Spectrum Analyzer Values for Non-Tech people

Status
Not open for further replies.

KD2JFA

Radio Guy
Premium Subscriber
Joined
Jul 18, 2015
Messages
113
Location
Northern NJ
Hey All, Quick question for you that I am attempting to ascertain some more information on. The agency I work for uses multiple antenna sites to receive radio transmissions from different agencies we work with. Ideally we would love to connect into their systems directly but that is not an option due to other issues.

So, take NYC for example, we have three radio reception sites (two in NYC proper and one in NJ near the bridge). Through these three sites we are able to receive all communications we need throughout the entirety of NYC. We are merely receiving and not transmitting whatsoever.

We are looking to move out to another area on the west coast, specifically San Francisco, and have acquired a USB Spectrum Analyzer from Signal Hound, some filters, antennas, a laptop, etc to go to potential sites to monitor channels to see how well they are received from locations. We should've done this in NYC as well but we got lucky with our previous method.

Anyways, onto my question, I understand the basics of how a Spectrum Analyzer works, I understand tuning into specific frequencies to see any interfering signals, and identifying "strong" vs "weak" signal sites. My issue is superiors want an actual numeric value to assign to these sites and documentation to prove which sites are better then others so we can decide where to go with. Unfortunately they have zero knowledge on radio systems, how they work, how radio frequencies actually transmit and receive, etc and Ive tried explaining multiple times but its not their forte.

What would you all recommend to use as a numeric value of interpretation that I can use in documentation that sticks out simply (with pictures from the SA software) that is easy for them to understand without getting too technical? Im a technical person and am just looking for some help, as well I do admit I don't know everything but would love to learn some more technical information myself to make me better at this. Thanks in advance.
 

mmckenna

I ♥ Ø
Joined
Jul 27, 2005
Messages
23,618
Location
Hiding in a coffee shop.
Decibels.

If you have a known value antenna, known value feed line and keep that the same across all your test locations, you can compare a signals dB level.
 

902

Member
Joined
Nov 7, 2003
Messages
2,620
Location
Downsouthsomewhere
Hey All, Quick question for you that I am attempting to ascertain some more information on. The agency I work for uses multiple antenna sites to receive radio transmissions from different agencies we work with. Ideally we would love to connect into their systems directly but that is not an option due to other issues.

So, take NYC for example, we have three radio reception sites (two in NYC proper and one in NJ near the bridge). Through these three sites we are able to receive all communications we need throughout the entirety of NYC. We are merely receiving and not transmitting whatsoever.

We are looking to move out to another area on the west coast, specifically San Francisco, and have acquired a USB Spectrum Analyzer from Signal Hound, some filters, antennas, a laptop, etc to go to potential sites to monitor channels to see how well they are received from locations. We should've done this in NYC as well but we got lucky with our previous method.

Anyways, onto my question, I understand the basics of how a Spectrum Analyzer works, I understand tuning into specific frequencies to see any interfering signals, and identifying "strong" vs "weak" signal sites. My issue is superiors want an actual numeric value to assign to these sites and documentation to prove which sites are better then others so we can decide where to go with. Unfortunately they have zero knowledge on radio systems, how they work, how radio frequencies actually transmit and receive, etc and Ive tried explaining multiple times but its not their forte.

What would you all recommend to use as a numeric value of interpretation that I can use in documentation that sticks out simply (with pictures from the SA software) that is easy for them to understand without getting too technical? Im a technical person and am just looking for some help, as well I do admit I don't know everything but would love to learn some more technical information myself to make me better at this. Thanks in advance.

You have a lot of variables to account for, first of which is the frequency range you're deploying in. "Agency" and across the country from where you've already deployed tells me you might have use to spectrum other than Part 90. That separation away from the "grass" and spikes you'll see matters. Do you use filtering (with or without preamplification)? What is your noise floor at the NYC sites?

