Well, as of about 8pm last night (conveniently before the end of 2010) I completed my Cheap Yagi for 439MHz. It was a fair bit more effort than intended, but the end result is good enough. I only intend to use it sporadically and mainly as a platform to try and improve the success/viability of a local 70cm Net on FM, and I think it'll fit that bill nicely. It also provided another opportunity to solidify some theory through some hands on activity – always an enjoyable element of the hobby.
In the end, 4nec2 tells me the design will only yield about 7.6dBi of gain. Not fantastic for an 8el yagi, but after all the compromises I made I'm happy enough with that. If I was intending it for some serious use, I'd probably start again. But that'll do pig, that'll do.
So, let's recap on what happened with this project:
1. There was a desire to have a WA5VJB "cheap yagi" that I could use to run a 70cm net.
2. So I came across the cheap yagi which was meant to be for 435MHz. I hoped it would have a large enough bandwidth to also be usable at 439MHz – the calling freq for 70cm FM.
3. Built it per design – other than larger diameter elements (0.25" rather than 0.125") only to find it was resonant way down around 430MHz.
4. Thought that was weird, so modelled it in 4nec2 which then revealed the design was actually not for 435MHz but instead was indeed resonant down around 430MHz – confirming build was correct, design not so much.
– so then it got more involved –
5. I decided to use the 4nec2 optimiser to attempt to make this suitable for what I needed. That gave me some new measurements for the reflector and DE.
6. I made the modifications, only to find I'd only improved the resonant frequency to sit around 432MHz – again, not what I was after, but what was the cause??
7. Realised my 4nec2 model was still based on the WA5VJB design and as such I still had the element diameter at 0.125" – not the 0.25" I was re-using. Reworked the model (significantly) and got my new dimensions and a better SWR curve.
8. Modified the beam again and tested with a direct coax connection with great results.
9. Made up a bracket to connect a socket (in this case – N socket) to the DE rather than the direct coax feed – this is a prototype for my planned stacked 2m 4el Cheap Yagis.
10. Connected bracket and socket and tested with great results – 1:1 SWR reading for 439MHz, but who knows the gain (hopefully at least 7dBi of the modelled 7.6dBi).
So in the end success, but a lot of extra fiddling around to get it right.
1. For all the modifications, I was limited to only changing the element lengths, not the spacing – too hard basket after having varnished the boom etc. This is why the gain probably suffers soooo much.
2. At UHF I've now seen the greater importance element diameter has. In this case it put out my implementation resonant frequency by several MHz, and it is reduced the gain – but also drastically increased the bandwidth. All this is obvious due to the greater ratio of element diameter to wavelength.
3. The bracket in the end has no soldered connections to the DE. The center pin should be fine, however time will tell if the bracket tightly mounted to the DE will be sufficient connectivity for the shield. But, the hot glue seems to be holding the bracket in place very tightly and the bracket is a snug fit anyway, so hopefully rubbing wont occur.
4. Also on the bracket, I'm surprised that it didn't totally destroy the SWR. It did however make a noticeable difference, but hopefully the gain is still reasonable – and the pattern.
5. Oh, and I've now realised how I've only considered horizontal mounting for it. I'll need to get a short section of tube for an outrigger mounting for it so that I can use it for FM as planned.
Finally, I've also attached a little PDF. It details the final dimensions I used; the SWR readings I observed (both prior to adding the bracket/socket and then with the bracket/socket) and also the SWR predicted by 4nec2 (reasonable difference there – probably due to the coax to the SWR meter etc); and there's also the charts from 4nec2 for the gain and radiation patterns.
Oh, and you can also see the NEC model is here for your perusal. It has all the dimensions as symbols so that you can play until your hearts content – or to just attempt to get it working on the frequency of your choice. (Although, I've just realised I was very lazy on comments, but hopefully the symbol names are self explanatory.)
All in all, an interesting project and hopefully I'll be running a 70cm net with it soon – as soon as I figure out the vertical mounting for it. Previous posts have some photos of the socket bracket etc., but maybe I'll take some photos of the final solution and post them up.