The models (not all but an example of each) are here (http://ghz24.freeshell.org/antenna/models/newermodels/parabolas/) now. I'll add more as I clean them up.
it looks like some material can be saved tapering the KISS reflector.
Yes if you look at the pattern you can actually see the spill over the top and bottom so taller and narrower would probably squeeze some more gain out of it
if the SWR holds out good the kiss reflector is the first one that has driven the swr of the lpda above 2. And if you cut 3 inches of each side the difference is insignificant.
One of the other interesting things I found was as I tried to optimize the trough similar to post 187. This model is optimized for width and focal point at 585 MHz (as are most the others)
troughlpda.nec (http://ghz24.freeshell.org/antenna/models/newermodels/parabolas/troughlpda.nec)
but if you edit the 2 variables with notepad to match these numbers
SY x=38.59413
SY rl=35
you'll see what I'd call a putty knife
http://www.digitalhome.ca/forum/picture.php?albumid=937&pictureid=4967

it has a 36 degree sweet spot where gain stays even within 0.15 dB
this pattern or close to it persists for almost 100 MHz and can be adjusted
with the reflector length ( rl ) variable.
I can think of uses for a pattern like that. the 5 feeds I tested are there now to.

Here is a zip file (http://ghz24.freeshell.org/antenna/models/newermodels/downloadparabolas.html) with the models used to compare plus a few weird test models thrown in. Most are combinations of the feeds and reflectors already posted but they are already pasted together and arranged more logicaly. I'll post some of the .out files here (http://ghz24.freeshell.org/antenna/models/newermodels/downloadparabolaout.html) IF it's not there yet try in a bit.

This model is optimized for width and focal point at 585 MHz (as are most the others)
troughlpda.nec
I think youre on to something there. Try optimizing it for like channel 15 as there is still decent gain up to channel 68 as is now. A lot are looking for a high gain low channel antenna, but with still decent gain (12+ dbi) at channel 51. With the lpda feed, I think its doable.

Right now, it has overall gain comparable to the smaller orig DBGH mesh model (which is 30" X 65"), but with better channel 14 - 15 performance. With more gain on the lower channels and about the same reflector size, it would be a worthwhile 48" X 80" build over say a larger DBGH10 mesh model (which is about 48" X 109").

Here is a zip file with the models used to compare plus a few weird test models thrown in.
Out of those, which in your opinion is the best KISS model ?

I think the best kiss model is probably the one posted here (http://ghz24.freeshell.org/antenna/models/newermodels/parabolas/kiss4p.nec) they are really the same model with the difference being the sy variables, in the zip file its name is tpBuild2128lpdachev4planevaroptTFx.nec
or just make the variables
SY x=34.638
SY rl=39
SY f=7.462125
SY t=10.56548

BTW the extra (IN)significant figures ( 10.56548 vs 10.56) denotes the optimizer at work.
A lot are looking for a high gain low channel antenna, but with still decent gain (12+ dbi) at channel 51 I've seen several posts regarding that.

For the record you're not talking about the trough /post 187 style like this one (optimized for 473)

tpBuild2128lpdadualoptlow.nec
I'll let the optimizer at the kiss model at 473 while I gone today and see what it gets.
Could you point me at the two models (GH) you referred to
I have a lot of gh models but I haven't studied them really yet.
They are hard to compete with and I figured to study them last.
BTW vhf performance unimportant for present purposes. So try for ones that get best uhf.
Or just pick the ones you like best.

I think the best kiss model is probably the one posted here
Pretty impressive, 19.33 dBi and 1.56 SWR at 585 mhz in a 78" X 78" package. Lowest gain of 17.33 dBi 1.67 SWR at 470 mhz. Gain at 698 mhz 18.62 dBi 1.82 SWR.

Should outperform the 7 ft diameter CM-4251 over the uhf range. Post #6 has the CM-4251 gain chart. From what I can determine, their dBi is dBd + about 1 rather than 2.15.

Using 2" X 4" mesh, a few more 10ths of a db can be expected. As a practical matter, 2X4 mesh is sold in 36 or 48 inch heights in that range. So for no waste, 72" X 78" or 96" X 78" would be chosen.

As with most 19 and higher dBi antennas, beamwidth is very narrow about 10 degrees.

