Need advice on homebrew passive VHF bandpass filter.

[Carl Wrbka]

Hi,

I recently built an Airband receiver and now want to add a passive band pass filter to the fronted. What reactive components work best on VHF (118-136MHz) band pass filters, capacitors or inductors or both? I've seen some designs use RC combinations and then there are also the LC variants. Is there a written rule regarding this or do you go with whats available in your inventory.

On a second note, does anybody perhaps have a good working design to share? I really don't want to go out and buy one from a dealer, as they are quite pricey.

73

Carl.

[IconicPCB]

Keep in mind the first stage of a chain determines noise figure of the entire chain.
SO no RC filtering if you care about noise figure.

At the frequencies in question reactive components ought to have a reasonable Q and therefore low insertion loss.

Depending on the filtering requirement an appropriate filter topology should provide less than a dB insertion loss.

What are Your requirements ( other than 118-136MHz)?

[VK5RC]

I have used the "ELSIE" (as in "L-C") filter design program before to design a low pass filter, it worked out pretty closely to the design parameters. It comes free if you get the ARRL Handbook (Ham radio book) or used to.
Do you want a band pass, low pass or high pass filter? What are you trying to filter out?
Do you need to pass the whole band? If so a cavity type filter is out.
Can you test it after for tweaking?

[retrolefty]

1, 2, and 3 stage helical filters are just size practical at VHF. There are (was?) manufacturers that made and sold these and they were adjustable over a narrow range. I think these might give the best and most robust input filter in your application.

 

[Carl Wrbka]

Hi,

Thanks for all the replies so far. I have attached a layout that I had in mind, but need help on determining the LC values. The first part of the filter will allow everything through up to about 140MHz and the second part will block anything below 115MHz. Typically its a low pass filter cascaded with a high pass filter.

I've been told that -20dB/Decade is a good "slope rate" for this type of filter, but I'm not sure what is meant by that.

[KJDS]

Some filter design equations

http://www.rfcafe.com/references/design-data/filter-design.htm

and some more

http://www.rfcafe.com/references/electrical/filters.htm

and some more

http://www.rfcafe.com/references/electrical/filter-denorm.htm

[chicken]

Here's a handy calculator for a LC bandpass that worked well enough for my purposes.
http://www.changpuak.ch/electronics/Direct-Coupled-Resonator-Bandpass.php

[German_EE]

See http://w7zoi.net/sa-stuff.html for a VHF bandpass filter used in a spectrum analyzer, this may be of use.

[IconicPCB]

In order to design a filter You need to specify its performance.
So lets assume we are talking 50 ohm input 50 ohm output terminations.
Also lets assume we are talking a band pass filter
Let us assume the pass band response is flat within a half a dB
Let us assume out of band attenuation is at least 40 dB
Is this filter to cover a number of narrow band channels?
Is group delay an issue?

Does this sound like a reasonable start?

[LaserSteve]

Mini-Circuits sells a nice airband  BP module and they do show up on Fleabay from time to time.
Steve

[G0HZU]

Although I don't recommend you use this topology, the basic filter topology you posted up in post#4 can be designed in two stages. If you want a BPF with about 18MHz bandwidth then you first design a 'tee' LPF with a cutoff of 18MHz. This has two series inductors with a shunt inductor in the middle.

For the filter topology you have in the image below this would produce values (for a 50R filter) in the ballpark of 360nH for the series inductors and 180pF for the shunt cap in the centre of the filter.

Then to transform this LPF to an airband BPF simply resonate all three of these parts at the centre frequency of your wanted BPF. So to resonate at 126MHz you need series caps of 4.4pF and the shunt inductor across the 180pF cap will be about 8.8nH. This now gives you values for all 6 of the parts in your filter.

You won't get very good performance from this BPF in terms of rejection at 108MHz or at 146MHz. You might get about 10dB with a typical design. Also, the typical Q of an 8.8nH inductor will not be that great at 125MHz so this will degrade the insertion loss a bit. Also, the series package/layout inductance associated with the 180pF cap will cause detuning issues. This is because it only takes a couple of nH package inductance to make the 180pF cap look like it is a much bigger capacitor than 180pF at 126MHz. So you would have to experiment with (maybe) 120pF here. A 120pF cap with a few nH in the leg/layout will grow in capacitance to look like approx 180pF at 126MHz. But this fudge with the 120pF cap will narrow the bandwidth of your filter slightly because it isn't the same as a pure 180pF cap.

