ppppenguin wrote:I'm sure Terry will tell me where I've got this wrong.
Come now, Jeffrey - you should have more faith in yourself!
Or Tool Satan, if that's more convenient - they seem to be opening up a branch at the back of every Wickes these days ...http://www.toolstation.com/shop/Electri ... ter/p66100
A problem these days is finding purely terestrial TV or CATV splitters as they all seem to encompass satellite distribution these days with a bandwidth up to 2.4GHz The penalty for this is increase through loss - typically 4.7dB compared with 3.5dB for devices with a bandwidth up to 860MHz or 1GHz. Not a problem here but, where devices are cascaded - as at the Dulwich museum - it could play merry hell with my RF planning losses!
Jeffrey - please take note! If you see any CATV splitters, grab them! I think we've virtually exhausted all the stock that I and others manage to source earlier on and the way that network keeps growing, you're going to need them!
However, back to the plot! A resistive delta or star splitter has a loss of 6db between all ports in any direction so that the isolation between ports is a maximum of 6dB. An inductive coupler, on the other hand, should feature an isolation of well over 20dB between ports.
AS jeffrey says, filtering shouldn't be a problem, although if you are going to use a distribution amplifier, be careful not to overdrive it! I did insist on a VHF diplex filter for Dulwich but was able to scrounge a CATV forward/return path filter which has a very sharp cut off above 65MHz for the low pass section and below 85MHz for the high pass section! No good for channel B5 but ideal for B4 and lower.
However, you could try making your own! I got a filter design program here - recommended my Michael Watterson, IIRC - called AADE Filter Design. Unfortunately, the website is currently closed 'Due to serious illness' and I don't know if I still have the source file availble,
However, (before I discovered that!) I knocked up a simple low pass filter - a Chebychef filter with 3dB max passband ripple like this:
which has values of (left to right) of 106pF, 138pF & 106pF for the caps and two 130mH inductors.
The performance is impressive ...
... with over 60dB attenuation at the bottom edge of Band III increasing to 75dB at the top end ...
Quite an improvement over the crude Belling Lee Diplexers of old - there's a picture on here somewhere which either I or Chris took; the main feature is an inductor stretched across thin air from a small PCB to one of the inputs!
I don't have time this afternoon but I will look for the original software package later on or perhaps, Michael or somebody else might have it.
You can tweak the capacitor values to preferred values and see the results and if that doesn't work, changing the corner frequency a little can work wonders. It even tells you how to wind the coils! It is well worth building one up because, even if you don't have equipment available to be able to align it perfectly or measure its response, it has to be better than nothing!