2024-05-17 Filter Design
Introduction
We are worried about strong out of band signals affecting us significantly at competition as this was an issue at last year’s competition. See 2024-04-22 RF Electronics for context.
We are addressing this for a higher priority 2.4ghz link with 2024-05-15 Digikey Order off the shelf bandpass filters, however, these are expensive and must be lightweight to mount on our drone and automated tower.
For the 1.3GHz video system there is no size and weight requirements so a larger more effective filter seems in order. This document covers the design of such filter.
This is also a fun learning opportunity for meeee!
Table of Contents
- 1 Introduction
- 2 Research
- 2.1 Filter Design
- 2.2 Our System
- 3 First Pass
- 3.1 Parameters
- 3.2 Ideal Dimensions
- 3.3 Materials
- 3.3.1 EMS Purchase
- 3.4 Actual Dimensions
- 3.5 Simulation
- 3.6 Assembly
- 3.7 Tuning
- 4 Conclusion
Research
Filter Design
scope cavity filter https://lea.hamradio.si/~s53mv/cavity/cavity.html
nice table given for designs with measurements
Need to use only ESMS rectangular tubing https://ems-stores.uwaterloo.ca/ems-site/aluminum-sq-rect-tubes
Actually good textbook: https://link.springer.com/book/10.1007/978-3-642-32861-9
chap 7 for cvity filters
lots of formulae are given here but not exactly what i wanna do
https://ieeexplore.ieee.org/document/4140435
very specific not very useful
this one is pretty dense but section 8.13 discusses these
calculators
https://en.wikipedia.org/wiki/Comb_filter
digital filter, not what im going for lol
Transition
Our System
Analog video bandwidth
We are currently using the following devices:
Devices from 2024-04-22 RF Electronics | Filtering Required RF Devices
VTX: Flywoo 1.2GHz/1.3GHz Video transmitter VTX-1G3 9CH
9CH (1080,1120,1160,1200,1240,1280,1320,1360,1258MHz)
each channel is 40 MHz wide so that’s the minimum bandwidth I need for this filter
CH0:910MHZ
CH1:980MHZ
CH2:1010MHZ
CH3:1040MHZ
CH4:1080MHZ
CH5:1120MHZ
CH6:1160MHZ
CH7:1200MHZ
CH8:1240MHZ
CH9:1280MHZ
CHH:1320MHZ
CHC:1360MHZ
CHd:1258MHZ
CHE:1100MHZ
CHF:1140MHZ
We’re using Channel 9 on rmrc so 1280 MHz center roughly
Band Information
1.24 – 1.246
1.252 – 1.258
1.264 – 1.27
1.276 – 1.282
1.288 – 1.294
https://www.rac.ca/1240-mhz-23-cm-page/
there are three legal ATV channels where we are
1240 – 1246
1252 – 1258
1276 – 1282
Legal (Possible) Spots
RMRC Ch 8 “1240MHZ” should be 1240 – 1246 MHz so legal
RMRC Ch 9 “1280MHZ” should be 1276 – 1282 MHz so legal
RMRC doesn’t have a channel for 1252 – 1258 so we can’t use this one
Passband should cover both legal spots
what’s in between hopefully isn’t an issue.
Just trying to filter out strong out of band signals with this filter.
First Pass
Parameters
Minimum Pass Band: 1236MHz - 1286MHz
Minimum Bandwidth = (1286-1236)/2 = 25 MHz
Target Center Freq = 25 +1236 = 1261 = 1260 MHz
Center Freq Wavelength = 299 792 458 / 1 260 000 000 = 0.2379 m = 23 cm
Center Freq Quarter Wavelength = 0.05948 = 6cm = 2.4 inch
Target Bandwidth = 40 MHz
Target Pass Band: 1220Mhz - 1280 MHz
Ideal Dimensions
Using the table from https://lea.hamradio.si/~s53mv/cavity/cavity.html for the initial dimensions.
“Fingers” are the three rods installed in the tube
“Probe” is the piece connected to the SMA
Tube Size
Ideally one dimension is 1/4 wavelength, will do my best
Targeting 2.4inch for one dimension based on above calculations
Finger Spacing
From table this is pretty clear I need to go with 40mm spacing
The distance between the “Fingers” is used to set the center frequency of the filter
Tube Length
From table 180mm seems optimal with this finger spacing so 7 inch
Finger Diameter
10mm from table so 0.39 inch = 0.4 inch
8mm also seems fine in table so = 0.31 inch
Finger Length
this seems to matter a lot based on center frequency
I will be able to sand these down worst case though this is also tunable
48mm on all of them seems reasonable so 1.9inch
2inch is 50.8mm so it’s kind of a lot but could sand it down
Materials
I am rushing this because machine shop closes for three day weekend!!
