First DIY dish (WOK used as reflector), need a little help

[rwgast_lowlevellogicdesin]

So I am setting up a small dish beam 2.4ghz wifi up a mountain behind my house to an ras pi, Acually the dish will most likely be connected to the PI and beamed down the mountain to a router running custom firmware with another homemade directional antenna, hoping I can get away with something much smaller on the router end.

Anyways the dish is is simple as I said I want to use it for wifi, and I may use it to some experimentation with other microwave frequency's by swapping the feed. Now I have seen people do all kinds of weird stuff online like taping a USB wifi card in front of a noodle strainer and claiming to get good results. Well I bought a 4 dollar WOK at wal mart and I would like to make the dish as efficient as something made from a WOK can be...

First of all this is where I am so far



A while back a guy I did some work on a lot of audio equipment for a guy, and besides payment he gave me four of these mic stands that make nice booming tripods for antennas, Im currently using two other ones for yagis. So you can see how I mounted the walk, first off is the fact that I punched that hole in the very center of the dish stupid? I mean that is the point where most of the RF energy is right, I plan to do this as an offset and I am under the impression the whole dish collects RF and then it is focused at the feed, so im not to sure how important the center of the dish  is.

Next the WOK was coated in teflon, which isn't a suitable conductor so Im stripping it with a drill and wire brush. I have to do some electroplating on a bunch of soldier tips soon, I was wondering if when I do the copper coating on the tips if I should throw the WOK in there... I am not sure what kind of metal the WOK is, it is magnetic so not aluminium, but to get the most out of it I would think plating it in copper (or silver if I could afford that) will give the collector a bit more oomph. Im not sure if that is over thinking and over engineering but I am already plating stuff so its really no big issue to do the WOK at the same time.

Im a bit unsure of one more thing when it comes to dishes... so you build your feed and the coax runs to it with the braid connected the ground part of the dish feed or antenna, and the signal part connected to the biquads or helix or whatever feed design you choose. What about the acuall dish though, should this either connected to RF ground or earth ground? Usually in a directional antenna the reflector is connected to ground, but I dont see that on dishes, or at least i haven't noticed it.

Lastly is there a good bit of software out there for simulating a parabolic reflector based on frequency? We all know any junk from a can wave guide to a WOK will work and even improve things, but it is no match for something proper... I think I have line of site to a friends house where I am going to put up a second remote controlled radio station, but he is pretty far maybe 15 miles or even a bit more. Like I said I think due to my house being at a higher elevation we can get line of site but Im sure a WOK is not going to do the trick to run a a full speed 802.11 G network between our houses.

[wraper]

Wrong geometry to be useful.

Next the WOK was coated in teflon, which isn't a suitable conductor so Im stripping it with a drill and wire brush.

[voltz]

This is a real 'thing'. Its called WokFi. https://en.wikipedia.org/wiki/WokFi

Get it right and you're looking at +10db gain.

A bit mad - and i like it good luck.

[Howardlong]

Probably going to regret coming in here ;-) but I'll give it a shot.

First, old satellite dishes are two a penny these days. Most are offset fed, which means the feed doesn't obscure the dish. You really need about 2' or more at 2.4GHz. Properly setup, you can expect about 19-20dBi of gain.

Second, the Pringles can is the wrong size for 2.4GHz, it's too small. Anything that professes to "work" at 2.4GHz with this is in spite of the Pringles can, not because of it.

Third, for a feed to work you need to consider the depth of the dish and the illumination, including losses due to overspill or under illumination. Any feed you make needs to be made with care and positioned correctly at the focal point of the dish. In addition any polarisation you choose should match both ends. That's the physical side. The transmitter/receiver should be positioned at the feed point as feedline losses are pretty nasty in the GHz, and care should be taken in matching the antenna feed impedance to the transceiver's.

Here's a design I made many years ago, still very valid today. http://g6lvb.com/60cm.htm but back then WiFi was only just starting up, we had the band to ourselves.

[Ian.M]

+1 for  re-purposing an old satellite dish.  You can also use a cantenna feed, but you need scalar rings to get the correct illumination.  See http://www.qsl.net/ki7cx/wgfeed.htm

[metrologist]

A long while back, I gave the 23cm ATV band a try using a C-band sat TV dish and a double biquad antenna I built for the feed. Lots of gain.