Now, take that data and compare it to your survey site. Do analyses without filtering and with, comparing the deltas. Your captures will result in values that are more than likely in dBm, which is a power density measurement in decibels referenced to 1 mW. These values can be corroborated against predictive mapping using a modeling utility that calculates predicted values based on clutter and antenna directional patterns and can also do service area degradation based on adjacent channel degradation. How deep your noisefloor can be measured is often based on a number of factors depending on your equipment. For example, using a $150 SDR receiver might not be as precise as a $30,000 test instrument. Also, settings matter.

I hesitate to give specific values because of the differences above, but -120 dBm (or lower, remember these are negative numbers) is much weaker than -30 dBm. -30 dBm is much stronger than -120 dBm. If you go wide, you can see a bump that usually tracks with the passband of any filtering (or even antenna system). But it all depends on how you have things set up. Non-technical people see the spikes and the noisefloor. It's up to you to tell them how that looks compared to a known quantity, such as baseline data from your New York deployments.

If you're confident in your antenna system losses and gains, you can calculate the free-space loss to the other sites (referring to data in ULS or other database) and derive the effective radiated power of the emitter. Then you can apply TSB-88 methodology to seeing how that may affect your receiver performance. But that's not necessarily the only thing to look at.

The other thing to site selection is an intermod study. Here, you take co-located and signals within a certain proximity to the site, and run algebraic calculations to see if the signals fall within the passband of any of your receivers (and, if you're a good neighbor, the receivers of the other users - generally the last one on the site is responsible for cleaning up problems).

Good luck. Send me an email if you want to talk off-line.
 

prcguy

Member
Joined
Jun 30, 2006
Messages
15,235
Location
So Cal - Richardson, TX - Tewksbury, MA
Good information has been posted on weak vs strong signal levels in dBm but you also have to set up the RBW and VBW filtering, otherwise you may miss seeing some signals. RBW is the Resolution Band Width filter and is basically the IF filter of the spectrum analyzer. The wider the RBW setting is, the wider the range of frequency it lets in as it sweeps across the screen.

For example, if you only had one CW signal on the screen, which has a tiny amount of band width, the analyzer will show that one signal but a wide RBW setting like 3MHz will show that CW signal as 3MHz wide when its actually less than 1KHz wide. In this case the spectrum analyzer is basically tracing the response of its own internal IF filter and showing you what you asked for, even though its not correct..

So for looking at radio signals in a crowded band with channel spacing at 25KHz or less, you need to use an RBW setting narrower than the expected spacing between radio channels. As another example, if you have a strong signal and two weaker ones maybe 12.5KHz either side of the strong one, and you are looking at them with a 100KHz RBW filter, you will only see one signal and you will loose needed resolution to see the individual signals.

For scanning two way radio bands I would use an RBW filter no wider than 10KHz and something like 3KHz would be best to not miss any signals. The down side of this is most modern spectrum analyzers will slow down the sweep automatically when you use very narrow RBW or VBW filters. That coupled with trying to look at a wide spectrum, like 100MHz to 1GHz could take 30 seconds or more for the spectrum analyzer to make one sweep.

So, you need to know what you want to look at and understand what the capabilities of you spectrum analyzer are before you start so you don't run into problems. There is a lot more to a spectrum analyzer that has not been mentioned and I have a comprehensive training course from HP/Agilent that I can email the OP if he PMs me. I also used to do beginner and advanced spectrum analyzer training at my work, so I am a little familiar with their operation.
prcguy
 

paulears

Member
Joined
Oct 14, 2015
Messages
770
Location
Lowestoft - UK
This is a bit odd - the bosses want data that they clearly do not understand. dB's are a good example, where 3dB difference doesn't sound much when you have a maximum range of maybe 100dB or so - but of course, being logarithmic it does! Your analyser, with an antenna connected will be showing relative readings - one signal being XdB upon down on another, plus a signal strength relative to some kind of reference - probably X microVolts. Antenna gain, feeder loss and position are all going to produce figures. Producing sensible data will be intriguing!
 

prcguy

Member
Joined
Jun 30, 2006
Messages
15,235
Location
So Cal - Richardson, TX - Tewksbury, MA
You are right that producing sensible data by and for people who do not understand will be very intriguing if not downright daunting.

