On 10/27/2013 2:01 PM, Jim Lux wrote:
On 10/27/13 10:25 AM, Roger (K8RI) on TT wrote:
On 10/27/2013 10:10 AM, Jim Lux wrote:
On 10/26/13 8:15 PM, Hans Hammarquist wrote:
All these "modern", solid state, PA have the same problem, they are
"protected" and the "protection" rolls back the power as soon as they
detect reflected power. Little depending on make and design they roll
back more or less. That's why the manufacturer offer built in tuners.
The "old days" with a pi-filter on the output could be tuned to most
anything below SWR of 1:3 or even more, and they didn't have (needed
maybe) the "protection". As long as you didn't kill the final tubes
by overheating them, you were OK. Do we like to have the "old" tube
final back? Maybe.
I would rather have "smart antennas" with the finals *at the antenna*
and the matching done there.
This gets back to what I want to do...sorta.
Particularly on 160 you don't have a lot of room to make frequency
excursions and that is to put the tuner "at the antenna", but that comes
with a location that is hazardous to the tuner's health. Another is just
how good are the remote autotuners? Will they take the SWR right down
to 1:1 which is important for SS amps, not because of power, but because
of deteriorating signal quality.
Whether it will take the SWR to 1:1 is a function of the step size in
the tuner design (if it's a switched L and C) and the control algorithm.
Understand.
Most autotuners stop when the SWR is below, say, 1.3:1 or 1.2:1.
Which is not good enough for me. I tune the amp into a non reactive
dummy load with an SWR of 1:1.00XX and expect it to see the same when
switched to the antenna.
What you might want is a way to manually configure the L and C, and
then store that for the frequency (most tuners can do this now.. the
AT200PC can, for sure), or have a computer that knows what frequency
you're tuned to, and then set it up.
The problem there is lack of ambition and programming with only one
hand. Before my medical inconvenience I could type over 70 WPM while one
handed typing is more than a little annoying.
I have a BS in CS with work toward a masters before a good job came
along and ended my quest for a Masters, so I enjoy working on these
types of problems. It's just the one handed typing really slows me
down. Write the Algorithm and then write a A Nassi-Shneiderman
diagram/flow chart. Makes multilevel programs much simpler.
My solid state rigs all auto tune to very close to a 1:1, so I expect
the same from an external tuner.
(this is what I was doing with my active phased array.. I used AT200s
as essentially computer controlled LC networks)
With a remote tuner, I want to match the antenna impedance, not just
move the resonant point. Yes, if I move the resonant point to cover the
entire band it does make life easier and I could take care of the rest
in the shack, but again I'd prefer to do this at the antenna so in most
cases I only need a small L network even for 160 if it's close to
resonance.
With the half sloper other than the difficult maintenance problem this
becomes rather easy although on 160 that makes for a lot of resonant
points.
putting the matching network at the antenna for a center fed, half wave,
sloping dipole is not practical although a single band tuner at the
tower using open wire line might.
I've looked at this strategy for portable operation with a 40 foot
extendable carbon mast. Basically, put the tuner at the base, run two
wires up the mast to the dipole. Sure, that feedline has some weird
impedance and varies, but because the Z is "high", the currents are
fairly low, and it's short, in any case. Modeling shows that it should
work pretty well (e.g. no significant loss compared to putting the
tuner at the feedpoint). The tuner takes care of whatever weird
impedance is presented.
Ice storms are common here spring and
fall, although there are far fewer in the fall but the make open wire
problematic and to me, reliability/durability is important because I
have to impose on others to get things fixed.
The SGC tuners are pretty rugged devices in a pretty rugged enclosure.
Sure, it *is* moving parts (relays), but I think that if we put our
minds to it, we can design this kind of thing and have it be rock
solid reliable. No, some cheesy breadboard in a rubbermaid box isn't
going to hack it.
And ruggedness adds weight, but I agree, this is no place for the bread
board approach, except during the design phase.
Sure, it's more complex than the historic Transmitter in
Shack/Feedline/Fixed Antenna, but life moves on.
For instance, I sketched out an interesting design for a form of Yagi
with all driven elements, using an array of magnetic loops, rather
than the traditional horizontal elements. The matching from low Z
semiconductors to the low Z of the magnetic loop is actually kind of
what you want. And you're doing spatial combining, so with 5
elements, each driven with a 200W module, you don't have the losses in
the power combiner you see in a "single output" SSPA.
Combine this with things like polar modulation, and you can get some
very interesting designs. It's almost like having the entire rig at
the top of the tower, and all you need is power and an ethernet link,
which could be wireless.
Sure, its nothing like ham radio in the past, but that's what ham
radio is all about: try new things.
I like the idea and to me it's far less different than a remote regular
station, controlled over the internet. You're just combining the rig
with the antenna. There might be issues with lightening, maintenance,
and cost though. Could it be made to match the big mono band Yagi for
performance?
Ah.. there's the rub. The single biggest factor in performance is
height above the ground. The second biggest factor is number of
elements (which sets the F/B ratio) and length of boom (which sets
gain). the element length isn't a big driver of gain.
An infinitesimally small dipole has a directivity of 1.6dBi, and a
full size dipole has a directivity of 2.15 dBi. So that 0.5 dB is
something you'd have to make up some other way (more elements, longer
boom); but fundamentally, there's no inherent advantage in full length
elements. In a Yagi-Uda, you also want full size elements so that you
can couple power among the elements easily, with the right amplitude
and phases, without having tight mechanical tolerances.
The other factor is efficiency. Physically small radiators often have
high losses, and the way the amateur regulations are written, that
presents a problem. If there was a "radiated power" limit instead of a
"transmitted power" limit, then it would be easier.
At least we no longer have amplifier efficiency paramount, but can move
toward class A for the best signal, although if you want an amp with a
clean signal you almost have to build it and there are some good tubes
out there.
If you're not concerned about wall plug efficiency, then you can
*radiate* just as much power from a small loop as a big dipole. All
either does is match free space to your feedline.
The challenge is in getting what's called superdirective gain (e.g.
gain more than N where N is the number of elements (5 dB for 3
elements + 2.15 dB for the dipole = 7.15 dBi).. To get superdirective
gain, you generally need to have mutual coupling among the elements,
which is how Yagi-Uda's "work". An all driven array can do this, but
you have the prospect of elements with negative power (something
familiar to folks building directional arrays in the broadcast business).
Most amplifier designs for RF don't have a way to feed power back to
the DC bus, for instance; although this is now standard practice in
motor drive systems (so called 4 quadrant drives). So your negative
power element would need to dump power into a load.
But hey, with these new fancy LDMOS devices with huge SOAs and very
fast transition times, we can start to look at RFPAs as more like
power switching systems, rather than narrow band tuned amplifiers.
73,
Roger (K8RI)
It's a radical design, but used much the same way as ham rigs have since
day one.
_______________________________________________
_______________________________________________
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
|