I've been mulling around a really easy build technique for this reflector.
I can't figure any way to model the exact antenna that would result from this build technique ( but it would be some where between the kiss and the fully developed parabola models for the same overall dimensions).


As a practical matter, 2X4 mesh is sold in 36 or 48 inch heights in that range. So for no waste, 72" X 78" or 96" X 78" would be chosen.


I recently helped a friend put up a fence it was a 4 foot tall welded wire version of what we used to call "2 by 4 non-climb" (only it was woven not welded)

As we would "hang " the pieces of fence from the end the widths would be either 6 or 8 foot as you pointed out but the height is expandable (just cut a longer piece off the roll) so the 96" wide model could be square at 96 " tall or even taller.

The dimensions I'll quote are for a 6'x6' model but here's the idea : hang or lay two 3 foot wide 6 foot long pieces of the fence so they make a square
(the openings in the fence should be 4" wide as you look at the square)
Wire the two pieces of fence together at the origin (dead center) then at top center and bottom center overlap the pieces of fence ~1 3/4 inch
so the overlap goes from 0 at the origin to 1.75 at the very top and bottom

Maybe this model (http://ghz24.freeshell.org/antenna/models/newermodels/parabolas/kissmeshlpda.nec) will help demonstrate what I'm getting at . This "tortured panel" like construction would make a smoother approximate of a parabola than the kiss with easier construction.

I've been mulling around a really easy build technique for this reflector.

Maybe this model will help demonstrate what I'm getting at .

Yeah, I get it. In cases like this, I would make a small scale mock up using cardboard and balsa wood. I just cut out a flimsy one out of graph paper and see the "diamond pyramid" problem in the middle.

The frame for the KISS antenna is going to need either a pyramid block at the apex, or the tubing bent with a heat gun and bolted/joined at the apex to form a pyramid. (I have an old aluminum umbrella type clothes line that would work perfect I think, heh)

For the 6ft by 6ft example, using 36" high 2X4 welded mesh, I would cut two 72" long pieces and bend them in the middle. They would be the top and bottom halves of the mesh.
The horizontal diamond pyramid hole left in the middle would be covered with 1/2" X 1/2" or 1" X 1" ratwire mesh. The hole looks to have a maximum height of maybe only 6 inches in your example model.

At the center of a parabola, a flat vertical diamond ratwire mesh panel of that small size may be just as effective as a diamond pyramid one ??

Umm, you drew up the 2X4 mesh using 1/4" material. Thats some strong mesh, heh. To save calculation time, you could segment mesh pieces at 1 segment per piece, that works fine Ive found.

Even quicker is to use SM and SC cards.

Sometimes it takes me a while to digest everything you say in your posts. Comparing sizes I see you were talking about optimizing the putty knife model for low uhf. I'd already done that it's in the trough models in the zip file but I'll post it separate here (http://ghz24.freeshell.org/antenna/models/newermodels/parabolas/troughlpdalowuhf.nec).
It gets 16.5 raw at 473 MHz and a low of 15.8+ at 677-689 MHz.
But it's a 5 feet wide and 6.5 feet high variant of post 187.
I just cut out a flimsy one out of graph paper and see the "diamond pyramid" problem in the middle. Yah I have several but
There shouldn't be a hole in the center.
Try this as a model. take piece of paper 8.5 x 11 cut it in half vertically from the top and bottom (where the Y axis would be on graph paper) but leave an inch or so of paper at the origin/middle. then overlap the very top of the paper ~1/2 inch . The seam should be a triangle 1/2 inch wide at the top going to no overlap at the center/origin.Repeat for the bottom.

Umm, you drew up the 2X4 mesh using 1/4" material. Thats some strong mesh Thats a reminent from the comparison models
mostly I built that model to find the 1.7 inch amount to overlap the fence(at 36") to make the parabola turn into something close to the kiss reflector.
The half inch gap is just to avoid intersecting wires errors I manually cut the wires that crossed over the z axis.If you look real close you can see the
small error this caused in the top center wires not staying straight.
The 1.7 inch estimate I gleaned is not accurate enough I'll have to start over to get more accurate estimate (or build a mockup to measure)

They would be the top and bottom halves of the mesh
wouldn't that be left and right?