[Whales]

Next the WOK was coated in teflon, which isn't a suitable conductor so Im stripping it with a drill and wire brush. I have to do some electroplating on a bunch of soldier tips soon, I was wondering if when I do the copper coating on the tips if I should throw the WOK in there... I am not sure what kind of metal the WOK is, it is magnetic so not aluminium, but to get the most out of it I would think plating it in copper (or silver if I could afford that) will give the collector a bit more oomph.

This shouldn't make a difference.  The non-metallic teflon should be transparent to the radio waves and most dishes have a coating of paint on them anyway.  Just as the air around us is (mostly) transparent to the RF you are using, even though it's not a metal.

[Ian.M]

While carefully lacquered or gold flashed  heavy silver plating is ideal for microwave work, even copper plating your wok would be gilding a turd.  Far more important is the accuracy of the profile - if the surface deviates from the ideal by more than 5% of the wavelength, its crap!   At 2.4GHz that means you need a Wok that's within +/-6mm of parabolic, which will be quite difficult to find.  You may have to cut down a larger wok, as off-profile edges will do more harm than good.

As the wave is reflected, 5% of a wavelength profile error results in a 10% of a wavelength path length difference.  50% of a wavelength path length difference would be antiphase - total destructive interference - so 5% of a wavelength profile error is quite a lax spec.

[CatalinaWOW]

Agree with all that the WOK won't help much.  Might get 3dB if you are lucky.  You can show yourself why.  RF, just like light can be view either as particles or waves.  The incoming RF beam can be view like a field of these particles (yes they are photons, just like at optical frequencies).  They are all going parallel to each other.  When they hit something conductive they bounce off, and a line halfway between the path of the reflected particle and the incoming particle will be perpendicular to the local surface.  If you plot a segment of your wok and plot all of the bounced photons you will find up that none of them end up at the same place, like they would with a nice parabolic reflector. 

If you are into metal work, and can make a good hot fire, you can pound your wok into an appropriate shape, but it is much easier to look around your neighborhood for an unused DishTV or DirecTV dish and talking them into letting you clean up their property.

[hendorog]

While carefully lacquered or gold flashed  heavy silver plating is ideal for microwave work, even copper plating your wok would be gilding a turd.  Far more important is the accuracy of the profile - if the surface deviates from the ideal by more than 5% of the wavelength, its crap!   At 2.4GHz that means you need a Wok that's within +/-6mm of parabolic, which will be quite difficult to find.  You may have to cut down a larger wok, as off-profile edges will do more harm than good.

As the wave is reflected, 5% of a wavelength profile error results in a 10% of a wavelength path length difference.  50% of a wavelength path length difference would be antiphase - total destructive interference - so 5% of a wavelength profile error is quite a lax spec.

Not suggesting this is practical, but given that the wok is the wrong shape, could it be corrected using a dielectric of the correct shape? i.e. using something like Teflon (just kidding it was too thin , PTFE, ABS, or PLA (i.e. 3d printable).

[Ian.M]

Theoretically, that could work, but you'd need to have a very accurate measurement of the wok shape, a supercomputer running custom raytracing software to iteratively refine the dialectric lens, and if you use the wrong plastic, dialectric absorption will loose a significant part of your signal.

In practice, if you've got the capability to fabricate the dialectric lens, it would be easier to simply produce a high accuracy plastic paraboloid (by CNC machining, 3D printing or spin-casting resin), and silver-plate it for use as a first surface reflector.

[Nuno_pt]

+1 for reconvert an Offset TV dish, 40 or 60cm ones are dirty cheap, even free sometimes.

[hendorog]

In practice, if you've got the capability to fabricate the dialectric lens, it would be easier to simply produce a high accuracy plastic paraboloid (by CNC machining, 3D printing or spin-casting resin), and silver-plate it for use as a first surface reflector.

Yes that is an interesting approach.

I've played with 3d printing a substrate and using copper tape as the ground plane and the tape or solar panel tabbing wire as the conductor. There is some loss which might be due to the many small internal voids and boundaries created by 3d printing. Probably would be better just melting the plastic into a mould.

