Author Topic: What are the technological options for homebrewing a 750MHz filter?  (Read 776 times)

0 Members and 1 Guest are viewing this topic.

Offline Yansi

  • Super Contributor
  • ***
  • Posts: 2274
  • Country: 00
  • STM32, STM8, AVR, 8051
Hi!

What are the technological options for homebrewing about 750MHz band-pass filter? I need to make a not very sharp roll-off filter, with preferably low insertion loss, as a crude pre-selector.  What are the options to make one?

Making a microstrip pattern - eh... well, that would be HUGE.
Such frequency is too high to do so using regular wirewound inductors and requires almost illegally small coupling capacitors :)
Making a custom SAW is I think out of the question.

What would be the preferable solution of filter technology, that would be "PCB friendly", rather then metal machining work?

Please mind I do not need the passband to be precisely set,  I just need to suppress out of band crap a little bit to help a receiver. I really have no idea how would I attack this one.

Any suggestions are welcome!
Thx Y.
 

Offline rfbroadband

  • Supporter
  • ****
  • Posts: 178
  • Country: us
Specs?
- pass band bandwidth ?
- out of band rejection?
- input/output impedance 50Ohm ?

Did you look at Minicircuits if you can buy one that meets your specs?
 

Offline DaJMasta

  • Super Contributor
  • ***
  • Posts: 1383
  • Country: us
    • medpants.com
What about an integrated filter?

https://www.digikey.com/product-detail/en/avx-corporation/LP0805H0750ASTR/478-10028-1-ND/6174533 - 751MHz center, 10MHz bandwidth

Then you add a coupling cap as a high pass filter inline with it.



I think this would be doable with discrete elements and small value 0603 or so sized parts without needing really high rated parts.  While it's certainly somewhat fast for discrete work, it should be manageable with standard SMD parts especially if you don't need/want a high Q factor.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 9027
  • Country: us
  • DavidH
It can be done with surface mount parts as many 1.2 GHz transceivers going back decades show.  A helical resonator would be too small, but a cavity filter is a reasonable size.  Quarter wave shorted stubs make fine bandpass filters but are pretty broad; they are also a short at DC serving as static and lighting protection.

 

Offline radioactive

  • Regular Contributor
  • *
  • Posts: 121
  • Country: us
Quote
I need to make a not very sharp roll-off filter, with preferably low insertion loss, as a crude pre-selector.

You can do this with lumped components.  Just a simple t-type filter with low-value caps on either side can give a decent UHF pre-select filter.  You didn't specify exactly what you need.

If you need something tighter,  I had very good results modifying a SIR microstrip filter for the 902-928 MHz band. It is very compact and high performance for these frequencies. I did find that FR-4 is not reliable enough between manufacturing runs.  FR-408 was fine.  (not much more expensive).   This is the paper I used as a reference:  https://rto.umac.mo/wp-content/uploads/docs/ruimartins_cv/publications/conferences/74.pdf  (Hoi-Kai Pang, Ka-Meng Ho, Kam-Weng Tam and R.P. Martins).  You wouldn't have to modify it much for 750 Mhz.

I used openEMS to design the SIR filter for my purposes, but there are other packages out there...
 

Offline radioactive

  • Regular Contributor
  • *
  • Posts: 121
  • Country: us
If you are working on a one-off, you also might find W1GHZ altoid tin filters  (interdigital type) interesting.  He lists length parameters on either side of 750MHz,  so you might get in the ballpark pretty easy.  I watched an eevblog mailbag segment recently where Dave received some Altoid tins.  Some tin-can bandpass filters would make a good video in my opinion...

Here is the link:  http://www.w1ghz.org/filter/Altoids_Tin_Filters.pdf
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 12783
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Yeah, coupled resonators are the way to go.  You didn't mention bandwidth, which is critical, but I'm guessing if you got absurd values for series caps, that was why...

And yeah, you need to use transformations on the ladder topology to get reasonable results at narrow bandwidths.  A series segment can be transformed to a tee segment with two series caps and a shunt resonator -- just as you transform a series crystal into a parallel resonator with the inverse topology.

