[hpsdr] Alternate Way To Calculate SWR

[John Westmoreland]

HPSDR,

I'd like to propose a fast method of calculating forward and reverse
(reflected) power on the HPSDR framework.

Using a fast LUT algorithm, it is possible to calculate the wattage to the
closest watt on PennyLane and then the values of forward and reverse
wattage can be transmitted to PowerSDR where the final floating point
calculation can be done for SWR.

This way, the algorithm is all integer on the PennyLane side and is fast -
and PowerSDR just needs to do the SWR calculation.

Some may say, well, this isn't really much faster perhaps than what we have
now; maybe that is true - but at the same time we can calculate with a fast
LUT to the nearest watt what the power is on the FPGA's without having to
do any floating point.  I have done some experiments with floating point
functions and we still have ~ 50% resources left on the Cyclone II on
PennyLane even after implementing all of those; I just want to mention that
we can implement a very fast LUT algorithm on the FPGA in which watts can
be transmitted for the SWR calc.

Maybe (I suppose) the system architects have already thought of this - but
just want to throw this out there.  For instance - 3V corresponds to 100W
on Alex (and hence PennyLane) via the ADC and SWR cable.
Full scale on the ADC is 12-bits or 4096 total counts.  100W then is ~ 3724
counts.  A lookup table can be correspondingly constructed to send the
actual wattage out via Nwire using this fast algorithm.

Anyway - maybe we can discuss tomorrow during the call.  I should be on the
call.

73's,
John
AJ6BC
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20120329/626996ef/attachment-0003.htm>