[Ian.M]

I'd suggest printing the dish backing with stiffening ribs and a honeycomb structure as hexagonal tiles that fit your printer bed,  gluing them all together to form the rough paraboloid, then spin-casting epoxy in it to form the precision paraboloid surface - which requires a carefully levelled hacked record player turntable with quartz speed control, then plating - probably mist with nickel paint to make it conductive, heavy copper plate and finish with silver then burnish it.  Done right, that can make an optical grade mirror.

[LaserSteve]

For feed design, Start Here:

http://www.w1ghz.org/antbook/contents.htm

The flat may or may not be a hindrance, depending on your design skills. If you had proper test gear, you can get some gain out of the flat Bottomed Wok, but not as much you could get from a true parabola.   One of the problems with having the flat there is that it is a reflective plane with no gain. The math is tricky for that when you look at the lobe pattern caused by the curved edge of this dish.  Until you have some experience with feed design, you might want to find a better Wok, rather then having to reinvent the " Resonant Cup" antenna.

There was no reason to strip the Teflon in most cases, and adding plating will not decrease your losses enough to be notable without test gear.

Here is some inspiration. Matjaz, S53MV  is a RF professional, so he has test gear both commercial and home made to validate his designs.

http://lea.hamradio.si/~s53mv/wumca/cavity.html

Steve

[rwgast_lowlevellogicdesin]

Thanks I will look at the link

For all the nay sayers this was in the wiki posted ny another memmber



this is the EXACT wok I bought I can tell by the handels I drilled out. He claims 17db of gain, ill take that with a grain of salt but im sure it did something for him, when I get a chance I will stick a dongle through the center hold I drilled like he did and see if I can duplicate a 17DB result... granted he is doing this the kind of half ass way without using a feed, if the WOK works Im probably going to plate it anyways more for aesthetics as I said I already have a batch to plate so no bigge...

If the WOK idea sucks as most people seem to agree on I am really only shooting the signal up a 100ft tall mountain behind my house and I am about 200 yards from the base.. Im not shooting a crazy distance I just want full bars and directionality so I am not all over the place... I have made quite a few decent patch antennas for L-band geo-sync stuff, and im even working on an 8 patch design that can be electrically steered using special RF relays.

Would a patch at each end be good enough? The only reason I am going the dish route is the higher directionalality (dont want a bunch of people seeing the network) and experimentation really. Secondly I have always wondered but never tried how much would a good LNA help, Wifi puts out such a low power signal it shouldnt cause any problems blowing the LNA on TX. I have only found one commercial LNA for  wifi and it was like 80 bucks and I dont think it specs were as good as a newer mini circuits chip, I dont think polluting the bands would be an issue with high directionality. This is all the first step in a 20 mile link, where I would like to use a sim and hand make a BBQ grill style antenna. I am pretty sure the houses are line of site I know theres nothing man made in the way, would the best way to figure this out before doing any work be to look at a topographical map or something, I did notice the highway between are house is at a bit of a curve so there could be a hill/mountain obstructing line of site not to visible wit the eye.

[Ian.M]

I've used a cantenna with a USB stick in it at one end and an ordinary unmodified  domestic router at the other for a 400m link before now.  Its only workable if you have a clear line of sight and low background noise.  Put a carefully aimed cantenna at each end of the link and you should be good for half a Km without any problems.

[CatalinaWOW]

Just a glance shows that the wok is not parabolic as the wiki claims.  Doesn't mean there won't be some gain.  Just putting your dongle in front of a flat plate, spaced for constructive interference will give 3 dB.  If one of the constructive interference points coincides with a "random" reflection from the rim a few more dB may result, but 17dB just doesn't seem possible.

If repurposing kitchenware floats your boat there are large stainless steel mixing bowls which aren't bad spherical approximations.  They would give better performance than a comparably sized wok.

I have played with the spin casting idea.  Never got close to optical grade, but it wasn't too hard to get plenty good enough for 2 to 5 GHz.

[Ian.M]

I have played with the spin casting idea.  Never got close to optical grade, but it wasn't too hard to get plenty good enough for 2 to 5 GHz.