Have a play with this topology, see if you get plausible values.  Mind the matching network can be transformed from a capacitor voltage divider to a current divider, which can sometimes give more reasonable values.
http://www.jrmagnetics.com/rf/doubtune/doubccl_c.php

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline Yansi

  • Super Contributor
  • ***
  • Posts: 2274
  • Country: 00
  • STM32, STM8, AVR, 8051
Well, I did not want to be any specific, just to look for ideas.   The coupling cap values are quite absurd even when working with much lower frequencies.

More specifically, I have a broadband signal source providing all sorts of unwanted crap in between like couple hundred MHz up to about 2GHz. I am only interested in signals specifically at 739 up to 749 MHz and want to get rid of the rest of the unwanted stuff, that may clog up the rest of the receiver. 

I have found that indeed some SAW filters do exist for quite cheap that would fulfill  this passband requirement, however, these are in absolutely insane packages, good only for an iSheep device (size 1x1mm LGA package), which is unfortunately out of the question for any kind of homebrewing.

T3sl4co1l - That is a very neat tool! However I am not sure if I could even 10% be sure it will work in real world. I would probably have to make a prototype, characterize it on a target PCB material and stackup and then pray it will work on the final PCB. But I will give it a try.

No, any mechanical altoids filters forbidden. I do not want to spend the rest of the year trying to teach a tuna fish can to pretend a filter.   :)

Thank you for suggestions,
Y




 

Offline eb4fbz

  • Regular Contributor
  • *
  • Posts: 87
  • Country: es
Helical coupled resonators.
 

Offline Yansi

  • Super Contributor
  • ***
  • Posts: 2274
  • Country: 00
  • STM32, STM8, AVR, 8051
Thanks, already thought about those. However it seems this is a rather obsolete/specific technology, not much of those really available (at the standard distributor channels). But yes, definitely a nice possibility there! Found a couple of interesting helicals on eBay. A bit on the expensive side, but quite handy.
 

Offline eb4fbz

  • Regular Contributor
  • *
  • Posts: 87
  • Country: es
Re: What are the technological options for homebrewing a 750MHz filter?
« Reply #10 on: May 16, 2019, 02:40:50 am »
Helical filters is not an obsolete technology, but radio technology has changed and nowadays there is not so much demand for VHF/UHF narrowband filters.
Murata bought Toko. I don't know if they have continued the helical filter catalog. Another major helical filter manufacturer is Temwell: http://www.temwell.com/products.php

However you have asked for a homebrew filter. You can't build a LTCC filter at home, so maybe you have no other option than this "obsolete" filter technology  :-DD
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 9027
  • Country: us
  • DavidH
Re: What are the technological options for homebrewing a 750MHz filter?
« Reply #11 on: May 16, 2019, 01:14:40 pm »
I didn't suggest a helical filter because at 750 MHz, the small size would make construction difficult.  A cavity filter would be larger making it easier.
 

Offline LaserSteve

  • Frequent Contributor
  • **
  • Posts: 790
  • Country: us
Re: What are the technological options for homebrewing a 750MHz filter?
« Reply #12 on: May 16, 2019, 08:15:26 pm »
Here is a good methodology for building PCB based  filters on the bench:

http://hparchive.com/Bench_Briefs/HP-Bench-Briefs-1989-01-03.pdf

It is too big to attach, or I would have attached it.

You might also want to look at figure 4 in this paper:

http://www.w1ghz.org/filter/Recipes_for_Printed_Hairpin_Filters.pdf

Steve
« Last Edit: May 16, 2019, 08:21:57 pm by LaserSteve »
"I've Never Heard of a Nuclear Meltdown Caused by a Buffer Overflow"  filssavi
 

Online Neganur

  • Supporter
  • ****
  • Posts: 1087
  • Country: fi
Re: What are the technological options for homebrewing a 750MHz filter?
« Reply #13 on: May 17, 2019, 03:26:25 pm »
DYI cavity filter made with copper pipe. This one is 1 GHz but you can increase the length of the resonators & feed.

http://www.scottyspectrumanalyzer.us/cavity.html
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf