On 11/9/19 12:04 PM, David Gilbert wrote:
I think that's a fine idea, Jim, assuming that the license allows it. I
wouldn't need it very often, but there are times when I'd like better
confidence in a model that has closely spaced wires or wires near (or
under) ground. It would even be cool to run a simple dipole model to
compare with a real antenna to see where actual RF ground might be on
this mostly dry rocky hillside.
I've spent the last few months doing this kind of thing modeling dipoles
of various lengths from 1 meter to 1 km and at various frequencies from
10kHz to 100 MHz against the lunar surface at various distances above
and below the surface. The fact that things like impedance vary smoothly
with parameter variations is what gives confidence that the numbers are
right.
I'm not sure how difficult it would be to implement, but it might be a
good idea to require an individual subscription of sorts with a
userid/password to be able to control potential abusers. Or maybe that
could be accomplished more simply with limits on compute time, frequency
of use, and number of elements.
Yeah, I'll have to contact Joe NA3T and Mark WM7D who have the AZ_PROJ
site to see what they have to say.
I'm sure there's some standard anti DDOS/hack stuff.
73,
Dave AB7E
On 11/9/2019 11:34 AM, jimlux wrote:
I've been thinking about putting up a server that would run NEC4.2 on
user submitted decks - The idea is you would upload your input deck,
and then, sometime later, your output would be there to download.
Single run at a time, some sort of queuing system, some sort of
"maximum run time". After some period of time, your output file would
disappear (depends on how much disk space I have)
(Just for context, I've been running a bunch of jobs recently for a
variety of designs with >600 segments, the GN3 Sommerfeld-Norton
ground, and run times are about 5 seconds per frequency)
It would be free, as an experiment. Sort of like the servers that will
produce a aziumuthal map centered at a user entered location.
The idea is that you can run your NEC2 jobs, and then, after
iterating, and you want to verify against the higher quality modeling
that NEC4 provides, you could do that.
Or, if your job is simple enough and runs fast enough, you could
iterate a design that has buried wires, etc.
I do need to look into the NEC4 license agreement to make sure it's
permitted. There's no export control issue - the executable (and
source) are export controlled (EAR, not ITAR), but input and output
data is not.
I'm looking for feedback on the idea - any interest in it?
Any suggestions on implementation or features?
Jim, W6RMK
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|