Its easy to produce good data with a spectrum analyzer if you know what your doing and understand the limits of the instrument. If you are designing a receiver distribution system with active amplifiers you also need to understand how to get aggregate levels so as not to overload any amplifiers in the system or create IMD. That is beyond beginner level spectrum analysis.
prcguy

This is a bit odd - the bosses want data that they clearly do not understand. dB's are a good example, where 3dB difference doesn't sound much when you have a maximum range of maybe 100dB or so - but of course, being logarithmic it does! Your analyser, with an antenna connected will be showing relative readings - one signal being XdB upon down on another, plus a signal strength relative to some kind of reference - probably X microVolts. Antenna gain, feeder loss and position are all going to produce figures. Producing sensible data will be intriguing!
 

mmckenna

I ♥ Ø
Joined
Jul 27, 2005
Messages
23,618
Location
Hiding in a coffee shop.
You are right that producing sensible data by and for people who do not understand will be very intriguing if not downright daunting.


Easy, management often likes colorful charts. A bunch of bar graphs embedded in overwhelming powerpoint presentations will easily overwhelm them with information and make them feel important.
 

jim202

Member
Joined
Mar 7, 2002
Messages
2,729
Location
New Orleans region
The agency I work for uses multiple antenna sites to receive radio transmissions from different agencies we work with. Ideally we would love to connect into their systems directly but that is not an option due to other issues.

So, take NYC for example, we have three radio reception sites (two in NYC proper and one in NJ near the bridge). Through these three sites we are able to receive all communications we need throughout the entirety of NYC. We are merely receiving and not transmitting whatsoever.

My issue is superiors want an actual numeric value to assign to these sites and documentation to prove which sites are better then others so we can decide where to go with.

Im a technical person and am just looking for some help, as well I do admit I don't know everything but would love to learn some more technical information myself to make me better at this. Thanks in advance.

Don't get me wrong here, but I would suggest that you start by doing an Intermod check of the site before going any further. This lets you have the computer digest all the frequencies at the site and it will spit out any conflicts with the frequency or frequencies your looking to use as receiver inputs. You don't even need to go to the site to do this. It does mean you need to work with the site manager to obtain all the CURRENT TX frequencies being used. I use the word "CURRENT" as in many cases the site records are not up to date.

Once you have run the intermod study, it will give you some feedback on what you may expect from the radios at the site. It also will show you any conflicts with any combination of the site transmitters. You may not be able to do this in real time because a specific combination of multiple transmitters may not happen while your looking at the spectrum analyzer. It could take a week or longer for just the specific combination of transmitters coming live at the same time to cause an intermod hit that could cause you problems.

Using the spectrum analyzer at the site will provide what the average noise floor looks like. What you find for the average noise floor will show you if your receivers can live with the noise there. It all depends on just what sensitivity your looking for the receiver to have.

I know this is not the answer you were looking for, but you have to look at it from a distance. You also have to know if the site is even feasible, based in an intermod study. This can save you hours if not days looking for problems that may not show up while your there.

It is much cheaper to do some homework before even turning the ignition key of your vehicle. Many of us have been in your shoes and I would suggest taking the smart route rather than spinning your wheels out in the field.

You may think the site is good. You go out and install your equipment and all the expense of the tower crew installing your antennas, interconnect back to your central point and the voting system. Then next week you find there is a problem now and then. So you get the gut feeling you missed something. Back to the site with the problem. You spend hours there and don't find anything.

Hope by now your starting to see the picture of how hard it is to locate an intermittent problem that probably could have been found using a computer program. Your going to do as you want, but take it from a person that has been through this before. Save the issue of standing in front of the management trying to explain why there is a problem with a site after spending all those hours and money installing the equipment and antennas.

I would rather have a pat on the back and hear "you did good".