For the record: How good was your speed control on the turntable?  Did you vacuum degas the resin before pouring?  Did you shield the resin from air drag by cling-filming across the top of the former?

I would expect optical grade spin casting to have a very steep learning curve, and it may even require exotics like UV cure resin so you can initiate the cure uniformly over the whole surface to minimise distortion.

[CatalinaWOW]

I have played with the spin casting idea.  Never got close to optical grade, but it wasn't too hard to get plenty good enough for 2 to 5 GHz.

For the record: How good was your speed control on the turntable?  Did you vacuum degas the resin before pouring?  Did you shield the resin from air drag by cling-filming across the top of the former?

I would expect optical grade spin casting to have a very steep learning curve, and it may even require exotics like UV cure resin so you can initiate the cure uniformly over the whole surface to minimise distortion.

I used plaster of paris in some experiments just to see how hard this was.  An initial article implied that it was falling off a log easy.  Used a phone turntable of average accuracy.  I concluded that your guess is right.  There is a lot of work getting optical grade results, though there are reports of using fluid mercury.  Todays folk are so afraid of mercury that few would try it.

By the way, speed control isn't the real problem, it is acceleration control.  Once the fluid in your pot gets out of balance it takes very solid mounting and a lot of time to get where you want to be.

[Howardlong]

Thanks I will look at the link

For all the nay sayers this was in the wiki posted ny another memmber



this is the EXACT wok I bought I can tell by the handels I drilled out. He claims 17db of gain, ill take that with a grain of salt but im sure it did something for him, when I get a chance I will stick a dongle through the center hold I drilled like he did and see if I can duplicate a 17DB result... granted he is doing this the kind of half ass way without using a feed, if the WOK works Im probably going to plate it anyways more for aesthetics as I said I already have a batch to plate so no bigge...

If the WOK idea sucks as most people seem to agree on I am really only shooting the signal up a 100ft tall mountain behind my house and I am about 200 yards from the base.. Im not shooting a crazy distance I just want full bars and directionality so I am not all over the place... I have made quite a few decent patch antennas for L-band geo-sync stuff, and im even working on an 8 patch design that can be electrically steered using special RF relays.

Would a patch at each end be good enough? The only reason I am going the dish route is the higher directionalality (dont want a bunch of people seeing the network) and experimentation really. Secondly I have always wondered but never tried how much would a good LNA help, Wifi puts out such a low power signal it shouldnt cause any problems blowing the LNA on TX. I have only found one commercial LNA for  wifi and it was like 80 bucks and I dont think it specs were as good as a newer mini circuits chip, I dont think polluting the bands would be an issue with high directionality. This is all the first step in a 20 mile link, where I would like to use a sim and hand make a BBQ grill style antenna. I am pretty sure the houses are line of site I know theres nothing man made in the way, would the best way to figure this out before doing any work be to look at a topographical map or something, I did notice the highway between are house is at a bit of a curve so there could be a hill/mountain obstructing line of site not to visible wit the eye.

330mm for 17dB(i?) gain at 2.4GHz? Sorry fella, that's pure bullshit I'm afraid.

Whether a patch works well or not depends on the illumination it provides for the dish so typically you will need a deep dish or it will over-illuminate leading to reduced Tx power but more importantly excess noise from the warm earth on receive. That is why a horn/can feed is used, or, more easy to fab and match, a helical of the right length, to reasonably illumnated the dish.

Get an old sat TV dish and make a feed as I described. I spent many many months on this all those years ago, I have offered you a real tried and tested solution which is easy to make, including getting a reasonable impedance match and focal point for example with nothing more than a ruler, bit of string and protractor.

Place your router very close to the dish with the shortest low loss feed you can achieve. An LNA is all very well but it will have losses due to tx/rx switching.

Feel free to experiment yourself, but please also take time to read and take time to understand the content of the links that I and others have provided. The content I wrote all those years ago, believe me, was the result of many hundreds of hours and lots of failure, I had a dumpster full of bent aluminium. What you see on the pages I linked are the results of what worked, was repeatable, and was tested on a calibrated antenna range. There was a ton of other stuff that didn't work, I am trying to help you avoid those same failures.