Maybe we are talking about different products, here is what I'd consider optimal.
http://www.redbrand.com/Products/OtherProducts/WeldedWire.aspx


But as you would build the fence like this (sorry for awful drawing)
http://www.digitalhome.ca/forum/picture.php?albumid=937&pictureid=4987
This orientation has twice as many vertical wires as horizontal so we would want to turn it 90 degrees like this
http://www.digitalhome.ca/forum/picture.php?albumid=937&pictureid=4986
That's why I'm thinking left and right not top and bottom.

The frame for the KISS antenna is going to need either a pyramid block at the apex, or the tubing bent with a heat gun and bolted/joined at the apex to form a pyramid
I was thinking of something like this
http://www.digitalhome.ca/forum/picture.php?albumid=937&pictureid=4993
The mast could be notched to let the two (overlapped) horizontal wires rest in, add T's to support the feed and spread the weight across the top (and spread torque across more area).
This would keep the center of mass more directly over the tower (than others I've considered)
I'm fascinated by remote mechanically adapting antennas so I would consider supporting the bottom with a dish actuator (or even a pulley from the top and some weed eater line) . Because hanging all the weight from the top would give a narrower longer shorter focal length dish (better at high frequency)than one supported partly from the bottom which would be shorter, wider, flatter and longer focal length (better at low frequency).
BTW
The 100 foot roll weighs 98 lbs for the 12-1/2 gauge 48 inch tall/wide type so for an 8x8 piece to make a reflector ~7.5 feet square would weigh in at
just under 16 lbs.
Thanks for the tip on SM and SC cards I'll have to look up how to use them.

That's why I'm thinking left and right not top and bottom.

Oops, your right. I had the graph paper model turned sideways, heh.

Yeah, meeting the two panel centers at the apex and overlapping the tops and the bottoms would be easy from a practical standpoint, and even make it stronger. The horizontal wires would be creating a shallow Vee, but that shouldnt change the gain more than a couple of 10th of a dBi.

I was thinking of something like this
Not bad, and I think it would work pretty well with 12 gauge 2X4 mesh. I think I would put in another T support near the bottom like you have at the top to make it even more twist resistant in the wind.

Thanks for the tip on SM and SC cards I'll have to look up how to use them.
I have a corner reflector example here post 861 :
http://www.digitalhome.ca/forum/showthread.php?t=83772&page=58
Ill have to see if SM and SC cards allow overlapping since theyre not really wires.

Found it :):
o If the sides between corners 1 and 2 and between corners 2 and 3 are
not perpendicular, the complete surface and the individual patches
will be parallelograms.
http://www.nec2.org/part_3/cards/sm.html

OK, so the panels of the SM/SC card model of the 6'X6' KISS antenna wont be 3'X3' squares but slight parallelograms. The corners of the KISS antenna SM/SC model will be slightly longer than in real life. But since the radiation pattern of the lpda feed doesnt really extend into the corners, the SM/SC card model should be very close. :)
Hmm, it doesnt look like parallelograms to me. Im not quite sure what its doing, but it looks great to me :

CM KISS Reflector using SM and SC cards by 300ohm
CM 6ft by 6 ft KISS parabola Reflector
CM SM panels
CM Density = 2 inch by 4 inch mesh
CM Size = 4 panels of 36 inch wide by 36 inch high
CE
SM 18 9 9.25 0 34.80 0 0 0
SC 0 0 6.75 34.80 0
SM 18 9 0 0 0 9.25 0 -34.80
SC 0 0 15.80 34.80 -34.80
SM 18 9 15.80 -34.80 34.80 6.75 -34.80 0
SC 0 0 0 0 0
SM 18 9 6.75 -34.80 0 15.80 -34.80 -34.80
SC 0 0 9.25 0 -34.80 0 0 0
GS 0 0 0.0254 ' All in in.
GE 0
GN -1
FR 0 1 0 0 585 0

This cuts the single frequency calculation time of kissmeshlpda.nec from 2327 seconds down to about 8.4 seconds, and a full 470 - 698 mhz 6mhz step sweep is done in 318 seconds, heh.

