Thanks for the last comments Spiney. I'd thought of the feedhorn variable, so in this new test I tried an Invacom with the oher commonly available offset feedhorn made by Gibertini. It is designed for an F/D ratio of 0.66. ChannelMaster and Prodelin do not publish the ratio of their dishes (I ought to go out and measure my dish, but it's raining hard)
I found that a ChannelMaster feed worked better on both the 1.2 and 1.5 Gibertini antennas.
The results confirm many anecdotal comments in forums over the past few years. The most commonly recommended LNBs have been MTI BlueLine and Invacom, and these tests confirm this. I only had one BlueLine to try, and I had to use it in an installation before the night-time testing, which was a pity. I had four MTI white 0.2 examples, they were all very similar, much worse than the BlueLine single 0.6 (so much for manufacturer's noise figures!).
The MTI BlueLine single was better than the twin. The MTI twin's rear port was better than the front.
What was really surprising was the poor performance of the Sharp 0.3. It was so bad I thought I had used a duff unit, so I went through the others on my box. (I've been removing these from customers' dishes for the past year, replacing them with Invacoms) See second attached image for another test on six Sharp 0.3 LNBs.
No Sharp LNB could produce a lock on 10773 H BBC transponder at 10-30pm. They varied quite a bit on vertical TP's, as you can see. They perform OK in the daytime but fail dismally to hold a signal when it weakens in the evenings (when most people want to watch TV)
It is hard to replicate a very weak signal by covering up part of a dish. I tried this and my meter went crazy. In this part of Spain, horizontal TPs are VERY weak late evening.
It's worth saying that Sharp is very popular in Spain amangst expat installers, no doubt because the trade price is so low (about £10), whereas Invacom and MTI are the most expensive. Draw your own conclusions. For me, installing one of these on a ChannelMaster is like putting a Reliant Robin gearbox in a Porsche.
I chose 5 transponders, a mix of hi and lo, V and H. I think the problem with 12304 H, FIVE, is that there is interference from adjacent transponders, not that it is particularly weak.
On my dish with the reference Invacom I get all channels all day in wintertime. In summer BBC2 can start to go about 11pm. Recently, since 10th October, sun outage day here, the 2D signal has been markedly stronger.
I intend to run this test again when I have bought some more MTI BlueLine, to get a night-time result.
In this test I have drawn a threshold line to represent a clear picture. I had a Pace 2600CI (since sold) alongside a Pace Sky+(TDS470n- same tuner as 430n) on the same LNB to try to get an idea of the level of signal each needed. You'll probably be aware already that the 2600CI has a lower threshold. On the scale my meter produces, the 2600 needs 33/4 or value of 85 to be clear, the 470n needs 41/4 or value of 86.
Calibration was set to 030 on the Wolsey meter.
For some reason I don't understand, the breakup threshold is lower on high band. Is this due to the frequency or the different data rate and error correction?
