Use tape to increase the range of ESP8266 | Hacker Day

2021-12-13 18:54:19 By : Mr. Eric Wan

For a long time, I have wanted to build a real and suitable radio telescope. To me, this means a large parabolic reflector, a feed speaker made of brass plates, a hanger wire, and at least in the initial experiment, an RTL-SDR dongle. I have completed the calculations, checked the old C-band antenna on Craigslist, and even designed one or two brackets that can point to the antenna. I have made enough plans, knowing that the results will not be great. After months of work, the best result I can hope for is a very low-resolution flat image of the Milky Way. With any luck, there may be a bright spot corresponding to Sagittarius A.

There are better ways to build a radio telescope in your backyard, but the thought of putting a huge parabolic disk behind and staring at the sky has always bothered me. I even designed a plate that can be easily disassembled and transported, because making your own plate is much cooler than buying West Virginia flowers from someone on Craigslist.

Recently, I was asked to design a futuristic space prop for the upcoming video. My customized, easy-to-transport parabolic antenna immediately came to mind. The idea of ​​a three-meter-diameter parabolic dish was rejected because it is more practical and cheaper, but I did do more calculations, opened a CAD program, and started the actual work design. As a test, I decided to 3D print a small model of this dish. When creating this model, I accidentally used 3D printed parts, a bit of epoxy and tape to create the perfect WiFi antenna for the ESP8266 module.

The design of the parabolic WiFi antenna is not much different from the design of the optical telescope. In fact, there is no difference other than the frequency of the light we are observing. A rule of thumb in optics is that the surface finish of a perfect reflector needs to be between 1/10 and 1/20 of the wavelength of the light used by the reflector. For a visible light telescope with a wavelength of 400 nanometers, this means that the surface finish of the parabolic mirror needs to be within 40 nanometers. This is the size of a virus, but it can still be made of cerium oxide, pitch, and a second piece of glass.

On the other hand, the wavelength of WiFi is 12.5 cm, and the surface of the parabolic WiFi antenna needs to be perfect within a millimeter or two. This can be easily achieved with a 3D printer.

In order to design this dish, I found some old free software on the Internet, which can measure the relative measurement values ​​of a disk-focal length and diameter-and spit out the XY coordinates of the parabolic arc. These coordinates can then be imported into the CAD software of your choice, extruded into a 3D shape, and then exported as a 3D printable object.

This project was initially just a test to see how I could build a large, easy-to-transport parabolic reflector. This requires some design considerations. First of all, this dish should be composed of segments, which should fit my car. I decided to build a hexagonal plate. Every sixth of the hexagon consists of the left and right half. This made the various parts of the dish small enough to fit in my car, and it gave me the added benefit of only needing to create two masters to remove the mold.

There is indeed a small disadvantage of making the plates into hexagons: I lost 17% of the plates. The area of ​​the hexagon inscribed by the circle is 83% of the circle itself. However, having straight lines on the outside of the dish can simplify any reinforcement or support I need to build. In any case, this is just an experiment.

Another design consideration is how to connect the antenna to the bracket. Ideally, you want to install it near the center of gravity. In the case of a circle, this is the center. For parabolic disks, this is not a problem; in any case, we do not need the center of the disk because it will always be hidden by the feed horn. Cut off the center part of the plate, we have a nice hexagonal hole as a mounting point.

Make some quick edits in Fusion 360, export to STL, and put it into Cura. I ended up with this:

This is 1/12 of a disk, and another 1/12 of the disk is copied and mirrored along the Y axis. No, one-sixth of this dish can be printed in only two days; I increased the layer height and reduced the filling, but it still takes about 10 hours to print these two parts. In total, this experimental disk requires 60 hours of printing time and two coils of filament.

In addition to the 3D printed disc part, I also need some other parts to turn it into a suitable WiFi disc. I need some way to mount this thing on the camera tripod, and need something to fix the ESP8266 on the focus of the plate. This is what I came up with:

The hexagonal block simply slides into the center of the plate. On the back, there is a 1/4-20 threaded insert for a camera tripod. In the front, there is a hole designed for 6mm pins, or in my case, a plastic knitting needle cut to the right size. I designed two mounts for the ESP8266 module. The first is the simple installation of the very popular and very cheap ESP-01 module. The second mount fits the Adafruit Feather Huzzah board, which is an ESP breakout board with some neat features, including a USB to serial adapter and a JST connector for batteries.