One final thing, using directional RF to try to hide yourself is like putting your house keys under a plant pot instead of the mat. I don't know what you want to hide, or why there may be an apparent paranoia about it, but RF isn't the way to achieve it.

[rwgast_lowlevellogicdesin]

I guess I shouldn't have said "hide". I realy meant I dont want everyone in the world seeing my SSID and I dont want to see theres... I want a noise free path that others wont crowd in on, or even know is there unless there actually looking for it. 2.4ghz is just overly jammed up and I figure the tighter the beam the less chance I have crossing some other 2.4ghz gear.

Also when I mentioned patch antennas I didn't not mean I wanted to use them as a dish feed, just cut some plane old metal and make a patch antenna which is pretty easy. I have no problems reading your info, and I have seen tons wifi stuff done with direct tv dishes, I am sure it works better than a the best WOK any day of the week, Im just trying to keep things a little bit compact easy to climb a mountain with and easy to mount up there and make sure it doesn't blow over when we have 20mph wind. I am 100% a cheapo Yagi from ebay on both ends would suit me just fine bet there is no fun in that... a have a 9db UHF 5 element yagi and that thing will receive as far as I can see in my town.

The reality is the WOK was 4 dollars if it works like crap i dont mind tossing it, and I definately will take your advice on feeds etc, but I would like to use a dish, just something smaller than a direct TV dish. I have no problem trying to sim a parbola for the size I want and them framing it out from hard drawn 14 guage then laying a soli flexible metal over that.

Also feed loss wont be a problem the dish will be connected to a RasPi 3 where the wifi chip antenna is desoldiered with a run of LMR 400 less than 10ft long.

[CatalinaWOW]

If dish size is your issue it is relatively easy to trim the sat TV dishes.  They are really not that much bigger than your wok.  Far easier than building a dish from scratch.  And from the penny pinching standpoint, it really is fairly easy to come up with free sat TV dish.

I do agree that carrying stuff up a mountain is tough, and reducing the load is worth it.  That said, those home tv dishes are already designed for far higher wind loads than you mentioned (when mounted appropriately).  The mount will be the biggest issue for any antenna you use.  Unless you have a mature and fortuitously placed tree, or an even more fortuitously placed rock formation, you will be carrying some structural material and concrete up your hill.  Maybe time to go old school and find someone in the area who wants to demonstrate how much better a mule is than any of those new fangled contraptions.

[rwgast_lowlevellogicdesin]

Im fourtante to have plenty of heavy rock at the top . You say you can trim the dTV dishes how would you do that without messing them up?

[Howardlong]

I did some calculations a while back but once the dish gets under a certain size it becomes a much less reasonable proposition: ISTR about 18" is the minimum size at 2.4GHz where it's time to look at other options.

This is all about aperture area, in short the more (carefully manipulated) metal you have the better.

One reasonably compact method is a patch array. I would hasten to add that patch antennas are inherently narrowband, and impedance matching them even on their own demands the right skills and equipment. The added complexity and losses of a power splitter/divider make this approach difficult to make work consistently unless you're well equipped. Care also needs to be taken to optimise the stacking distances.

That is why I am a proponent of the axial mode helix as a dish feed, which is broadband, and if you are careful following the matching dimensions of the 1/4 wave impedance matching taper section in my notes, it is hard to go wrong. You can also tailor the beamwidth to best illuminate any given dish by changing the number of turns. Be careful of following instructions that leave the helix immediately attached to a former such as plastic/PVC pipe: as well as introducing unknown losses, this has a very significant dielectric effect and _will_ dramatically detune the antenna unless the dielectric effect is known and taken into account in the design at the outset, mostly it is not.

Furthermore, avoid using supports that are off-centre, whether metal or not, they too significantly move the beam off-centre as well as introducing destructive interference on the off-centre beam anyway, reducing the gain by many dB. Typically, longer non-self-supporting axial mode helices are best constructed with a relatively small centre support with small insulating standoffs to support the helix. For dish feeds which are typically only a few turns, the helix can be self-supporting.

One other option of you want portability is to use an umbrella dish. I've made quite a few of these and they perform excellently. You can also sometimes find commercial ones on ebay. Be careful of wind loading though.