NEC file with the lpda feed :

CM KISS Reflector using SM and SC cards by 300ohm
CM 6ft by 6 ft KISS parabola Reflector
CM oldsparks uhf lpda feed
CM Density = 2 inch by 4 inch mesh
CM Size = 4 panels of 36 inch wide by 36 inch high
CE
SY x=34.6
SM 18 9 9.25 0 34.80 0 0 0
SC 0 0 6.75 34.80 0
SM 18 9 0 0 0 9.25 0 -34.80
SC 0 0 15.80 34.80 -34.80
SM 18 9 15.80 -34.80 34.80 6.75 -34.80 0
SC 0 0 0 0 0
SM 18 9 6.75 -34.80 0 15.80 -34.80 -34.80
SC 0 0 9.25 0 -34.80 0 0 0
GW 1 21 11+x -6.2204724 0 11+x 6.22047244 0 0.03188976
GW 2 21 8.83464567+x -5.5511811 0 8.83464567+x 5.5511811 0 0.03188976
GW 3 19 6.90551181+x -4.9606299 0 6.90551181+x 4.96062992 0 0.03188976
GW 4 17 5.21259843+x -4.4488189 0 5.21259843+x 4.4488189 0 0.03188976
GW 5 17 3.67716535+x -3.976378 0 3.67716535+x 3.97637795 0 0.03188976
GW 6 15 2.2992126+x -3.5433071 0 2.2992126+x 3.54330709 0 0.03188976
GW 7 15 1.07874016+x -3.1496063 0 1.07874016+x 3.1496063 0 0.03188976
GW 8 13 -0.023622+x -2.8346457 0 -0.023622+x 2.83464567 0 0.03188976
GS 0 0 0.0254 ' All in in.
GE 0
EK
TL 1 11 2 11 -125 0
TL 2 11 3 10 -125 0
TL 3 10 4 9 -125 0
TL 4 9 5 9 -125 0
TL 5 9 6 8 -125 0
TL 6 8 7 8 -125 0
TL 7 8 8 7 -125 0
EX 0 8 7 0 1 0
GN -1
FR 0 1 0 0 585 0
EN

the 12-1/2 gauge 48 inch tall/wide type so for an 8x8 piece
BTW, approx how much more gain does the 8X8 KISS have over the 6X6 KISS ?

Just an observation,
Looking at the old sparks lpda, it looks like it has average horizontal beamwidth of 65 degrees and an average vertical beamwidth of 110 degrees over the 470 to 698 mhz range. Fuzzy logic would indicate a KISS reflector height about 1.7 times the reflector width. For a 72 inch wide reflector, that would be about 122 inches. I dont think it needs to be that high, but it probably should be higher than its width. You probably dont need to buy 48 wide inch material. (unless going for 96" wide by 163" high, heh)

Ill have to see if SM and SC cards allow overlapping since theyre not really wires.

Yep, It allows for overlapping panels, with no Geometry or Segment Errors/Warnings. :) The 36" X 36" panel example above didnt really have overlapping panels as I just basically followed your kissmeshlpda.nec dimensions. On the 36" wide by 48" high panel example below (making a 6ft wide by 8 ft high KISS reflector), I specifically forced overlapping panels and all is good with tight seams, heh. Take a look in the 3D viewer.

CM KISS Reflector using SM and SC cards by 300ohm
CM 8ft high by 6 ft wide KISS parabola Reflector
CM SM panels
CM Density = 2 inch by 4 inch mesh
CM Size = 4 panels of 36 inch wide by 48 inch high
CE
SM 24 9 12.4233 0 46.3644 0 0 0
SC 0 0 9 34.7733 -2.41154
SM 24 9 0 0 0 12.4233 0 -46.3644
SC 0 0 21.42330 34.7733 -43.9529
SM 24 9 21.4233 -34.7733 43.9529 9 -34.7733 -2.41154
SC 0 0 0 0 0
SM 24 9 9 -34.7733 2.41154 21.4233 -34.7733 -43.9529
SC 0 0 12.4233 0 -46.3644 0 0 0
GS 0 0 0.0254 ' All in in.
GE 0
GN -1
FR 0 1 0 0 585 0

Allowing for overlapping panels makes it a bit easier to get coordinates on the KISS reflector panels.
What I do to get the coordinates for the SM and SC cards is to draw up panels using wires in Geometry Edit. (panels having dimensions of even multiples of the mesh size used so as to make for even cuts) I then rotate and move the panels into place, in this case 15 degrees, and let them overlap naturally.

