A Quadruple Helix for AO-40
Created 20 August 2001
Updated 23 August 2001 - Audio added comparing 60cm dish & Quad helix
Updated 27 August 2001 - Added Telemetry Waterfall comparison between 60cm dish & Quad helix
Updated 29 August 2001 - Added excerpt from RSGB VHF/UHF Manual about stacking and why a very good match is so important
After an astonishing run of luck with a 70cm homebrew 8+8 element crossed Yagi and a 13cm 60cm dish, fate has once again proved that antenna building not only requires luck, but often an enormous dose of patience.
I refer to my attempt to build an alternative to the 60cm dish. The problem with a dish is that they're rather cumbersome to mount on the az-el rotator as they're heavy and they take up precious real estate on the cross boom. So here's an alternative.
'Your dinner's on the table love'
There's no doubt: if you have a dish, building a feed for it takes a tenth of the time that this quadruple helix takes. I estimate that I spent thirty hours building and rebuilding this antenna, compared with a couple of hours on the dish.
Cost's also an issue - the dish cost a few spares out of the junk hoard. This quadruple helix cost four N sockets, eight N plugs some new fibreglass PCB (I ran out), a few trips to the DIY store for the aluminium stock, nylon screws and plastic spacers, and over 10m of expensive solid conductor coax. Oh, and a four way splitter. Total guesstimate ~GBP120, US$160. Ouch.
But this was a mission.
So I built four G3RUH style helixes with some differences...
Used single sided fibreglass PCB instead of Aluminium for reflector and spacer (same 1.6mm thickness, but easier to cut)
For the N type, I used sheet metal punches to make a 16mm hole in the spacer and a 12.5mm hole in the reflector, which is slightly larger than necessary. I used a 3mm drill and brass M3 nuts & bolts to mount the N type.
Used 32mm between turns (Tnx K9EK)
Each turn is 146mm long, marked out on the wire with a CD-R marker pen before turning. (Tnx K9EK)
Used a 40mm diameter PVC waste pipe as the initial former for the helix.
Used 500mm, 10mm x 10mm Aluminium helix support booms. Bolted to PCB reflector with 30mm M8 bolt and locking washer.
Used 7mm hollow plastic supports with 20mm M5 Nylon screws.
Checked the spacing with a 1m piece of 32mm PVC waste pipe marked every 32mm. (Tnx K9EK)
2.8m Westflex W-103 Coax 3mm solid copper inner used for each helix.
If you use PCB rather than aluminium, use the copper foil (from the 100% shield coax outer) for 'plating through' the PCB and wrapping the 1.6mm thick N type spacer. When plating through, make the copper side as flat as possible with the foil completely flattened down and the minimum of solder!!
Cut down the solder 'shoe' of the N type socket with a junior hacksaw to make the same sort of shoe but much shorter to take the helix wire.
Change the matching section: 1/4 turn of 8mm wide copper foil, 1.6mm spacing at the start, 9.6mm at the end - this way you don't have to do the 10 degree bend, AND you can use some PCB as a spacer to measure exactly the 1.6mm when you solder the helix to the N type. Don't ya love microstrip.
At Ib OZ1MY's suggestion, use an F9FT power splitter. Accept nothing less.
Be very careful of RTV. Because of its dielectric properties it messes the matching big style - both due to the differing velocity factor and vastly different impedances. Therefore only use your antenna when it's dry ;-) Seriously, I had enormous difficulties getting a decent match - and messing with RTV when it's setting is very messy. And trying to resolder with RTV everywhere isn't fun. So I decided eventually to forgo this to begin with and see how I get on.
Feed cables _exactly_ the same length - I used 400mm of RG-213 from the same reel, carefully pruning each to the same length after putting the first N type plug on one end of each cable.
Helix separation 1.8 wavelengths, 225mm between centres. (Tnx OZ1MY)
The most expensive part - the four way F9FT splitter, eight N type plugs, four N type sockets.
