, although you may be able to find some cookbook examples where you just need to plug in the values.

quote:
One is supposed to pass 174 to 230 MHz but using an analyzer, it is passing 162 to 250 MHz. I'd like to lower the passband on this output for scanner use and have it pass from 140 to 230 MHz +/- a few MHz either direction.
I'd increase the values of C12 and C13 by about 10% - 15%, and measure the circuit again.

quote:
The low band VHF output is supposed to pass 47 to 90 MHz. My analyzer shows it is only passing 83 to 115 MHz.
I'd like to also lower the output range of this section to better allow reception of the typical VHF low band range of 25 to 60 MHz +/- a few MHz.
That may require tinkering with the input section and changing the Ls in the output section (in addition to changing the caps) so I can't offer any help on that.
--
Wacky Races 2012!

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
kudos:1
reply to kruser

said by kruser:

supposed to pass 174 to 230 MHz but ... passing 162 to 250 MHz. ... VHF output is supposed to pass 47 to 90 MHz ... only passing 83 to 115 MHz.

Component tolerances, age and of course how you actually build the circuit will all conspire to change the real world filter performance.

You could "fine tune" using varactors to complement some of the fixed capacitors. Most probably you know about them but here is a easy and funny intro: »www.qsl.net/k/ko6bb/varactor.html


tschmidt
Premium,MVM
join:2000-11-12
Milford, NH
kudos:9
Reviews:
·G4 Communications
·Fairpoint Commun..
·Hollis Hosting
reply to kruser

As others have posted RF design at these frequencies is art as much as engineering.

Given the frequencies may want to see about a custom/semi-custom filter for your needs. These guys are often used for Over the Air TV, but I have no direct experience with them.

»www.tinlee.com/index.php

/tom



THZNDUP
Deorum Offensa Diis Curae
Premium
join:2003-09-18
Lard
kudos:2
reply to kruser

Other than the low band VHF being way out of whack, aren't most filters and such rated for the bandpass at the -3db points?

Are you using a tracking generator for your analyzer?
--
one should not increase, beyond what is necessary, the number of entities required to explain anything



leibold
Premium,MVM
join:2002-07-09
Sunnyvale, CA
kudos:10
Reviews:
·SONIC.NET
reply to kruser

At the risk of stating the obvious, here is the breakdown into functional blocks:

C15, L11, C14, L10, C17, C16, L12 are the T-section bandpass filter for the 47-90MHz output.

C13, L9, C12, L7, C11, L8 are the T-section bandpass filter for the 174-230MHz output.

The schoolbook examples for T-section bandpass filters always show symmetry in the input and output L-C circuit. This would mean in your case:
L11 = L10, C15 = C14, L9 = L7, C13 = C12 and while you didn't provide capacitor values we at least know that this isn't true for the coils.
See here for an example.
The most likely reason for the lack of symmetry in the actual circuit is probably due to the influence all those filters in close proximity have on each other. This makes calculating the appropriate capacity and inductance values much harder. Nevertheless, increasing the capacity of C15,C14 (C13,C12) which are in series with the signal or decreasing the capacity of C16, C17 (C11) which are parallel to the signal should shift the frequencies of those two bandpass filters down.

C4, C5, C6, C7, L1, L2, L3 are the highpass filter for the 470-900MHz output
(it doesn't actually limit the upper frequency) and most likely not relevant to your problems.

L4, L6, L5, C8 and C9 are a T-section band block presumably for the FM radio range ?

C1, C3 and C10 appear to be DC decoupling capacitors. Higher values may be needed if you want to use lower frequencies.
--
Got some spare cpu cycles ? Join Team Helix or Team Starfire!


kruser
Premium
join:2002-06-01
Eastern MO
reply to kruser

Wow, thanks everyone!
I knew L/C circuits were not easy hence the reason I often do not understand them and decided to draw the schematic on paper and post here!
I now have some good places to look for making changes but I do need to read the current cap values first of course.

This circuit is in an old radioshack splitter/combiner. I assume most used it for using band specific antennas combined into the single output. I'll be using it in reverse and attaching a wideband antenna on the combined in/out port.
It is a double shielded box. Cast zinc or some crap for the main box (typical of cheap radioshack splitters) and then it does have a nice and tight fitting formed tin plate secondary shield inside.
The board itself is fastened to the cast box with several screws that offer very good multi point grounding.
Then of course the typical aluminum plate held on the bottom with an epoxy.
By the time I muck this thing all up and it either works or it does not, it would have been much easier to just have bought a commercial combiner like those used on today's multiband two way commercial radios. They are not cheap though plus not many offer low band (below 50 MHz or so) so I figured I do not have much to lose by trying to "adjust" this thing!
Plus being used for a receive only application, it does not need to be perfect.