Thank you for that model! I was plugging numbers and pulling it apart as soon as I saw it ( one example is worth more than 12 pages of directions to me especially an example that uses numbers I can recognize)
by replacing the numbers I recognized with variables now the optimizer can find the best angles of the planes.
I'll get around to following the links you posted (see if I missed something) I'm having way to much fun playing with the new toy you gave me to read the directions yet.:D
But just playing I had new adjustable ~6x6 and ~8x8 foot models within a half an hour.
[BTW, approx how much more gain does the 8X8 KISS have over the 6X6 KISS ?
the ~ 6x6 (actual ~69 )gets about 18.7 raw 18.5 net @ 585 MHz the model I made with your cards/example ~8x8 at about 93 " is hitting 20.6 raw and about 20.3 net @ 585 MHz
I think I would put in another T support near the bottom like you have at the top to make it even more twist resistant in the wind

I agree : that would be the hard part of using an actuator to support the bottom (and remotely change the shape of the dish) loose enough to slide up and down but tight enough to not twist/torque in the wind adds complexity and this is supposed to be a "simple" design.
But the weed-eater string could be useful to tune /decide what height to put the bottom T for your actual site/channels.
But the patch model (6x6) has already shown me that optimum feed distance at least is not linear with frequency. I'll back burner the dynamic adjustment ideas (for now)

Fuzzy logic would indicate a KISS reflector height about 1.7 times the reflector width... it probably should be higher than its width

Nothing at all fuzzy about that logic I'm sure your right. Don't know the ratio but that looks like a good estimate. I was going to move in that direction next. All the plots I look at I see the big lobe at about -15 degrees (on a vertical plane plot) where the gain is leaking over the top and bottom of the reflector.
(unless going for 96" wide by 163" high, heh)
I'm on it ! I think I've found my antenna. Wait how tall? OK how about 144" so the reflector still weighs > 25 lbs.

BTW have you noticed how you can see the shape/type of feed in the gain pattern from the back side of the reflector?

The only problem I see with this style of modeling is I need to polish it more so the fence stays 36 or 48 inches even as the angles change. now it varies slightly with different angles.

I may have killed the quick calc. times but here are the 2 models. I used a GX reflection card to help me make the planes all move the same.

They are so eloquent and small (thanks again 300ohm) that I can post both here.
8x8
CM KISS Reflector using SM and SC cards by 300ohm
CM 6ft by 6 ft KISS parabola Reflector
CM oldsparks uhf lpda feed
CM Density = 2 inch by 4 inch mesh
CM Size = 4 panels of~ 48 inch wide by 48 inch high
CE
SY x=45.02226
SY y=11.47579
SY z=10.50109
SM 24 12 y 0 46.4 0 0 0
SC 0 0 z 46.4 0
GX 400 011
GW 1 21 11+x -6.2204724 0 11+x 6.22047244 0 0.03188976
GW 2 21 8.83464567+x -5.5511811 0 8.83464567+x 5.5511811 0 0.03188976
GW 3 19 6.90551181+x -4.9606299 0 6.90551181+x 4.96062992 0 0.03188976
GW 4 17 5.21259843+x -4.4488189 0 5.21259843+x 4.4488189 0 0.03188976
GW 5 17 3.67716535+x -3.976378 0 3.67716535+x 3.97637795 0 0.03188976
GW 6 15 2.2992126+x -3.5433071 0 2.2992126+x 3.54330709 0 0.03188976
GW 7 15 1.07874016+x -3.1496063 0 1.07874016+x 3.1496063 0 0.03188976
GW 8 13 -0.023622+x -2.8346457 0 -0.023622+x 2.83464567 0 0.03188976
GS 0 0 0.0254 ' All in in.
GE 0
EK
TL 1 11 2 11 -125 0
TL 2 11 3 10 -125 0
TL 3 10 4 9 -125 0
TL 4 9 5 9 -125 0
TL 5 9 6 8 -125 0
TL 6 8 7 8 -125 0
TL 7 8 8 7 -125 0
EX 0 8 7 0 1 0
GN -1
FR 0 1 0 0 585 0
EN
and 6x6
CM KISS Reflector using SM and SC cards by 300ohm
CM 6ft by 6 ft KISS parabola Reflector
CM oldsparks uhf lpda feed
CM Density = 2 inch by 4 inch mesh
CM Size = 4 panels of 36 inch wide by 36 inch high
CE
SY x=33.50276
SY y=8.183808
SY z=6.75675
SM 18 9 y 0 34.80 0 0 0
SC 0 0 z 34.80 0
GX 400 011
GW 1 21 11+x -6.2204724 0 11+x 6.22047244 0 0.03188976
GW 2 21 8.83464567+x -5.5511811 0 8.83464567+x 5.5511811 0 0.03188976
GW 3 19 6.90551181+x -4.9606299 0 6.90551181+x 4.96062992 0 0.03188976
GW 4 17 5.21259843+x -4.4488189 0 5.21259843+x 4.4488189 0 0.03188976
GW 5 17 3.67716535+x -3.976378 0 3.67716535+x 3.97637795 0 0.03188976
GW 6 15 2.2992126+x -3.5433071 0 2.2992126+x 3.54330709 0 0.03188976
GW 7 15 1.07874016+x -3.1496063 0 1.07874016+x 3.1496063 0 0.03188976
GW 8 13 -0.023622+x -2.8346457 0 -0.023622+x 2.83464567 0 0.03188976
GS 0 0 0.0254 ' All in in.
GE 0
EK
TL 1 11 2 11 -125 0
TL 2 11 3 10 -125 0
TL 3 10 4 9 -125 0
TL 4 9 5 9 -125 0
TL 5 9 6 8 -125 0
TL 6 8 7 8 -125 0
TL 7 8 8 7 -125 0
EX 0 8 7 0 1 0
GN -1
FR 0 1 0 0 585 0
EN