The parts are printed out and everything is assembled. There is still a problem with this experimental disk. It is relatively transparent to radio. Although I could have used metal filament in this build, I really didn't want to spend that much money. Instead, I found a better solution: tape.

In the United States, the term "tape" has multiple meanings. By far, the most common is cloth-backed tape coated with plastic, which is a close cousin of tape. This tape or "duck tape" may look like metal, but it is not. This is also something you should never use in plumbing work. Don't put these things on the HVAC system.

I use something you should use in plumbing, not a cloth-backed "duct tape". This "duct tape" is actually thick aluminum foil with a pressure sensitive adhesive on one side. It has radio reflectivity, strength and thickness, so it can be polished smoothly. I'm using tape to reflect radio waves, and doing so has created the best clickbait title ever.

I live in a van by the river in the middle of the woods, so it is not easy to test the range of the WiFi connection or the gain of the antenna. Ideally, I would go to the desert, go to the top of the mountain, and then drive with a friend on the walkie-talkie, dragging a WiFi router. The second best option is to build high-rise buildings in densely populated cities. Fortunately, I attended the Shmoocon at the Hilton Washington DC a few weekends ago. This is how I test my dishes.

After the Dishes and Donuts party at Shmoocon, I took my dishes to several floors and pointed to it on the vast townhouses and embassies. The experimental setup consists of Adafruit ESP Huzzah, programmed to report all visible SSIDs (with RSSI) via a serial connection. Without this dish, I can see about a dozen SSIDs. With this dish, I saw the largest number of about 150 SSIDs. This dish works very well, and it works very well.

The performance of this dish is very suitable for something pieced together with 3D printed parts and five-minute epoxy resin. This is a 19-inch diameter disc with an F/D ratio of 0.5. In theory, if it is circular and the center of the dish is not in the shadow of the horn, it should have a gain of 18.1 dBi. My hexagonal dish is not circular, and the entire area of ​​the dish is not illuminated, but I should still have a gain of about 16 to 17dBi. Compare it with a commercial Cantenna of 12dBi and remember that the decibel is on a logarithmic scale, I did a great job here!

This was originally an experiment to determine if it is possible to build a parabolic reflector that can be easily disassembled and reassembled. Did you make it? Yes. But there are some caveats.

I originally planned to metalize the plate, including 3M spray glue and aluminum foil. This is a complete failure. Although the surface of the plate may be only a millimeter or two from perfect smoothness, there are too many wrinkles and I don't want to deal with it. The best solution for a relatively wrinkle-free metal surface is tape.

I still have to describe this dish correctly. I know that the gain is good enough, and theoretically it should be much higher than ordinary, cheap, commercial antennas or WiFi patch antennas. Quantifying this difference is another matter entirely. I am open to suggestions on how to do this, but please remember that there are trees on top of all the peaks around my house. The sight is a luxury.

Like all my projects, I have put all the details of this project on This includes all .STL files used to print yourself. I would like to know if anyone has made their own 3D printed parabolic disk, and even more so if the design can be improved. If you manage to print out one of them, please send it to the Hackaday prompt line.

Or solder a wifi before 9 wifis 2, 4 GHz Wifi scan after welding 37 wifis

I mean the antenna *see picture

This seems to be a very suboptimal way of doing what you do. You are still connecting the original antenna, soldering the coaxial cable to its end, and then unshielded at any length. A better method might be to cut off the original antenna and solder the coaxial cable to the feed point, and place the shield as close to the ground as possible. Or better yet, just get the ESP8266 module with U.FL connector.

+1 For example, ESP07, which has the same footprint as ESP12.

Can I get the PCB file and component value? I have always wanted to print one for my wroom-02, but haven't found time to design it yet.

Where is the tape? I mean antenna* see picture

On the subject of real radio telescopes; I and many others have old steel C-bands? Sitting on a plate in our backyard to collect dust is usually an eyesore. Most even have electric coaxial drive actuators. Some older people may even remember how they work :). I bet you can see some stars with a 3m dish!

Especially if you put it on your head.