Use of the copper coax shielding to make a 'through hole plate'
Getting an accurate microstrip matching transformer is one of the main secrets to success of this antenna. If you don't believe me see this excerpt from the VHF/UHF Manual 4th Edn (RSGB). I've had a some good and a lot of bad experiences with matching sections, so here's some tricks to getting that match perfect.
An important point is to notice the change to the standard 1/4 turn matching section. This change really makes obtaining the perfect match child's play. To make things easy, I used the copper foil 100% shield from the outer of the Westflex W-103 coax to replace the brass shim. Note that the copper foil is about ten times thinner than the 0.2mm thick shim it replaces, so you must be careful that it doesn't get accidentally bent and stays parallel with the reflector surface.
To mark out the matching section I used a CD-R marker (see picture) directly on the copper foil to get the outline and then, with a mixture of freehand and a ruler, expanded the marking to 8mm wide. I then set to with the kitchen scissors.
Using a CD-R marker to get the right shape onto the copper foil
To get the exact measurement of spacing for the impedance transformation correct, I used 1.6mm at the start (you measure this using a piece of 1.6mm PCB between the base of the helix and the reflector when soldering it to the N type) and let the 32mm helix pitch take it up to 9.6mm. Using a microstrip calculator that I found on the web, I discovered that for 50 ohms at 1.6mm you'll need an 8mm wide section. By luck you end up with a darned close match to 140 ohms at 9.6mm 1/4 turn later - with 32mm helix pitch, 1/4 turn is at 9.6mm, so you don't need to bend the first 1/4 turn anymore.
When soldering the helix to the N type, temporarily use a piece of PCB to get that 1.6mm spacing exactly right
If you use PCB material as I did instead of aluminium, I discovered that for a consistently very good match (which in my experience is important if you want the power splitter to work with any certainty) you need to 'plate through' the PCB to the spacer and N-type socket. I took a 30x30mm piece of the copper foil, cut a cross in the middle, and pushed the four flaps made from the cross from the back of the reflector through to the copper side. Fold the flaps as flat as possible with the copper PCB surface (so as not to upset the matching section) and use an absolute minimum of solder.
You'll also need to wrap the 1.6mm spacer in copper foil (if you made it out of PCB) to make sure that the dielectric properties of the PCB doesn't affect the matching.
Using the new 32mm spacing and matching section method, three of my helixes tuned in with absolutely no problem at 1.1:1 VSWR, and one was 1.2:1. I did absolutely no adjustment with any of them. Tuned right in. Combined VSWR at power splitter 1.1:1. This is the first time I've managed to reproduce this microstrip matching section so consistently.
Believe me... I had dozens of aborted attempts before this with RTV mess ups, attempts to better the N-type feed point and realisation that if you're gonna use PCB instead of aluminium, discovering the 'plating through' issue and the value of a conducting spacer.
Matching section piccy #1 - also shows the 'through hole plating' soldered to have the minimum surface effect on the copper.
Matching section piccy #2 - also shows copper foil covered 'spacer' to make teflon of N type exactly level with the PCB's copper surface
On AO-40 at MA 29 on S2 on 23 August 2001 at about 1918Z with a squint of 24 degrees, a (feeder loss + coax switch loss + NF) = -6dB before the downconverter, I recorded this. The first 10 seconds are the calibrated 20dBic 60cm dish. The second 10 seconds are the quadruple helix. Can you tell the difference?
They're connected via a coax switch (when I switch over that's when the signal disappears for a second). To check that the feeders and coax switch weren't telling lies, I switched the feeders on the antennas, and I got the same results.
More audio at range 50,000km squint 5 degrees, 6dB losses (feeder + coax switch + NF)...
Single 16 turn helix
Quad 16 turn helix
60cm dish (reference 20dBic)
My antenna comparison setup
Dish (left) and Quadruple Helix (right) waterfalls
Mail Howard, G6LVB