Picking a tower has never been easy. It takes some skill and some expertise to do a good job the first time at a tower site for it to work out. It may even take installing some pass and notch cavities to be able to fit in. Don't forget if your installing a transmitter, you could cause a problem to others all ready at the site.

Another couple of suggestions. Use a good receiver that has decent shielding and a tight front end. most of the current receivers out on the market today have a front end that is as wide as several barn doors. This will not help in keeping other signals out of it at a site with many transmitters.

Don't use any of the LMR coax type cable transmission lines for your antennas. This cable may have good specs for signal loss, but typically will fail you at any high RF environment. This cable has 2 dissimilar metals used for it's shielding. The aluminum foil is the first shield around the center conductor. Then you get a copper braided shield. Over time the cable will get moisture inside the outer jacket. Once this happens, the 2 materials start to make diode type connections as corrosion sets in.

You will also have people saying as long as you weather seal up the outside connections really good, moisture will not get into the shields. Yup, have heard this many times. Yup, after a year or two, these same people are fighting noise on their receiver input. Why even put yourself into this position. Use a good quality cable that doesn't have this shielding problem that is well known.

Enough of the technical feedback. Let the group here know what you do and the results of your efforts. we all learn more from the feedback of others.
 

KD2JFA

Radio Guy
Premium Subscriber
Joined
Jul 18, 2015
Messages
113
Location
Northern NJ
Easy, management often likes colorful charts. A bunch of bar graphs embedded in overwhelming powerpoint presentations will easily overwhelm them with information and make them feel important.

I feel like this is going to be the option that I end up going with in terms of presenting them "data". Trust me, My face when I was told to present them "non-technical data" to help make the choice easier was hilarious I'm sure.


Also, Thank you all for the replies and Information. This is a great deal of information to absorb and Ive reached out to a few for more information. Ill let everyone know what happens, and If I have anymore questions.
 

mmckenna

I ♥ Ø
Joined
Jul 27, 2005
Messages
23,618
Location
Hiding in a coffee shop.
I feel like this is going to be the option that I end up going with in terms of presenting them "data". Trust me, My face when I was told to present them "non-technical data" to help make the choice easier was hilarious I'm sure.

As part of my job, I have to spend a considerable amount of time explaining complex technical issues to not technical people. I used to be not very good at it, and eventually with enough feedback I developed a pretty good set of skills at doing it.
- No one likes to be talked down to. "Dumbing" it down isn't the correct attitude, as the audience will know that. Simplifying the subject matter and avoiding the use of acronyms and technobabble is key.
- Drawings/charts work well.
- Photos work better. People like to see what they are getting.
- Answer questions in a non-technical way.

I've watched peoples eyes glaze over when presented with technical subject matter that is outside their grasp.
I've watched a lot of technical people that don't have these skills. I've had carefully choose people to do technical presentations/training to non-technical people.

It's not always easy, but you'll get much better feedback from the audience. In the end, that's what you are after.

Far too many people go into presentations with the "I'm going to show them how smart I am" attitude. It's never worked in my experience, and it's horribly painful to watch.

Sorry, sort of wandered off topic there. But, yeah, keep it simple. A bar graph can show differences in signal level without getting off into the technical weeds, so to speak. Trying to explain the logarithmic nature of Decibels won't work with non technical type. In fact, what works best is similar to the "bars" signal level display on a cell phone. That's pretty much generally understood by all.
 

RFI-EMI-GUY

Member
Joined
Dec 22, 2013
Messages
6,859
A bar graph is fine, but you need a reference level of what signal strength is acceptable. Otherwise the suits are going to want to know why you chose site x (most expensive) versus site y (cheapest), if all the signals show on the bar graph. An analog narrowband system versus a P25 system have different thresholds. You also need signal margin so that fades don't take the signal out. Then you have gains and losses associated with the antenna, feedline and multicoupler system. What about simulcast? Is a site in a problematic overlap zone? If so, how do you plan to mitigate the distortion?


Good luck explaining simulcast distortion to anyone who is a non technical manager.