both optimized for 585 MHz

Thank you for that model! I was plugging numbers and pulling it apart as soon as I saw it ( one example is worth more than 12 pages of directions to me especially an example that uses numbers I can recognize)
by replacing the numbers I recognized with variables now the optimizer can find the best angles of the planes.
Yep, it really cuts down on calc times for mesh. But at the end of optimization, before building, a wire model should be made and run with the correct wire sizes and material to double check the results.

that would be the hard part of using an actuator to support the bottom (and remotely change the shape of the dish) loose enough to slide up and down but tight enough to not twist/torque in the wind adds complexity and this is supposed to be a "simple" design.
Oh wow that would be cool having the panels move for changes in frequency or F/B ratio. :p Using stranded wire to electrically connect the movable panels would work fine. Of course like you said, then its no longer a Keep It Simple Stupid antenna reflector, heh.

I used a GX reflection card to help me make the planes all move the same.
Cool, I havent studied or played around with the GX card yet. This looks like a good use for it.

BTW have you noticed how you can see the shape/type of feed in the gain pattern from the back side of the reflector?
Yep.

Oh wow that would be cool having the panels move for changes in frequency or F/B ratio.
Hmm, those old worm-gear motors from adjustable beds would do the trick once waterproofed. :p

I enlarged the height of the 8x8 four planed kiss model (to 12 foot total) and optimized for 585 but found less gain than the square ( equal z and y dimensions) 8x8 model. I'm pretty sure we need more vertical chines/planes to take advantage of the lpda pattern.
We lost a whole dB converting the vertical curve to a line vs .6 for the horizontal change. More sensitive to imperfections to the vertical curve.

old worm-gear motors from adjustable beds
They won't be wanting for power!
There kinda heavy and would need to run A/C up the tower.

In my recycle/reuse/re-purpose hayseed-engineering day dreams I use an old cordless drill motor chucked to a threaded rod and a couple of brazed/welded nuts. would need to add limit switches (from microwave oven door) and a dp/dt center off toggle switch (or 4 more door switches) to control it (reverse polarity) some cat5 cable should carry enough dc power to run the drill motor at intermittent duty.
If there is to much current/voltage drop then 6 of the 8 cat-5 wires could charge the drills battery and the other two control relays.
My daydreams reserve the bed motors and old garage door motors to elevate the whole mast from within the tower and stymie the zoning Nazi's 40' max antenna height rules.;)


I've thought about using a similar system to change the reflector length of the putty knife patterned trough model to achieve that pattern across the whole uhf band and alternately getting a 16-17 dBi narrower optimized pattern at any frequency, like this one (http://ghz24.freeshell.org/antenna/trparab187.html) (with a slightly inferior feed). The lpda feed model optimized at 16.55 dBi @473 MHz.
To see the performance of the adjustable trough parabola I'd need to re-optimize for each channels center frequency. Might eventually do that and record/graph the values of reflector length and feed point distance.
Just as a curiosity.
BTW those kiddie cars (electric barbie jeeps ect.) rarely get battery replacements and are usually thrown out with lightly used high torque geared DC 6-12 volt motors. (cause I don't know if a standard rotor will turn an 8x8 dish)

I enlarged the height of the 8x8 four planed kiss model (to 12 foot total) and optimized for 585 but found less gain than the square ( equal z and y dimensions) 8x8 model. I'm pretty sure we need more vertical chines/planes to take advantage of the lpda pattern.
I think youre right, Ive basically found the same thing on the 6' wide by 8' high model. I didnt get less gain, but got no significant increase.
So it was fuzzy logic after all, heh.

Ive found that vertical wires in a horizontally polarized world tend to narrow the beamwidth a bit without increasing gain on most antennas where the reflector is used in a normal back reflector setup. Since the parabola would be a front facing reflector in regards to the feed element, and a normal back reflector in regards to the signal, it does make sense that vertical wires have an even more dramatic effect.

motors??? you could use a satellite actuator they are 36vdc or 24vdc and are waterproof... even ice proof if you install a boot on the ram.

you could use a satellite actuator they are 36vdc or 24vdc and are waterproof
Yeah actually that was my first suggestion/thought back in post #308 and you are right to suggest them vs the drill motor they are superior in that they have built in limit switches (plus weatherproofed) and would make the proposed build easier for the average DIY builder (if he even chose to include this option).
But my hayseed engineer would use any actuators he gets to point the big brother of this build (perfected) at a satellite.
His main duty is actually to moderate arguments between the head engineer and the CFO.
All kidding aside I'm trying to remember how those actuators worked.

If both ends are not kept from rotating does it extend or just turn and not extend ? And can you twist the extending rod when device is off.
Normally both ends would be attached to something that would stop both ends from turning.
Because in this application if the actuator resists turning when moving and stopped then the wind wont be able to twist the bottom of the antenna.
Which would be optimal.

got no significant increase.
So it was fuzzy logic after all, heh.

Regardless of those results I still believe the logic is sound.
I have a 2 patch card 8x12 optimizing now that's hitting 22.3 dBi. Now there are 8 planes total instead of 4 and this model is flawed (the two planes separate, but it hints at significant improvement.
I kind of regard the kiss models as a worst case model, very hard chine compared to the actual desired product or the intended build.

I'm going to return to wire models again (even though these are useful ) I need more control to do a rectangular example.
here's the flawed model so you can see the problem
SY x=54.52699
SY y=8.168075
SY z=8.81461
SY w=46.4
SM 24 9 y 0 46.8
SC 0 0 z w 0
SM 12 9 y 0 46.8 15 0 71
SC 0 0 21.3 w 71
GX 400 011
GW 1 21 11+x -6.2204724 0 11+x 6.22047244 0 0.03188976
GW 2 21 8.83464567+x -5.5511811 0 8.83464567+x 5.5511811 0 0.03188976
GW 3 19 6.90551181+x -4.9606299 0 6.90551181+x 4.96062992 0 0.03188976
GW 4 17 5.21259843+x -4.4488189 0 5.21259843+x 4.4488189 0 0.03188976
GW 5 17 3.67716535+x -3.976378 0 3.67716535+x 3.97637795 0 0.03188976
GW 6 15 2.2992126+x -3.5433071 0 2.2992126+x 3.54330709 0 0.03188976
GW 7 15 1.07874016+x -3.1496063 0 1.07874016+x 3.1496063 0 0.03188976
GW 8 13 -0.023622+x -2.8346457 0 -0.023622+x 2.83464567 0 0.03188976
GS 0 0 0.0254 ' All in in.
GE 0
EK
TL 1 11 2 11 -125 0
TL 2 11 3 10 -125 0
TL 3 10 4 9 -125 0
TL 4 9 5 9 -125 0
TL 5 9 6 8 -125 0
TL 6 8 7 8 -125 0
TL 7 8 8 7 -125 0
EX 0 8 7 0 1 0
GN -1
FR 0 1 0 0 585 0
EN

Because in this application if the actuator
Looking at the Optimizer and Evolver as it was running, I only noticed maybe about a 1.5 dBi difference at various variable y and z distances. An actuator setup would be a big expense in time and money just to tweak for an extra 1.5 dBi.

