RF output power for ISM bands

[rx8pilot]

This is a general question about how to measure RF output power of various ISM band devices.

Purpose: The goal is to make relatively ballpark measurements of a variety of ISM band devices operating in a dense environment to see how close they are to the maximum allowed transmit power. Most of the time, there cannot be a direct connection to the transmitter so it would be a radiated measurement of some sort. This kind of thing is well outside my comfort zone but hoping to learn more since I am planning to incorporate RF transceivers into some of my products this year. The industry that I am in (film/tv) is overrun with RF for just about everything from WiFi to lighting control to very high-speed video links. Now that there are a variety of brands from unknown manufacturers claiming better range and stability - many of the known brands are starting to fail far more often when used in the same location. One theory is that the new brands are simply blasting a RF beyond 30dBm TX or using antenna gains for EIRP beyond 36dBm. I want to do some on-site non-invasive power measurements to gauge how much manufacturers are pushing the legal limits of tx power. I have no expectation of being metrology grade at all, just want to see what it takes in terms of equipment and knowledge to make these measurements of actual transmissions.

EDIT:
I am not looking for the cheapest DIY option possible, but I do not have an unlimited budget either for equipment.

[Gribo]

For the equipment, you need a spectrum analyzer and reference antenna. For the environment, you need a quiet area, preferably a rural area without too much background noise.

For the test methods, follow the relevant FCC regulations (Part 15.245 etc). They specify distances, limits, antenna position, and orientation.
Take into account that you really can't use this data and complain to the FCC, but it is a start. 

Mind you, it is not only output power, but also power spectral density, channel hopping, and duty cycles.

[rx8pilot]

For the equipment, you need a spectrum analyzer and reference antenna. For the environment, you need a quiet area, preferably a rural area without too much background noise.

For the test methods, follow the relevant FCC regulations (Part 15.245 etc). They specify distances, limits, antenna position, and orientation.
Take into account that you really can't use this data and complain to the FCC, but it is a start. 

Mind you, it is not only output power, but also power spectral density, channel hopping, and duty cycles.

Thank you for the response - 

I have no expectations to file a complaint based on my rough measurements - only looking for a ballpark idea before going further. Agreed that output power is only the tip of the iceberg. The next hypothesis is channel hogging is being used to bump conflicting traffic. Of course, that analysis will likely require a professional that is equipped and skilled in the art (maybe it's easier than I think). Even if I had the gear, I do not have enough time to self-educate and be confident in any resulting conclusions. My intent is not to throw anyone under the bus, but rather make an informed decision about whether or not to get into that market from a business perspective. If I have to compete with products that are willing to break the rules - I may just consider something totally different. Not sure it is even big enough of a market for the FCC to care, they are too busy investigating comedians language. Launching a new product is difficult under the best circumstances - dealing with shady competition makes it considerably harder.

[Lord of nothing]

so for outsite of the us in the most country the official dont care about the output powert so much until something happen like some hobbyist measure the outputpowert or the device inter fear something.

Now that there are a variety of brands from unknown manufacturers claiming better range and stability - many of the known brands are starting to fail far more often when used in the same location.

hmm the question is its possible to enhance anythink?
Like Wifi some company claim with 360degree Antenna Setup and using refection the can enhance the Signal.

When I look on PMR446 device some cost +100€ per device and the get use on Band where many people use the cheap Baofeng who are illegal with an "huge" Output power.....  So what to do? Use some illegal Radio to with an huge output power or use "good quality" Radios with the legal output power but who cant use because the cheap interfear.

[rx8pilot]

Here is one example -

http://www.amimon.com/broadcast-production-market/sky-link-family/

The claim is 3Gbps data 1000 meters on 5Ghz ISM - 3 Gigabits per second seems unlikely without some compression. 1000 meters seems like a generous estimate of range. Even 1024QAM would struggle to do that data rate with any sort of reliability right? Maybe some new fangled modulation?

[dmills]

3Gb/s, so that will be SDI then.

The range at 5GHz is not the issue, that is trivial, you just chuck a big dish on the RX, been doing that between hilltops with analogue PAL since the klystron days.
I would have thought that the 5GHz band is a little on the low side, but 3G-SDI is reasonable at 24GHz.

Lets see, if they do not have significant latency which they claim, then there is no channel coding happening, the 5GHz band extends from 5.725 - 5.875GHz so is 150MHz wide meaning the bare minimum SNR for the Shannon-hartley limit is  3Gb / 150MHz = 20 =  log_2 (1+S/N). 2^20 = 1 +S/N (Where S/N is a power ratio) = ~10^6 = 60dB.

In a 150MHz bandwidth the thermal noise floor is around 92dBm so theoretically we would be looking for -30dBm at the receiver (If the channel efficiency was 100%!), and the preamp had no noise contribution, a huge ask.

Now if they play fast and loose with the specs and go for say 300Mhz instead then 2^10 = ~30dB SNR  a little more reasonable, but anything else in that band is still going to stuff things up pretty hard.

Radio marketing guff always comes up with range estimates that assume free space with no interference, no rain and no Fresnel losses, you can generally divide by at least 10.

I have to admit to not quite understanding the use case here, put the recorder on the camera already, and if need be cart a 270Mb SD  stream over the air for preview, it will be orders of magnitude more reliable.

The production market values reliability over all things, and a length of Belden with 75 ohm BNCs, or a deployable fiber with expanded beam connectors from Lemo  is reliability that is hard to beat. I would be very cautious of selling RF in the ISM bands into that space, something with narrow front end filters in a LICENSED band is usually not a big deal to those guys.

73 Dan (Who plays SDI for work).

[rx8pilot]

3Gb/s, so that will be SDI then.

