Radio Astronomy

A 1421MHz Interdigital Filter for H Line Receivers

Parent Category: Radio Astronomy

This has been quite a long time coming. 

My Hydrogen line receiver design is described elsewhere.

The Interdigital filter is a metal item, no electronic components here..

Interdigital filters are extremely efficient and comprise of a number of tuned 'rods' that are resonant at the desired frequency. Each rod interacts with the adjacent rod, thus improving the rejection of unwanted signals. The result is a clean, flat passband with steep sides providing more and more attenuation.

Design

I can't claim any originality for the design, I used an online calculator that can be found here.

My requirements were for a 5 pole filter (the more poles the greater the rejection but also the greater the insertion loss), centred at 1420MHz. I specified 0db ripple as I wanted to make measurements within the passband and not have to worry about any ripple effects. 

After inputting the data into the on-line form, I was presented with the results:

Interdigital Bandpass Filter, based on work of Jerry Hinshaw, 
Shahrokh Monemzadeh (1985) and Dale Heatherington (1996).
www.changpuak.ch/electronics/interdigital_bandpass_filter_designer.php
Javascript Version : 09. Jan 2014
-------------------------------------------------------------------------
Design data for a 5 section interdigital bandpass filter.
Center Frequency          :    1420 MHz
Passband Ripple           :    0 dB
System Impedance          :    50 Ohm 
Cutoff Frequency          :    1415 MHz and 1425 MHz
Bandwidth (3dB)           :    10 MHz
Fractional Bandwidth      :    0.0070
Filter Q                  :    142
Estimated Qu              :    1961.04
Loss, based on this Qu    :    2.034 dB
Passband Delay            :    103.007 ns
-------------------------------------------------------------------------
Quarter Wavelength            :    52.78 mm or 2.078 inch
Length interior Element       :    47.54 mm or 1.872 inch
Length of end Element         :    48.37 mm or 1.904 inch
Ground plane space            :    19 mm or 0.748 inch
Rod Diameter                  :    6 mm or 0.236 inch
End plate to center of Rod    :    6 mm or 0.236 inch
Tap to shorted End            :    2.75 mm or 0.108 inch
Impedance end Rod             :    67.004 Ohm 
Impedance inner Rod           :    83.597 Ohm 
Impedance ext. line           :    50.000 Ohm 
-------------------------------------------------------------------------
                     **** Dimensions, mm (inch) ****
 #    End to Center      Center-Center        G[k]         Q/Coup
 0    0.00 (0.000)
 1    6.00 (0.236)      32.86 (1.294)        0.618        1.000
 2    38.86 (1.530)      37.08 (1.460)        1.618        0.556
 3    75.94 (2.990)      37.08 (1.460)        2.000        0.556
 4    113.02 (4.450)     32.86 (1.294)        1.618        1.000
 5    145.88 (5.743)     0.00 (0.000)         0.618        0.618
 6    151.88 (5.979)      

                      **** Box inside dimensions ****
Height : 52.78 mm or 2.078 inch
Length : 151.88 mm or 5.979 inch
Depth  : 19.00 mm or 0.748 inch

The program also predicts the performance of the filter:

So far, so good. All I had to do was to convert this design into metalwork and align it.

Construction

The filters are typically manufactured from brass or aluminium, with the brass parts optionally silver plated. 

Mine was built from 3/4" x1/4" brass strip for the walls, 6mm brass rod for the tuning elements and SMA connectors to interface to the outside world.

This is the type of SMA connector to use. It has an extended PTFE sheath that fits into a 4mm hole in the wall of a box, ensuring that the 50 ohm characteristic impedance is maintained.

Soft lead free solder was used to hold the rods in place, the rest was simply screwed together.

10 months later, I made a start. Much of the delay was due to other activities, but the nagging worry was the alignment. Aligning Interdigital filters (or any filter for that matter) is tricky without the correct tools.  To meet the alignment requirement, I acquired a NanoVNA SAA2 which has a 3GHz upper frequency limit and also a copy of the excellent "NanoVNAs Explained" book, authored by Mike Richards G4WNC and available from the RSGB.

The metal was sourced from a Model Engineering shop, Macc Models in Macclesfield (UK). SMA connectors came from stock.

The tools required in addition to normal hand tools were:

  • Pillar Drill with Machine Vice
  • 6mm, 4mm, 3.5mm and 2.5mm SHARP drill bits
  • Blowtorch
  • M5, M3 Taps
  • Digital Vernier Calliper

Manufacturing steps are broadly as follows:

  • Cut the strip bars to length (151.88mm) as accurately as possible. The lop and bottom bars are most critical, cut oversize and fille down to the correct size, making sure that the ends are square.
  • Cut the side bars to length adding 0.5" as these will overlap the top and bottom bars..
  • Using a metal scribe, mark out the positions of the rods in the top and bottom bars. This will mean 5 holes cut in the long sides.
  • Clamp to the two long strip bars together and drill through both using a 4mm drill. (Ideally a 4.2mm drill).
  • On one of the sides, open out the holes in the 2 outer and centre to 6 mm. Tap the remaining two holes to M5.
  • On the second side, do the opposite Tap the two outer and centre hole to M5, drill the remaining two holes to 6mm
  • Cut and fit the resonator rods. 
  • Solder the resonator rods in place. Before soldering, use the Vernier Calliper to set the exact internal length. Note that the inner 3 resonator rods are a different length to the outer pair. Use a blowtorch to heat the assembly, fitting and soldering one resonator rod at a time. You may need to use additional electronic quality flux (not plumbers flux) to ensure that the solder flows freely. Use minimal lead free solder, it is very lossy at these frequencies. File off all traces of solder that may have strayed.

  • Mark and drill the side  pieces  - 4mm for the SMA connector and 2 x 3.5mm clearance at each end for a single M3 screw used  to attach the side bars to the top and bottom bars.
  • Solder the SMA socket into the end pieces, avoiding heating the connector directly with a flame.
  • Align the end bar with the top and bottom and trim the SMA connector pin so that it just touches the first tuning element. 
  • Mark the position of the end securing screw and drill the lower bar with a M2.5 drill 12mm deep and then tap to M3.
  • Repeat for the other end.
  • Secure the end pieces with a single M3 screw
  • Position the top bar and adjust the width to 52.78mm. Mark the top screw hole and drill through to the tune screw hole with an M2.5 drill. Tap to M3
  • Repeat for the other end
  • Fit short screws to secure the top bar to the side bars.
  • Solder the SMA connector pin to the end resonator rods. Avoid getting the blowtorch flame on the SMA connector. Remove any surplus solder.
  • Mark and drill (M3.5) the side plate screw holes in the completed frame. 2 holes at each end and 4 holes in the long bars.

 

  • Using thin brass sheet (thickness not critical, I used 0.032"), cut the sides to the correct size. Clamp in place and drill through the frame through the side plates.
  • Secure the side plates with 12 x 25mm M3 cap bolts and nuts.
  • Fit the M5 Tuning Screws with locknuts.

Alignment

Without any alignment tools, this is very tricky. However, the availability of very affordable nanoVNAs have brought complex alignment tasks within reach of the average amateur constructor.

This is not intended to be a tutorial on using a nanoVNA, they are all slightly different. 

The Interdigital Filter is a 2 port symmetrical device, either end can be defined as input or output. The nanoVNA has an output port, (port 1), and an input port, (port 2).

We will measure 2 characteristics of the filter, S11 configured as VSWR and S21, configured as through.

Calibrate your nanoVNA as per the setup instructions

Set the Start and Stop frequencies to 1401MHz and 1441MHz respectively with a marker set to 1421MHz.

Adjust the Tuning Screw nearest the input port (Port1) until the VSWR suddenly drops.

Swap the cables round and repeat for the opposite end Tuning Screw.

Slowly screw the 3 inner tuning screws until the S21 trace resembles the passband illustrated above.

Tweak for a nice flat passband response from 1415MHz to 1425MHz and a flat VSWR response.

The blue trace shows VSWR. It looks horrendous but actually only varies between 1.8:1 and 2.2:1. Not ideal but should be acceptable.

200MHz Span. After further tweaking, I managed to get the insertion loss down to 5dB and the VSWR less than 2:1

2024-11-09

<End>