toggle quoted messageShow quoted text
I have been following this thread with interest. My question is how when the interface is built it is then confirmed to be as per the design. My understanding would be that a signal generator producing a signal at the right level would be fed into the receiving unit and a Spectrum analyser would then record the output at different frequencies to produce a graph . I do have an RF explorer and wondered if this would be good enough.
On 8 Oct 2019, at 15:55, Andy TALBOT <andy.g4jnt@...
It was some time ago, but I imagine I simply labelled the component values as those that came off the AADE filter design software, rounded to 1 decimal place.
I always use several capacitors in parallel for applications like that - not least as there's a higher chance of tolerances cancelling out for a more accurate average.
But at those values strays often intrude, anyway.
Occasionally on diagrams I've specified the parallel equivalents for what was actually used, depends at what stage the diagram is actually annotated with values.
As I did the original filter design for Sam G4DDK's 144MHz Anglian transverter I wondered about your design of the 144MHz bandpass filter. So I have modelled it.
I assumed Q values of around 200 for air wound coils and 500 for bog standard SM capacitors.
I get around 2dB insertion loss and indeed an excellent response.
But I don't see the need for 12.2pf and 15.7pF capacitors, you have used quite a few SMDs on your board to get to these values.
If you change to 12pf and 15pf values with suitable small changes in the values for the inductors(which are tweaked) you end up in almost the same place.
likewise dropping the 0.8pF value to 0.75pf has little impact on the response and allows the use of a single component.