I ran the Optimizer and Evolver on the 6 X 6 model for a freq sweep on the x, y, z variables from 470 - 527 mhz because I wanted to squeeze every last bit of low end gain out of it, even if I had to sacrifice top gain. It looks pretty good and compact. Average Net Gain over 470 - 527 mhz of about 17.4 dBi. Lowest Net Gain of about 16.7 dBi at 470 mhz. Highest Net Gain of about 18.2 dBi at 638 mhz.
Much, much better gain than the little smaller Finco P-5 parabola (post #75 of this thread). And better low end gain and about the same high to ch 51 gain of the 7 ft CM-4251 (post #6. Add only 1 to their dBd numbers to get dBi)
And of course it blows away anything here:
http://www.hdtvprimer.com/ANTENNAS/comparing.html

NEC file:

CM KISS Reflector using SM/SC and GX cards by 300ohm and GHZ24
CM 6ft by 6 ft KISS parabola Reflector
CM oldsparks uhf lpda feed
CM Density = 2 inch by 4 inch mesh
CM Size = 4 panels of 36 inch wide by 36 inch high
CM Optimized for 470 mhz to 527 mhz range
CM AGT = .98 (-0.1db) at 585 mhz, Autosegmentation = 21
CE
SY x=26.93947
SY y=9.022773
SY z=7.227386
SM 18 9 y 0 34.80 0 0 0
SC 0 0 z 34.80 0
GX 400 011
GW 1 21 11+x -6.2204724 0 11+x 6.22047244 0 0.03188976
GW 2 21 8.83464567+x -5.5511811 0 8.83464567+x 5.5511811 0 0.03188976
GW 3 19 6.90551181+x -4.9606299 0 6.90551181+x 4.96062992 0 0.03188976
GW 4 17 5.21259843+x -4.4488189 0 5.21259843+x 4.4488189 0 0.03188976
GW 5 17 3.67716535+x -3.976378 0 3.67716535+x 3.97637795 0 0.03188976
GW 6 15 2.2992126+x -3.5433071 0 2.2992126+x 3.54330709 0 0.03188976
GW 7 15 1.07874016+x -3.1496063 0 1.07874016+x 3.1496063 0 0.03188976
GW 8 13 -0.023622+x -2.8346457 0 -0.023622+x 2.83464567 0 0.03188976
GS 0 0 0.0254
GE 0
GN -1
EK
EX 0 8 7 0 1 0 0
TL 1 11 2 11 -125 0 0 0 0 0
TL 2 11 3 10 -125 0 0 0 0 0
TL 3 10 4 9 -125 0 0 0 0 0
TL 4 9 5 9 -125 0 0 0 0 0
TL 5 9 6 8 -125 0 0 0 0 0
TL 6 8 7 8 -125 0 0 0 0 0
TL 7 8 8 7 -125 0 0 0 0 0
FR 0 0 0 0 585 0
EN

http://img404.imageshack.us/img404/2533/kiss6x6lowchannelrawgai.jpg
http://img20.imageshack.us/img20/4674/kiss6x6lowchannelswr.jpg

I think thats pretty much the max low end gain I can strain out of oldsparks' lpda feed. Now to do a wire model of it. :p

here's the flawed model so you can see the problem
Yeah, the top sections mis-aligning.

Oh, and Jase88 had posted pictures of a parabola with an lpda feed in posts #215 and #217 of this thread. Interesting to note that the lpda length on it is also 12 inches, like oldsparks lpda.

Very great work / very interesting work and ideas guys ... on the KISS (Keep It Simple?) parabolic reflector for UHF.

The Kathrein-Scala type antennas, I noticed, space the reflector rods closer near the center of the parabola. Increase the reflector rod spacing some, as they move furthur from the center.

(that must be an attempt to optimize material use - more rods near center of reflector give more returns on the investment of material use. Also due to the illumiation pattern thing...the way that works)

So idea - would adding a few extra rods or wires horizontally near center add anything?

Or .. overlap fence mesh for a little further between top & bottom parts - so that you get half fence spacing for the first few rods. That might add strength too - near the center.

Just some thoughts from what I observed from other antennas.

Does that idea make any sense?