so… DELOCK 89583 in the house - well actually on it yet - though granted it’s a literally “quick fix” … any comments on how legit/effective it is to mount it that way? (the metal tube obviously goes down the whole chimney, ~9m - but there’s no oven connected anymore)
Above observation is right, the shortening or ‘k factor’ makes a dipole shorter than the theory.
For low frequencies (HF) above picture can be used for correction.
On this link you find some interesting facts for VHF/UHF antenna’s. Halfwave/Diameter is usually below 100 here.
A shortening of about 7%, as you observed must be right. Exact shortening will depend very much on the thickness of the radiator.
this diagram
shows that when the wirediameter is 1.7 mm, the K factor for 100 is about 0.96 , so 4% shorter than expected.
The endeffect will make the k factor some more shorter.
Best is to measure your antenna with a reflection bridge for best VSWR and resonance.
Even those diagrams will not give you the ultimate solution.
OK, here is a comparison of 3 different RFM96 modules, with the 1/4 wave vertical element being trimmed.
Note that the 3 modules have different responses and slightly different peak signal lengths.
The variation was actually less than I had expected.
Somewhere around 7.9cm seems a good compromise, around 91.4% of the actual 1/4 wave length at 868mhz, elements were 1.5mm diameter copper wire.
Measuring in the lab can provide useful information, but my shed (lab) is way to small to be good for tuning antennas, due to presense of nearby objects, the metal roof etc.
By measuring out of doors you can eliminate the effect of nearby objects and test how the antennas will perform in realistic situations, a poor mans anechoic chamber if you like.
OATS. Open Air Test Site. The standard by which all anechoic chambers are measured.
I went back to the dipole test, its made as in the picture below;
When trimmed the performance of this was slightly better (0.6dB) than a tuned 1/4 wave with radials.
However as the tuning graph below shows, there is no one point where the antenna is an obvious best match;
Which is perhaps an expected issue with low cost modules, the point (length) of best match may vary.
Please note that in your dipole design the coax feed is part of the antenna as well because there in no balun.
You might want to read this section of a Wikipedia article about dipoles.
Apologies for not mentioning it, there was no co-ax used to connect the dipole, it was mounted direct on the module, same as for the 1/4 wave vertical, keep it simple.
It was done this way specifically to avoid potential issues with the co-ax causing confusion etc.
I was also aware that with the theoretical impedance of the dipole being 75R if you chose to address this with baluns etc, it would significantly complicate the installation.
Yet despite all this, the ‘simple’ dipole performed as well and the more difficult to build 1/4 wave vertical with radials.
Considering the low power used in LoRa applications the dipole design without balun seems to be a good and simple to build antenna.
diy swr tool
I suggest a competition to establish the worst performing ‘868Mhz antenna’
To start the ball rolling today I tested an ‘868Mhz 5dbi antenna’ I got off eBay, the RF output was -18dB down on a 1/4wave vertical with radials.
I suspect this ‘868Mhz 5dbi antenna’ was actually a WiFi antenna.
That reminds me to the “Dragino LoRa-Bee” antenna i accidentally decapped 
Size matters…
I carried out some tests on the very simple dipole for 868Mhz that I posted earlier, the results are posted here;
http://www.loratracker.uk/?p=690
A simple clamp on ferrite choke seems to eliminate most of the issues caused by using co-ax cable to connect up the antenna.
A larger cap performs better? 
Just to add to the discussion.
I’ve come across some nice vids and links regarding the use of RTL-SDR and/or AirSpy to perform antenna analysis for those of use who don’t have access to Vector analysers and VSWR meters that cover ISM bands.
Looks like a low cost way to check DIY/Chinese antennas and perform some fine tuning.