Getting connectors for this via 2024-05-15 Digikey Order
ensured they have the dielectric on the output to keep the 50 ohms
went with ones with just two bolts hoping easier to secure
Box Tubing
https://lea.hamradio.si/~s53mv/cavity/cavity.html
uses 60x40mm tubing so 2.4 x 1.5 inch exterior for similar center frequency designs
https://ems-stores.uwaterloo.ca/ems-site/aluminum-sq-rect-tubes
they really dont have anything that close
Alum Rect Tube 2 X 1 X .125 Wall so 2 x 1 inch
seems like the closest, inner diameter will be 1.8 which is actually quite far off
hopefully I can make this solve working off of the table provided
Alum Rect Tube 3 x 1 x .125 Wall so 3 x 1 inch
going from the inner diameter this will be closest on the 3 inch axis but the 1inch on the other one is gonna be tough to work with, will see if it solves, kinda gotta send it
Wall thickness
the .125 wall thickness is a bit thicker than the 0.01 inch used so
https://ems-stores.uwaterloo.ca/ems-site/steel-tubing
they have this in better sizes than aluminum
steel has less conductivity (~ 5x less) so it’s really not viable
not even close
Length
7 inch, I can pick this so yea will just tell them
Finger pieces
https://ems-stores.uwaterloo.ca/ems-site/aluminum-solid-rounds
diameter
must go with 0.25 or 0.5 inch
i have no idea what’s gonna be better but 1/4 inch seems like the way to go here because it’s closer
length
1.9inch if they can do it, otherwise 2inch should be tunable with the screws and sanding down
End Cover
0.6mm sheet aluminum was used (could be thinner tho if they have?)
as thin as possible at EMS
Dimension options to cover 1x3inch tube with an extra 1/2 inch (reference used 0.4inch on each end so 0.8inch total) on bent part gives
1.5x3
this one seems preferable
1x3.5
https://ems-stores.uwaterloo.ca/ems-site/aluminum-sheets
Alum Sheet - 5052 H32 20 GA. - .032
Fasteners
4x for SMA conn bolts M3X3
3x tuning bolts M3X15
3x tuning counternut and lockwasher
3x finger bolts
8x end cover bolts M3 (M3X6 ?)
8x end cover nylock nuts M3
copper/brass rod for probe
EMS doesn't have anything good for this, will need to get inventive here, not sure where to get brass rods or thin copper tube. guide uses UT-085 semirigid shield but I don’t have any of that
We have steel rods that are for little red yard flags
I can solder to these easily
Can easily cut myself with wire cutters
width should be close enough hopefully
EMS Purchase
Website and sign says they close 4:30pm friday, but 3:45pm they were closed. This means not gonna happen for competition.
Alum Rect Tube 3 x 1 x .125 Wall
7 inches
1x
Alum Solid Rounds 1/4 - 6061
3x
1.9 inch (2inch ok)
Alum Sheet - 5052 H32 20 GA. - .032
1.5x3 inch
2x
Metric Steel Socket Head Cap Screw 3X10MM X 0.5
20x
Metric Steel Socket Head Cap Screw 3X20MM X 0.5
10x
Metric Steel Socket Head Cap Screw 3X16MM X 0.5
10x
MET NUTS M3 X 0.5
20x
MET LOCK/SPRING/WASHERS M3
20x
Actual Dimensions
measure and sharpie it out
Simulation
As the EMS is closed for the weekend I have more time to simulate this circuit to hopefully get closer to acing it on the first try even if I can’t get the perfect tube size.
Simulate multiple tube sizes and spacings of components
Software options
I need to CAD up my thing before I can sim it
Assembly
taps from EMS?
should be able to drill with stuff in the bay and measure it out myself mostly
soldering rod to SMA connector
Tuning
Borrowing VNA from university
https://lea.hamradio.si/~s53mv/cavity/cavity.html
Screwing in the tuning screws should decrease the resonant frequency of each finger by approximately 5%.
“The useful tuning range of the complete filter is approximately +/- 3% around its nominal center frequency.”
using epoxy or hotglue to secure the tuning bolts once done as I don’t have the recommended paint
sharpie on parameters
Conclusion
No VNA yet, no materials, low on time to simulate. Will save this for later. It is good fun, cost effective, and useful so seems worth pursuing.