But this filter topology isn't really the best you can choose here because of the wide variation in component values required with this design approach. As you can see, you end up having a very small shunt inductor and a very large shunt cap. Other filter topologies will use component values that behave better at 126MHz meaning the overall design experience will be much smoother with much more predictable performance.

Also, I think many airband users will need a higher order filter than this. A typical spec might be to achieve about -30dBc at 108MHz and at 144MHz. This would reduce overload issues from FM broadcast stations on 88-108MHz and also from hams and pagers across approx 144-155MHz. But this will require a filter with a lot more parts than the one shown below.

[Wolfgang]

I have made good filters for several frequencies using ChangPuaks filter designer page:

https://www.changpuak.ch/electronics/chebyshev_bandpass.php

[rcbuck]

I used Elsie to design an 18 MHz Chebyshev band filter centered on 126.7 MHz. I allowed 0.2 dB of ripple in the passband to give steeper skirts. The two end resonator sections of your schematic have values of 542.6nH and 2.9pF. The center section yields values of 7.74nH and 203.8pF.

The simulated plot shows the filter down 19 dB at 108 MHz and 12.4 db at 144 MHz.

Unless you have VHF design experience and test equipment to confirm the filter's response, you will not be able to construct such a filter. A one inch piece of copper wire has roughly 25nH of inductance.

I built a 7th order high pass filter for the 2 meter ham band last month. I wound air core inductors with copper wire for the highest Q. The inductors had values of 64nH and 40nH. The filter was built on a 0.9 inch x 1.4 inch piece of copper clad board material. Out of curiosity I measured the end to end inductance of the copper ground on the circuit board before placing any parts on the board.  It was over 40 nH.

I had to use my spectrum analyzer to adjust the inductors and capacitors for proper operation. I tried using capacitors with fixed values that Elsie had calculated for the center capacitors. It was impossible to get the correct response from the filter. I ended up replacing the capacitors with variable SMD parts that I could adjust. I was able to use fixed values for the end capacitors as those were large values (82pF). I placed grounded double sided PCB material between the inductors for shielding purposes.

I was ultimately able to get the filter adjusted for 0.25 dB loss from 143 MHz to 149 MHz with no visible ripple. Between 149 MHz and 200 MHz, I had as much as 2 dB of ripple.

I made the final adjustments after mounting the board in a 2 inch x 3 inch aluminum case. The PCB ground was connected to the case at 5 points with screws and nuts. I used RG-174 to connect the input and output of the filter to SMA connectors on each end of the case.

VHF construction is critical. As was suggested, check out W7ZOI's site here: http://w7zoi.net/mixed-bag/mixed_bag.html That will give you an idea of the task you are facing.

[whalphen]

It's not so hard to do if you use surface mount components with very short connections between them.  I highly recommend using the Elsie design software.  It's free and very interactive.  You specify your requirements and the software will determine the component values.  It took me about two minutes to determine the values for your filter.  But you need to do that yourself so it fits the performance characteristics you want it to have.  Here is some more information I previously posted:

https://www.eevblog.com/forum/rf-microwave/breadboarding-at-rf/msg1411346/#msg1411346
https://www.eevblog.com/forum/rf-microwave/breadboarding-at-rf/msg1412425/#msg1412425

I'm glad to assist if you need more help with the design.

[David Hess]

1, 2, and 3 stage helical filters are just size practical at VHF. There are (was?) manufacturers that made and sold these and they were adjustable over a narrow range. I think these might give the best and most robust input filter in your application.

A helical resonator based filter would be good for a narrower bandwidth; 118 to 136MHz is awfully wide for one.  The older Amateur Radio Handbook had design tables for constructing them.  Below is a photograph of one I built for the 2 meter band which does not really cover the whole band despite being over-coupled.

[Lord of nothing]

I use this one: ABF128-SMA for my perspective the best I ever got.
There is another Company who makes good RF Products: https://www.ebay.com/str/gpiolabs