Thanks again everyone for the suggestions and advice and detailed info explaining what each cap or coil is used for. I'm glad I at least remembered to come up with numbers for the caps and coils.
I've seen much worse construction in the little FM Traps that radioshack sells. I've never found two alike and some of those traps do not even fall within the FM broadcast band! One must get the back off and spread coils and move them around to get them to block the FM band like they are sold to do. Some use hot melt glue to affix the coils in place after the so called alignment while others I've opened had actual insulated coil supports that kept everything in place should you drop the thing onto something hard. Most of them that I open use hot melt glue to keep everything in place.
I'd say the little Chinese guys just stuck the coils and caps in place and as long as they looked proper, the shot hot glue around everything and sent them out for packaging and sale!
I doubt many of them actually go through an analyzer to see if they even trap the FM Broadcast band. I've found quite a few that trap the entire civil air band very well though! Or they trap the same amount of bandwidth but below the FM broadcast band.
I'd guess there is no quality control on the FM traps judging from what I've found. I bought a dozen of them off a guy on eBay several years back. They were the original design or first batch. Those were the ones that were built much better and used some form of real supports for the coils. No PC board at all in the FM Traps that I've realigned. They just solder everything together and drop it in the box and solder the in and out points to the F connectors and call it a day! The mess of soldered wires from the coils and caps is what held everything in place but some did have supports located in the middle of the tin can while most just had hot glue injected around the parts so they could not move.
I've often wondered if the reason many of them were so far off frequency was due to the hot melt glue melting during their ship ride from China and then the free hanging coils and caps would realign themselves!
The ones that used some type of supports reverberated when you tapped them with something heavy while the ones that used hot melt glue just sounded dull when tapped.
I also think the original models were from Japan and of course today's models are from China.


kruser
Premium
join:2002-06-01
Eastern MO

2 edits
reply to THZNDUP

said by THZNDUP:

Other than the low band VHF being way out of whack, aren't most filters and such rated for the bandpass at the -3db points?

Are you using a tracking generator for your analyzer?

Sorry, I missed your question.

Nope, I'm using a very crude setup with what I have at home here.
I inject a very low level signal from a signal generator and watch the signal level on an Icom R9000 receiver tuned too the gen's frequency.
It's crude but it works.

I then lower the gen frequency a few MHz and watch for a large decrease in available signal on the R9000. When I see a noticeable loss of signal, I bypass the splitter/combiner and hook the input and output together and make sure the level on the R9000 does go back up to normal.

Hey, it's crude but it gets the job done! It is a bit slow though as you must manually change everything on two devices but what the hey, it works.

kruser
Premium
join:2002-06-01
Eastern MO

1 edit

1 recommendation

reply to leibold

said by leibold:

L4, L6, L5, C8 and C9 are a T-section band block presumably for the FM radio range ?

C1, C3 and C10 appear to be DC decoupling capacitors. Higher values may be needed if you want to use lower frequencies.

That all seems correct. I tinkered with shorting some of the coils at L4, 5 and 6 and it did raise the signal levels in the FM broadcast band.

The caps at C1, 3 and 10 all measure out at .001uF.
I may try swapping them for 0.01 and see what happens but I'm going to try raising the values at C12 and 13 first and see if I can get that section to pass lower frequencies.
C12 reads as 60 pF and C13 reads as 15 pF currently but that was in circuit readings. I wanted to read them in circuit in case I ruin them when I attempt to remove them.
I'm tempted to try some variable caps I have for some as they are in range of what I'm measuring and should allow at least a 15% increase in capacitance. They are tiny little guys that should fit nicely back in where the surface mount fixed caps are.
Plus the caps are on the side of the board you can access while it is still soldered and screwed in place inside its cast housing. All the coils are on the other side but nobody has suggested changing the coils so that simplifies things greatly!

I'm using a cheap little BK cap meter and it shows a floating capacitance reading of about 3.5 pF when the leads are not connected to anything. I find that reading low value caps in the pF range sometimes need me to subtract the meters error or about 3.5 pF from the measured reading. Higher value caps like in the uF range do read correctly on the meter.
You just need to make sure and use super short test leads of maybe two inches when working with the small values.
That creates a challenge getting the meter in close enough plus your hands affect the readings also.
Sometimes I solder a very short stub of wire at the caps which gives me a point I can attach a meter probes clip too so my hands do not alter the readings.

edit: thanks for that link about LC circuit theory. Interesting stuff and it did help me understand a T filter verse a Pi section filter!
Now I know exactly what I'm doing ha!

kruser
Premium
join:2002-06-01
Eastern MO

1 edit
reply to tschmidt

said by tschmidt:

As others have posted RF design at these frequencies is art as much as engineering.

Given the frequencies may want to see about a custom/semi-custom filter for your needs. These guys are often used for Over the Air TV, but I have no direct experience with them.

»www.tinlee.com/index.php

/tom

Thanks for the tinlee link. I see they sell 50 to 75 ohm matching transformers.
I know many that say it does not matter to mix 75 and 50 ohm stuff when using it for receiving purposes but I like to keep everything balanced. My antennas and my radios are all 50 ohm so why they heck would I throw 75 ohm equipment in the middle! The transformers they sell would allow the use of 75 ohm stuff which is often very cheap and also common. Many times you cannot find a device you need made for 50 ohms but you can find it made for 75 ohms.
I may order up some of their 50-75 matching transformers.
Heck, they even change the connecter from F on the 75 ohm side to BNC on the 50 ohm side!
I like that.


SmokChsr
Who let the magic smoke out?
Premium
join:2006-03-17
Saint Augustine, FL
reply to kruser

said by kruser:

Nope, I'm using a very crude setup with what I have at home here.
I inject a very low level signal from a signal generator and watch the signal level on an Icom R9000 receiver tuned too the gen's frequency.

FWIW, using that setup the filters may actually be working closer to the desired frequencies than you think. The 3db points where the filter is rated is likely not that obvious on the receiver.