Ramp up production!
I’ll be your first customer.
Thanks for the offer but it’s a limited production run of one …
I’ve done a spectrum sweep using RTL Panorama https://www.rtl-sdr.com/rtl-sdr-tutorial-measuring-filter-characteristics-and-antenna-vswr-with-an-rtl-sdr-and-noise-source/
Noise source attached directly to dongle
Noise source through filter
My pre-order for next production run: 2 pieces 
That’s very encouraging although the spikes around 920 and 960MHz are only 8dB down. Perhaps it’s phone signals getting direct into the RTL.
Suggestion for your tuning screws:
Brass screws may or may not be a better material, but are easier to work with. If you wind the nut onto the screw and then squeeze it in a vice, you’ll have a self locking nut that you can solder to the case that will prevent the screw moving after tuning.
Are you using one of the BG7TBL noise source?
For your plots, did you have the gain setting on auto or fixed?
Edit: comparing your build with with Nigel’s pic / TomMuc’s link, the feed point should be 3mm from the bottom. It’d be interesting to adjust your feedpoints and see what that does for your spectrum sweep.
.
I think yes, because after triggers purchased it and posted about it in 2015, I also purchased same/similar model.
.
.
Yes a BG7TBL but running at 5v rather then 12v due to overheating problems.
Gain setting was Auto (I think?).
I’ll try brass nuts and bolts.
I’ll re-run the scans this evening and make a note of all the settings and compare against the coax stub filter.
1 Like
The calculators suggest 5.7mm for the tap. Happy to try 3mm
Ah - ok. I was reading the diagram above. The 3mm position is for 12mm elements after all. Might still be worth trying a different position to see what it does to the response curve.
Whilst trolling through the internet I came across combline filters which use capacitors. These have the elements all grounded at one end rather than grounding at alternate ends.
@TomMuc - thanks. I do have a directional coupler somewhere in my box of tricks. I’ll get it out and give it a try.
I found this “I tried a prototype with adjustable input taps, and found that the input and output tap positions on the lines were not too important.” from this document http://www.marwynandjohn.org.uk/GM8OTI/proj23cmBPF/proj23cmBPF.html
I can’t see how those tap can be adjustable (can anybody?)
To me, with the seams unsoldered, it looks half built.
If your prepared to post it, i have have access to a VNA.
Hi theresjam,
Thanks for the offer My brother lives in Farnborough so maybe I could arrange to drop it off with him?
I want to swap the nuts and bolts for brass ones and generally tidy it up before putting it on test again. I’ve been busy this last week so have done nothing with it.
yeah i can make that work, just give me a nudge when your ready, farnborough’s only 10 mins away.
Hi theresjam,
I think I’ve sent you a PM. The filter will be with my brother on Xmas day.
I didn’t use brass nuts & bolts as my local hardware store only had steel. Love to know what the VNA makes of it. I have enough copper sheet to make another.
This is the calculator I used Interdigital Bandpass Filter Designer
Here are the dimensions
I purposely made the bandwidth 80MHz to try to eliminate any errors in dimensions & build.
Here is the scan from RTL_Pan
2 Likes
Good work!
Finally I’ve got round to posting the results that theresjam sent to me a month or so ago. He tested the two filters I’d built on professional equipment. His opinion was that both were OK to put in the RF chain - woo hoo! The 3 element is currently being used.
This is a long post & it looks like I need to edit some text. This Markdown is difficult …
Many thanks to theresjam for the testing.
I used this https://www.changpuak.ch/electronics/interdigital_bandpass_filter_designer.phpdesigner with the following parameters.
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 3 section interdigital bandpass filter.
Center Frequency : 1090 MHz
Passband Ripple : 0 dB
System Impedance : 50 Ohm
Cutoff Frequency : 1065 MHz and 1115 MHz
Bandwidth (3dB) : 50 MHz
Fractional Bandwidth : 0.0459
Filter Q : 21.8
Estimated Qu : 2712.84
Loss, based on this Qu : 0.14 dB
Passband Delay : 12.732 ns
Quarter Wavelength : 68.76 mm or 2.707 inch
Length interior Element : 61.15 mm or 2.407 inch
Length of end Element : 61.74 mm or 2.431 inch
Ground plane space : 30 mm or 1.181 inch
Rod Diameter : 8 mm or 0.315 inch
End plate to center of Rod : 15 mm or 0.591 inch
Tap to shorted End : 6.26 mm or 0.247 inch
Impedance end Rod : 88.549 Ohm
Impedance inner Rod : 93.734 Ohm
Impedance ext. line : 50.000 Ohm
**** Dimensions, mm (inch) ****
| No | End to centre | Centre to centre | G[k] | Q/coup |
|---|---|---|---|---|
| 0 | 0.00 | (0.000) | ||
| 1 | 15.00 | 37.04 | 1.000 | 0.707 |
| 2 | 52.04 | 37.04 | 2.000 | 0.707 |
| 3 | 89.08 | 0.00 | 1.000 | 1.000 |
| 4 | 104.08 |
**** Box inside dimensions ****
Height : 68.76 mm or 2.707 inch
Length : 104.08 mm or 4.098 inch
Depth : 30.00 mm or 1.181 inch
3 Element Interdigital filter
Around 0.2/0.3 insertion loss and around 88 MHz bandwidth at 6 dB.
Ignore the -6.00 dB marker.
VNA results
My results with a $15 noise source, rtl_power and heatmap.py
Out of tune
After tuning
I also built a 2 element and the results are
Ignore the -3.01dB text
The expensive equipment that theresjam used
My kit
5 Likes