Sent from my SM-T350 using Tapatalk
 

northzone

Member
Premium Subscriber
Joined
Dec 19, 2002
Messages
502
Location
Northern California
There are many people in the San Francisco area that have lots of local radio knowledge. You could state what you are trying to listen to and get feedback on what sites would work. This comes from lots of talent on here with many years of experience.

I had a remote site in the Hayward Hills that heard almost everything in the Bay Area.

You might want to tap into this knowledge base before running a much of measurements.

You can also listen to most agencies by way of live feed. I sometimes listen to 8 feeds at once.
 

zz0468

QRT
Banned
Joined
Feb 6, 2007
Messages
6,034
Anyways, onto my question, I understand the basics of how a Spectrum Analyzer works, I understand tuning into specific frequencies to see any interfering signals, and identifying "strong" vs "weak" signal sites. My issue is superiors want an actual numeric value to assign to these sites and documentation to prove which sites are better then others...

What would you all recommend to use as a numeric value of interpretation that I can use in documentation that sticks out simply (with pictures from the SA software) that is easy for them to understand without getting too technical?

There's some good sound advice on previous replies. But not a lot of it directly addresses your specific questions. You're tasked with making a lot of highly technical data digestible for non-technical (but assumed intelligent) people. Here's how I would tackle it...

First, identify all the specific signals/agencies that are needed to be monitored. List frequencies/systems/sites/talkgroups that you require the ability to listen to.

Try to understand the architecture of those systems, where the transmitters are located, etc.

Presumably, you have some proposed receiver sites identified. Try to plot out those locations along with the known transmitter locations.

With some work, it can be possible to predict expected signal levels from the transmitters to all the proposed receiver sites. Compile that data as best you can, and then you're ready to set out and make actual measurements.

Once you have made the measurements, you can lay the data out in a table for each site showing predicted and measured received signal levels. Just stick to received signal level in dBm. The data can speak for itself, and management can make their decisions based on raw data, not opinion.

That's how I would approach the problem. Hope that helps some.
 

paulears

Member
Joined
Oct 14, 2015
Messages
770
Location
Lowestoft - UK
I'm rapidly getting the impression that the management want data to enable decisions to be made, but seem to have a big problem. Firstly, they want data that they can understand, but they don't actually seem to have framed the essential question - what do they actually want, and then have parameters for collecting the data agreed. The poor person who is tasked with this doesn't have the full story to base his attempts to collect data on? Is there something preferred that the data is designed to prove/disprove? Is the real plan to consider alternatives - maybe an increase or decrease in sites as a result of complains or expenditure.

If you plot the polar diagrams around specific transmitter sites you can produce all kinds of useful data - but it needs to be filtered - what is it to be used for? Is there a need for a specific point in space to have data collected on signal strength in say, microvolts from specific sites, or is the idea to generate a map with red/green areas?

So much of our advice points to some of these kinds of data - but I'm not sure that meaningful data can be produced by anyone without some specific skills in this type of field. You can buy equipment, but accurate data relies on the skill of the technician collecting it. I saw in my own area a survey conducted with proper comms vehicles - their predictions were very wrong. Nobody told them that the results would be then used to produce a system using cars with magmount antennas - lower than the test vehicles with their antennas on a large roof. Moving to mag mounts in vehicle convenient positions, with distorted polar patterns produced a rather poor system with many blackspots. The survey company had not been told the purpose of their survey - just asked to produce, like this one possibly, a signal strength plot of an area.
 

jonwienke

More Info Coming Soon!
Joined
Jul 18, 2014
Messages
13,416
Location
VA
The only solution is for the schmuck collecting the data to have a conversation directly with the manager(s) to find out what they actually want to know andwhy tney want to know it, and figure out what data to collect that will actually address their question(s), and agree on some way of presenting the data that will be meaningful to the managers.

In large organizations, it is common for the employee to be tasked with something completely irrelevant to the decision the CEO is trying to make.
 
Status
Not open for further replies.
Top