Hacker News new | past | comments | ask | show | jobs | submit login
Inspired by aircraft IFF interrogator antennas, I decided to build my own array (techhub.social)
75 points by luu 12 months ago | hide | past | favorite | 12 comments



I had never looked into how IFF worked until this post left me with way more questions than answers. Like WTF is this even doing? I had very much incorrectly assumed IFF was a coded challenge-response setup, so this antenna array was confusing me. After reading the WikiP[0] article on it, it made a lot more sense about what is happening when the hand held device is used near the receiving antennas.

Yesterday, there was an article on signal reflections that spun me down a rabbit hole on wave reflections, 1/4 wave attenuators, dampeners, etc. Today, it's tuned antennas and playing with transceivers. Thanks to HN for helping fill my "learn 1 new thing a day"

[0] https://en.wikipedia.org/wiki/Identification_friend_or_foe


For anyone confused, "identification friend-or-foe" (IFF) is the military term for what's more generally called "secondary surveillance radar" (SSR) in the civilian aviation community. This is the backbone of modern air traffic control radar.

The Wikipedia article on SSR is a good read if this topic is new to you: https://en.wikipedia.org/wiki/Secondary_surveillance_radar

In the linked video, it looks like the poster is trying to build a small-scale beamformed directional antenna, which is somewhat related to larger larger phased array radar systems. See: https://en.wikipedia.org/wiki/Phased_array

> I had very much incorrectly assumed IFF was a coded challenge-response setup

It can be, for Mode 4/5 transponders used by the military. The civilian world uses Mode 3 transponders, which are unencrypted.

See: https://en.wikipedia.org/wiki/Aviation_transponder_interroga...


radartutorial is a great site to learn about this stuff IMHO. Here's their IFF section: https://www.radartutorial.eu/13.ssr/!sr01.en.html


thanks. it's a good resource, but maaaan, that website. there are some old websites that instill a feeling of nostalgia, then there are sites like this that make you frustrate to no end and make you so thankful modern UI has evolved. basic things like navigating to the next page should not be this annoying. the crap we put up with back in the day. at least it's not flash, but this has the feel of a site converted from flash with the same elements and emulates the flash site as close as possible


> but this has the feel of a site converted from flash with the same elements and emulates the flash site as close as possible

To me, this feels more like university lecture powerpoint, complete with white text on dark blue background.


maybe I should have tried hitting the next button on my presenter's remote?


Is there an explaination of what's going on in the video, or is it really just a two antena array with the fields interacting with each other and the panel?


The bit about IFF is kind of irrelevant. What the GIF is showing is a basic beamforming network where the two antennas are combined 180 degrees apart in phase, which result in a “difference” or “delta” pattern. This is seen in many more applications than IFF. Many antenna arrays use beamforming networks.


I don't see how it's irrelevant when it was the very thing that inspired the project in the first place. After all, I had no idea what I was looking at based on the video and brief description, but after reading about how IFF works, I understood what the video was showing

I also saw that it had the GIF tag, but is that really a GIF? If so, that's the best quality GIF and maybe the longest one I've ever seen. Something seems amiss with that tag


The two antennas on the desk are set up as an array (Input A and Input B). The beamformer (that little disc) creates two outputs, a sum channel (A+B) and a delta channel (A-B). Where the -B part is essentially inverting the phase of B and adding that to A. The delta channel is plugged into the network analyzer for display (or oscilloscope -- I didn't pay too much attention as to which it was).

The diagram to the left shows the theoretical beam patterns of the output: the orange is the sum channel and the blue is the delta channel. The delta channel has a deep null where the sum channel has a peak.

The poster turns on a transmitting antenna and sweeps it in angle past the array. The output is very low when the transmitter is directly between the receive antennas -- as expected (ideally it would be zero, but any phase or amplitude mismatches between the two antennas would reduce the depth of the null).

The first pass he makes doesn't show as large of a dip as expected -- probably because he's in the near-field of the array. You'd typically want to be in the far field and there's a rule of thumb based on the size of the array as well as the wavelength. He moves the transmitter back a bit on subsequent passes and it shows a bit deeper of a null.

I would call this a "poor man's monopulse" (monopulse being a keyword if you'd like to search further). If you have, say, a very large array and you're using it for radar, you would often apply some sort of amplitude and phase weighting across the receive elements on the "sum" channel to control sidelobes. Otherwise, strong signals coming in from angles not associated with the mainbeam can be mistaken for signals in the mainbeam. You would ideally match your delta channel weighting to your sum channel weighting.

Monopulse in radar is used to improve angle estimation. Radar is very good at ranging: you transmit a signal and time how long it takes until you receive it and the speed of light is a well-known quantity. But the cross-range error is poor: you want a very large antenna to give you a very fine beam and to translate that to cross-range error. For example, you may easily be able to range something to within 10 meters, but with a 2 degree beam width at 20km, your cross-range error would be on the order of 750m. Monopulse, for nominal signal-to-noise ratios, can give you angle accuracy roughly 10x better than real-beam -- so in our example you might have 75m cross-range error with monopulse. Still not great, but much better than 750m.

So you'd perform your detection on the sum beam, and then you'd essentially look at the ratio of delta / sum to get a better estimate of the angle within the mainbeam.

Note: I'm not an antenna engineer, but I do make use of antennas in my field.


to me, it shows the system has been tuned so that when the hand held unit is operational that the signal is generating the spike on the scope. i'm not sure what the power output is that it needs to be so close, but it's a good PoC of understanding of the concept and being able to build a functioning system.


Yeah, I think that's all there is here.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: