TV Fool Discussion Thread  

Several issues have been reported regarding the new experimental site at http://www.tvfool.com/ and the signal analysis tools within, so I'm starting a new thread specifically to address those issues.

First of all, thanks everyone, for all the useful feedback. Everything on the site is quite new and it certainly still has some rough spots. I'll do my best to address the issues reported here and hopefully we'll end up with something that everyone can use.

I'm mostly staying focused on the coverage map and radar plot software development, although any web related feedback is also welcome. I can forward any web-hosting related issues to the friend that is helping me host the site. BTW, he says "Thanks!" for all the encouraging comments that people have been sending in.



Recent changes:

16-Dec-2007
- Added online coverage map browsing feature.

17-May-2007
- A brief synopsis of the search criteria used in the analysis is now included on the plot. The exact addresses and coordinates are still obfuscated for privacy reasons, but there should be enough rough information (e.g., zip code) to get an approximate idea of where the analysis is focused.
- Custom titles are now possible. If you want to include a brief description or keyword to help you remember what the report is about, now you can enter your own text to be included on the report.
- Azimuths are now color coded. This should make it easier to quickly visualize which transmitters are clustered together. A color ring has been added to the outside of the radar plot to correspond with the azimuth values.

13-May-2007
- The reports now include the network affiliation of most broadcasters. This should help avoid the hassle of having to reference other sites to get the network identities for each callsign.
- Several recommeneded formatting changes have been added. The most notable are 1) decreased radar plot size, 2) ressurection of transmitter power column, 3) expansion of the transmitter details table, 4) general formatting improvements for clarity and easier reading, 5) graph limits to prevent very strong signals from drawing over the text area.


6-May-2007
- The radar plot tool has been changed to now include a graph of channel vs. signal strength. This feature is designed to show the distinction between VHF and UHF channels and hopefully make it more obvious what kind of antenna is needed to receive each channel. This is still experimental, so any feedback on this feature is greatly appreciated.
- The HTTP referrer check has been removed. This should allow the tool to work even for those people that have disabled referrer reporting in their browsers.
- All distance and height values have been changed to miles/feet.


2-May-2007
- The FCC data extraction has been changed to a less aggressive approach. The previous approach would try to pick the most up-to-date FCC records, but in many cases, this ended up picking records for "future" (not built yet) transmitters. The new approach will hopefully do a better job of picking the transmitters that are really installed and operational today. As always, this can be improved further by individual "overrides" and feedback from all of you, so if you know of any entries that need to be corrected, please let me know.
- Another category of typos in the FCC database has been detected and fixed. This caused a small number of transmitters to be seen as having zero transmit power. Those records should now show up correctly.


29-Apr-2007
- Several people have reported formatting and screen cut-off problems in Internet Explorer 6 and a few other browsers. The site's style sheets have been changed to hopefully get the formatting right on a broader range of browsers. If you were seeing problems before, you might have better results now.


23-Apr-2007
- The coverage map processing has been increasingly automated. This makes the map generation go faster and allows me to release more metros at a time. For anyone still waiting for their metro to show up, at least now your wait won't be quite so long!

Upcoming changes:

- Maybe add the option for higher resolution (larger print) reports.

Everyone has been providing great suggestions, so please keep 'em coming!



Best regards,
Andy

Thanks for the feedback!

The invalid referrer check has been removed, so hopefully things will work for you now.

The purpose of such as check is actually to prevent abuse of the URL from third-party "unscrupulous" web hosts creating direct links into the tool interface. I will leave the check disabled unless there's a significant rise in server abuse.




The choice to use dBm over dBuV is mostly for ease of use. The conversion between the two is trivial, so from a processing point of view they are interchangable.

Field strength (and power flux density) is more useful when it comes to antenna and waveguide analysis/testing. It is the preferred unit for broadcast engineers and facility managers because it's the easy to measure and understand in the scope of what they're doing. Power is more useful when it comes to analyzing individual tranmitter and receiver devices (e.g., silicon devices, tuners, power amplifiers, LNAs, filters, etc.). As for propagation modeling, the units actually doesn't matter, because a dB of loss in dBuV is also a dB of loss in dBm.

Since I mostly deal with designing and testing devices (e.g., cell phones, GPS, TV, etc.) rather than the antenna systems, I find it more convenient to deal with power units. Most of the lab equipment (e.g., spectrum analyzers) and other devices that I deal with provide specifications in terms of power. With everything specified in terms of power, it's easier to estimate things like heat dissipation (for transmission paths), SNR relative to thermal noise floor (in receive paths), and device signal handling capacity.

The relationship between dBm and dBuV is , so you're right that dBm does vary with impedance, but for consumer TVs, we mostly just care about the 75 ohm case (btw, there's only a 1.8 dB difference between 50 and 75 ohm systems).