The range at 5GHz is not the issue, that is trivial, you just chuck a big dish on the RX, been doing that between hilltops with analogue PAL since the klystron days.
I would have thought that the 5GHz band is a little on the low side, but 3G-SDI is reasonable at 24GHz.

Lets see, if they do not have significant latency which they claim, then there is no channel coding happening, the 5GHz band extends from 5.725 - 5.875GHz so is 150MHz wide meaning the bare minimum SNR for the Shannon-hartley limit is  3Gb / 150MHz = 20 =  log_2 (1+S/N). 2^20 = 1 +S/N (Where S/N is a power ratio) = ~10^6 = 60dB.

In a 150MHz bandwidth the thermal noise floor is around 92dBm so theoretically we would be looking for -30dBm at the receiver (If the channel efficiency was 100%!), and the preamp had no noise contribution, a huge ask.

Now if they play fast and loose with the specs and go for say 300Mhz instead then 2^10 = ~30dB SNR  a little more reasonable, but anything else in that band is still going to stuff things up pretty hard.

Radio marketing guff always comes up with range estimates that assume free space with no interference, no rain and no Fresnel losses, you can generally divide by at least 10.

I have to admit to not quite understanding the use case here, put the recorder on the camera already, and if need be cart a 270Mb SD  stream over the air for preview, it will be orders of magnitude more reliable.

The production market values reliability over all things, and a length of Belden with 75 ohm BNCs, or a deployable fiber with expanded beam connectors from Lemo  is reliability that is hard to beat. I would be very cautious of selling RF in the ISM bands into that space, something with narrow front end filters in a LICENSED band is usually not a big deal to those guys.

73 Dan (Who plays SDI for work).

Pretty much all of that :-)


Most of these systems claim to be able to put 4+ systems on at the same time - 12Gbps total plus a control channel. It is really a stunning amount of data to stuff into such a narrow band. So far, I have not found anything that hints at any kind of line compression or maybe line groups to keep the delay to a minimum and reduce the bitrate a little. I am not familiar enough with digitally modulated RF to have much of an opinion - only that this feels impossible without some secret compromises.

As for the use case, it is a situation that is dominated by non-technical creative people. They believe that RF works by magic fairies and unicorns. It only makes sense that the image coming from a low-cost RF system should be the same as a hardwired signal. Of course in reality, RF video (especially the lower-end <$5k) is a fiddly mess that destroys the 5Ghz ISM band for all other uses. The cameras all have 5Ghz WiFi for command and control, lens controls are 5Ghz, digital wireless mics, a myriad of computers on 5Ghz Wifi etc.

Sometimes a cable is impossible when shooting car to car or in very difficult locations. I see crews using wireless pretty much all the time even when they don't need it and struggling with drop outs and other devices o the set. Everything that can possibly be wireless is not wireless whether it needs to be or not. It is almost funny to watch the struggle unfold sometimes. Just plug in a cable and all your issues go away, lol.

side note: I don't think Lemo ever made expanded beam fiber connectors. I used Tyco expanded beams for years and they were fantastic. Lemo made the hybrid cables that had 2 fibers and copper power so you can power a camera from 1000' away. Those systems saw very little use in film and TV shows relative to broadcast shows.

[dmills]

I thought it was Lemo, may have been Fisher, one of those guys that used to do the Triax camera stuff anyway, and yea expanded beam is magic.

Looks like the film industry is more divergent from broadcast then I at first thought, in broadcast the guy from the uplink truck (Who is usually technical) would have provided some fairly sweary education on the subject of the realities very, very quickly, and he has a spectrum analyser available to back it up.
It is strange because you would think that the cost of a day on set would be high enough to make doing it properly cost effective just is savings on having expensive talent standing around while people muck with intermittent faults.

For car to car stuff, send the timecode (It fits over an audio channel), not the video!

For fun, consider that sooner or later someone is going to try to stuff quad link UHD over 4 of these things, that would be amusing from a safe distance.

There are things that can only be done wireless, but the theatre industry has a saying that "A $5,000  radio system works nearly as well, is nearly as reliable and sounds nearly as good as a $50 cable", words to live by.   

Is it wrong of me to respond to seeing a film shoot in my street by firing up the SDR and trying to hack the camera controls?

73 Dan.

[CJay]

Would removing detachable antennae count as invasive?

If not then perhaps a simple wideband power detector in series would do just what you need instead of a full blown spectrum analyser?

Would also probably be usable with a 1/4 wave antenna and no direct connection to the equipment but you do need a 'quiet' area to acquaint the detector with the offending equipment so as to obtain fair measurements.

A spectrum analyser would of course also tell you if the equipment was 'dirty' and what bandwidth it's using which could also be part of the problem.

Though if there's budget for test gear...

[dmills]

The US limits allow 1W transmitter power output and 4 watts EIRP (More under other rules for a fixed point to point link), other countries are likely similar.

Free space path loss at 5.8GHz for a 1km path is ~108dB.

So assuming a somewhat directional transmit antenna having 6dB of gain, we have 6 - 108 = -106dBW at the receiver antenna, +30dB to convert to dBm + whatever antenna gain we have, = -76dBm at the receiver + antenna gain.

noise floor in a 150MHz bandwidth is ~-92dBm, leaving only 16dB SNR at the rx, so the antenna will need to be both very high gain and very low noise temperature, to even theoretically be able to pull this off Shannon-Hartley says we need a 60dB SNR to fit 3Gb/s into 150Mhz of bandwidth, and I would want 20dB or so of margin.

Someone is not playing by the rules.

[Gribo]

They specify 20 and 40MHz channel bandwidth, so they must be using some compression.