"There is a better way to build a radio telescope in your backyard, but the thought of putting a huge parabolic disk behind and staring at the sky has always bothered me."

Become a topic in the city. Make it look like a huge inflatable eyeball facing the sky.

According to the photo, it seems that the tape can be used for a little attention with polishing tools and some elbow grease. This may reduce scattering and slightly increase gain.

This may not make much difference. The optical tolerance rule of λ/40 is actually only applicable when the physical size of the antenna is much larger than λ. In this application, the size and wavelength are measurable, so the diffraction effect will be significant.

We call this metal tape "100 mph" tape.

We had a similar tape in the 1980s that was used for aircraft "war damage" repairs. I don't remember the details now. If you work in aircraft maintenance, you can actually take a course (in Clark, we have an old Navy F-4 to practice). Most of the repair methods we have learned will make MacGyver proud.

Obviously it is different from the 100mph tape that the U.S. military calls because it is cloth-backed. It needs to meet a large number of ASTM and MIL specifications. Metal backing tapes usually do not meet the tensile strength or temperature range requirements, but they do provide heat resistance up to 200F/98C, which is very good for some applications.

Among the guards, we tore open the soft top on the Hummer, and there were 90 miles away when it rained. Yes, its 100 mph tape is definitely more like 70 to us, but replacing the missing top of nearly square meters, there is no problem.

Why is it called a "feed horn"? From the photo, it seems that he has placed an omnidirectional antenna at the focal point of the parabolic reflector. It is impossible for him to obtain a gain of 16 dB.

In any case, this is click bait. Who cares, right?

This. It may be the biggest failure of the project, and it is also an easy problem to solve.

I think he was saying that a second reflector is needed before feeding. The second reflector can be as simple as a 4-inch curved metal plate placed at 1/4 wavelength in front of the ESP8266 module. If there is no second reflector, ESP8266 will be in the path of constructive and destructive interference between the incident wave and the reflected wave. Therefore, the main lobe gain and side lobe suppression will suffer up to 6dB of performance.

Or maybe there is no 3D printed chaotic wire fence? Remember, the 2.4 Ghz wavelength is not that small after all. But anyone who owns a 3D printer is looking for a way to justify its purchase.

The required surface accuracy of the reflector is approximately lambda/30, where lambda is the frequency. For 2.4GHz Wifi, this is approximately 4 mm. It is difficult to get a wire mesh to do this, but it is easy for 3D printed parts.

Edit: where lambda is the wavelength.

An error of wavelength/10 will only cause an attenuation of 1 dB. We can keep it within + – 12 mm, which is a good 1 inch tolerance.

Lambda/10 is more like a 7dB drop, where lambda/30 is 1dB away from the ideal value. Huge gap. But you are right, it is the accuracy of RMS. Under lambda/30 or even lambda/20 (only 2dB), the overall peak-to-peak error is 11-16mm. This is still difficult to do with barbed wire. If you can still do it within an inch, then you obviously haven't tried to make anything with barbed wire.

um, yes. Very good, often. But sometimes it takes a lot of effort, and 3dB is 2x pwr, to be accurate, you will have a dish antenna of almost 1-6 feet instead of two conjoined dish antennas. 2 The way to go. Good or big. choose. Sometimes, no matter...

Neat! You can also consider using copper tape instead of aluminum tape so that it can be welded to the seams

Out of curiosity, why don't you use an offset feed design for the antenna reflector?

ESP-01 will not cover the plate at all. Even things from Adafruit Huzzah won't have much influence on this dish. Ordinary (as opposed to offset) dishes make design and assembly easier. just because of.

It makes sense, especially since there are no feed horns that add shadows.

clickbait rubbish headline...bad antenna connection, bad center pole, ridiculous reflector (look at commercial antennas, do they give you any hints?)

Maybe their family was overwhelmed by an unsafe C-band TV disc.

There is nothing wrong with this. Well done. It can be improved by more complex mathematics and offset feed. The ESP antenna is more suitable as a horn, and its beam pattern fills the dish-shaped area. At present, because the small PCB antenna is omnidirectional, it loses a lot of directivity.

The design uses more plastic than needed. RF does not need a solid reflector, only a reflector with enough small holes.

It is easier to go to local appliances and cooking shops. Pick up the largest stainless steel cooking bowl and make a parabolic dish antenna. The beam width will be narrower and tighter. It is troublesome to use any type of tape.

Yes, I was just thinking about this. A standard wok. Or a cheap stainless steel mixing bowl. A plastic bowl with foil lining. Abandoned home satellite antenna...

The required metal tape is usually called flue tape and can usually be purchased at wood stove supply stores. It is essentially a very thick aluminum tape with a good adhesive on the back. Peel and glue things.

Also used for HVAC pipe work

There is no need to use both metal and cloth tape. The cloth will only make it rough and increase scattering.

There is an omnidirectional antenna on the module, which requires a second small parabolic reflector to capture the radio waves reflected from the main antenna, but there is no antenna to focus them on the front side of the antenna. Then the waves emitted from that side will be reflected to the main disc.

There is no reason to use a large enough printer to round the edges of the plate. In order to save the filament, please use the equal grid pattern on the back.

This is the McDonnell Douglas Isogrid design manual. Go to page 42 and you will see the dimensions of the grid walls and floors used in Skylab.

The isogrid stuff looks interesting, maybe someone should do something on Hackaday!

MechEng students are here, thank you! Will study it

FWIW, I wrote a spreadsheet that I used to create a paper craft sphere, in which the generated chart is half a bloodstain. If anyone is interested, it is not difficult to change the formula used to create the parabola. Although it consists of a (user-defined) number of flat plates, its advantage is that it can (laser?) cut and scribe from flat materials, which makes construction and transportation faster and easier. Everyone is welcome to use (and improve) it as you wish.

Wouldn't it be easier to just buy a satellite dish? Or I don’t know to get one for free at the scrapyard? This reminds me of 10 or 15 years ago, when I saw a brand new small (half-meter) parabolic satellite antenna in a store for about 15 euros, I bought it just because I thought it would be cool to own it.

After 5 years, I can’t receive wifi signals on the floor, so I want to buy a USB cable. Instead, I plug the wifi pen (without an external antenna, maybe the black asus 54bg 8 years ago) directly into the desktop USB port. It is the focus of this dish. I can get better reception.

Of course, this is not a hell hack I manage, but now I can always think of French fries dripping and hammering it into a parabolic shape, or any other small dot mesh/metal mesh. But just buying a high gain antenna is easy and cheap.

I did something very similar to my project, but I just used one of the $50 parabolic grid antennas on Amazon. The grid is a vertically polarized filter, so the surface needs to be covered with foil because the integrated ESP8266 PCB antenna is not vertically polarized. I haven't tried to remove my pattern because there is too much scattering in the room. To characterize the pattern, you need a LOS environment without multipath obstacles. The 40-foot separation should be good enough, but remember that the ground is also a source of multipath, so the DUT antenna and feed antenna should be as high as possible. The feed antenna should also be co-polarized with the DUT antenna. You can try to use RSSI for data collection, but the resolution may be quite poor, and the update speed of the wifi beacon is very slow. Expressive does provide a certification test firmware that can be used to put ESP8266 in continuous wave mode. Using the CW mode allows you to use the spectrum analyzer with the feed antenna for more accurate signal strength measurement. You should be able to reasonably characterize the main lobe according to my description, but accurate side lobe measurement requires an anechoic chamber or strict multipath correction technology.

I want to know if chrome spray paint can solve the problem. I think it has a good metal composition, right?

"...Cooking your own food is much cooler than buying a West Virginia flower from someone on Craigslist"

What did azaleas do to you?

You should use mylar/space blanket instead of tape. This is less wrinkled than aluminum foil, but has a higher reflectivity

Big enough to remember how they work? ;>) Now we see dish antennas placed in most open spaces, but I built a sled antenna for 2.6GHz to capture and decode HBO using a coffee can feed horn. It took one night to manually calculate and cut the circuit boards. Ah... past days. Ha ha! The non-circular edges give it a simple and rustic antique feel.

Since you are focusing on energy, not images, this concern about surface specular reflection seems to be more troublesome than its value.

Please be kind and respectful to help make the comment section great. (Comment Policy)

This website uses Akismet to reduce spam. Learn how to handle your comment data.

By using our website and services, you explicitly agree to the placement of our performance, functionality and advertising cookies. Learn more