[CatalinaWOW]

Im fourtante to have plenty of heavy rock at the top . You say you can trim the dTV dishes how would you do that without messing them up?

I would first glue a stiffening rim on back (or front if you can come up with a material with low enough dielectric constant) at the size you want, and then use a cut off wheel in an angle grinder to do the actual surgery.

But I would also recommend Howardlong's post just previous to this.

[Ian.M]

I don't see the problem with toting a 20" (50cm) dish up the mountain.  Unless the slopes are heavily wooded , with dense undergrowth, even a 24" (60cm) dish wouldn't be too difficult to get up there.  Unassembled, all the smaller dishes fit in a fairly small flat-pack, and the weight and volume to be backpacked up the mountain won't be significantly different for a small dish or a woktennae.   The bulk of the cargo-packing will probably be the solar panels, battery, tools and readymix cement and water  to grout the installation into place.

There is therefore little point in cutting the dish down and compromising its performance, especially when the objective is minimum beamwidth to make it more difficult for unauthorised users to access the system.

[rwgast_lowlevellogicdesin]

No heavy woods, live in the desert the mountain is just steep and unaccesible even by a two wheeler dirt bike. This is good becuase if you can get to it on a dirt bike someone will destroy steely it. The terrain is rocky and its gonna be 110+ for two more months. On top of that the pie will have 6 more SDR's plugged in for an electronically steered lband patch array, ADSB, ACARS, ARPS , a Discone made from copper tubbing, and two SDRs using 800mhz band antennas for unitrunking public scanning stuff. This set up also requires me to lug a few solar panels and APC SLA trype batteries not to mention RasPi 3 with enough coax to space the antenna systems properly and USb Cables. Im definately making more than one climb but I want to pack what I can in each hike hike with a 30-50lb pack. And in reality I think acward sizing will be my problem Here is my physical challenge



I also have two more properties I will be duplicating most of the station on, one is 15 mile north over that hill at higher elevation on a 40ft mast and another 15 miles east at lower elevation on a barn. I will probably drop ACARS,ADSB on those stations and maybe L-Band, not really sure if having L-Band arrays that close will make any difference really. I already know how to RX and have sucessfully decoded iridium text, but I want to be able to transmit too (im brewing a wideband transmitter for each site), im not sure how to get an iridium or imersat subscrition to TX without buys 200 dollar modems or a sat phone. The idea of this is to be able to drop projects in the middle of no where of even try a bslloon and be able to communicate with it. Each progect and RF site will have GPS, this is a two fold thing #1 to make sure if my stuff gets ripped off I can track it over L Band (instead of wasting to much on GSM) and #2 to is to compute distances and bearings between any projects and base stations.

One last thing sort of unrelated im always hearing if you want good HF you antennas have to be WAY higher than the ground like 100ft+, well doe that mean the actual ground? If you just live under a 100ft mountain and then stick a vertical up there is that the same thing as being a hundred feet over the ground, im guessing no...

[Richard Crowley]

They make outdoor speakers disguised as rocks.  And cell-site towers disguised as evergreen or palm trees.
Perhaps you could make a mottled-gray color radome that looks like a rock to hide your reflector inside.

The surface-finish (or lack of) makes no difference to reflecting RF. 
You could argue that the Teflon (or whatever) coating will extend it's life exposed to weather.

But the shape (cross section) will make a huge difference.
There is a very good reason for a parabolic shape.
That angular shape bowl may be little better than a flat reflector.

[1design]

If you are into making antennas from cookware you might want to build one of these:

http://lea.hamradio.si/~s53mv/wumca/archery.html

BR

[rwgast_lowlevellogicdesin]

Well you guys are right I have been doing a bit of simulation and some testing with the WOK it doesnt do to much, im sure directionalty aside a simple col-linear would do a better job.

I was wondering if I flipped it up side down would it make a good ground plane? I mean its flat and angles down a bit more than 45 degrees but still. I mean it was only 4 bucks and I dont mind throwing it away but it seems it should have some use?

Wow that "Archery antenna" looks pretty cool, that guy put a lot of effort in to that thing!