Feel the Power!

Well,

Since the progress I discussed last time:

http://g0mgx.blogspot.co.uk/2012/10/its-becoming-real-power-struggle.html

with this power meter project; I've had (yet another) build of the directional coupler:

This time its been constructed from double sided PCB material and uses FT-82-67 toriods.

Connecting the forward and reflected ports to my 'scope (terminated with 50R) with a RF generator (this one is made my Kenwood) connected to the transmitter port and a dummy load connected to the antenna port, here's what I found on the 'scope (I've got the setup wrong and the blue reflected port reading on the 'scope needs to be divided by 10):

I then connected the coupler to the ports on the power meter and also included a 30dB attenuator. The attenuator was constructed using this site here as a Pi attenuator:

http://www.microwaves101.com/encyclopedia/calcattenuator.cfm

The setup for the next phase looks very much like this:

 

So, once this was put together, I then constructed an excel table with frequency across the top and the AtoD down the side, this table looks like this for the forward port of the coupler:

So, once I had these values I can the look into the equation that will take me from the A-to-D reading back to power - this is basically what the power meter needs to do - here's the equation:

So solving this equation as a pair of simultaneous equations - in fact a pair or more of simultaneous equations - I see this:

and this:

so basically I get different constants depending on what lines of the table I use as the input to my simultaneous equations.

However, despite the different constants, if I then use the equation and the constants to re-generate my powers from the ADC values the curve seems to be a very good fit. Here is the original readings plus the two equation outputs plotted:

So, all in all, I'm rather confused about the whole thing! But whatever the confusion, the device certainly seems to be very frequency stable i.e. very little variation in AtoD values as frequency increases. It's all good.

The software I have for this meter contains a huge set of lookup tables for the AtoD reading to power conversion; clearly these all need to change!

Cat's not helped much:

 All good though egh? I'd love to know what you think.