In a non-engineering sense, I also think that power means more to the average person than field strength. Most of us have an instinctive feel for what it means to transmit 5 megawatts of power, but don't have a good feel for what 41 dBu means.

Best regards,
Andy

OK, all the units in the radar plot tool have been changed to miles and feet. Hopefully, this will cause less confusion for everybody.

Best regards,
Andy

D'OH! I forgot to change that. Thanks for catching this. It's been fixed now. I guess that's what I get for coding late at night.

Best regards,
Andy

Yep. I originally developed my radar plot tool because I always felt that antennaweb was dumbed-down too much. Colored wedge charts give very little real information, and I'm not sure that it's helping make the consumer's life any easier. As an electrical engineer, I felt I needed more detailed information about each signal to really understand how to optimize reception and antenna selection. When I started generating plots to help answer other avs members' questions, people seemed pretty interested in the analysis tool, so I started work on making an online version available.




I know BitTorrent is new to a lot of people. The Google Earth files are quite large and demand for them is fairly high, so torrents were the only way I could distribute the files with reasonable speed given the limited bandwidth of my network connection. BitTorrent is much more efficient with bandwidth utilization because it aggregates shared bandwidth from many servers simultaneously. Pieces of each file can be downloaded from multiple sources at the same time until it assembles a whole copy of the original. If many people are downloading the same file, they can also help each other because each person will download some pieces from my server and also get some pieces from other peers. This way, everybody shares a fraction of the bandwidth load and everybody ends up getting the file faster.

I am searching for better hosting facilities that have enough storage and bandwidth to allow for direct dowloads, but I don't expect anything to be ready for quite a while. The collection of files is already several gigabytes and I still have a lot of metros to add.

Given my current setup, BitTorrent is the only viable option for now. The good news is that the latest BitTorrent clients have become quite good. For Windows users, uTorrent is nice because it does not require any installation (it's just one executable), it's very small and efficient, and it's rich with features. Azureus runs on any platform because it's Java based, and is also rich with features.

It takes some time to warm up to BitTorrent, but once you get the feel for it, it's not that hard to use. It lets you download many large files unattended (just pick them up when they're done) and the improved network efficiency is a nice side benefit.

Best regards,
Andy

Thanks! These are fantastic suggestions. This is definitely a work in progress, so it's the perfect time to make changes that improve the software. This started as a very technical engineering tool I used for personal reference, so I know I need to clean up the UI for general consumption. I like your ideas a lot.

I think I will end up removing the LOS_h and Min_h columns altogether.

I also think it's safe to shrink the radar plot, leaving more space for the table, and keeping the layout roughly the same.

FYI, there's a subtle requirement of trying to make everything fit within an 800x800 image. That's because most vBulletin based forums restrict image uploads to that size. My hope was to make it easy for people to upload their plots when posting to forums like this when asking for help. By encapsulating all the really useful information in a single image attached to a post, it makes it easier for people like me to respond to people's questions. If this constraint begins to be a problem, then I have no problem with giving up on this idea.

My philosophy is 1) make it work, 2) make it work well, and 3) make it faster/cheaper/better.

Best regards,
Andy

Thanks for the pointers! Let me expand on these one at a time:


The levels are actually only about 10 dB too low. That's because I'm using F(99,99) parameters in the Longley-Rice modeling. Those are really worst case conservative conditions, so it definitely predicts power levels lower than the average. I will probably shift all the values up by 10 dB to bring things in line with reality (based on what I've measured in field tests).

I suspect the other 20 dB difference you're seeing is probably because of your receiver setup. Keep in mind that the L-R modeling is predicting the field strength "in the air". If you are measuring something on a meter or spectrum analyzer, you're actually including gain/loss from your antenna, amps, cables, and connectors.

I don't expect anyone to actually be able to reproduce the Rx_dBm numbers via measurement because we generally don't work with perfectly tuned dipoles or isotropic antennas. The data is supposed to show what raw signal you have to start with, and then it's up to individual selection of antennas, amps, cables, and receivers to make the most out of what's there.




I think I will need to place an upper bound on the bar lengths / positions on both the radar and spectrum plots. This will prevent the graphics from running into the text.




I will be experimenting with expanding the table. It's probably the most informative part of the analysis and deserves the most attention. You may see several variations over the next few days while I try out different ideas.



These are all excellent observations and I'll try my best to make things better!


Thanks a bunch.


Best regards,
Andy

Thanks!

Yep. I'm quite familiar with the FCC TV guidelines. They've already added fudge factors to account for building penetration losses, antenna gain, and receiver sensitivity to come up with these numbers. However, as many avsforum members know, every situation is different, so sometimes those assumptions can be way off.

It's easy to be critical of the FCC, but I also have to give them credit. Afterall, if they hadn't made the transmitter database available for public download, I wouldn't be able to make any of these tools. We also have to keep in mind that a lot of these guidelines are based on work originally started in the 1950s and 60s. It's only recently that computer processing power has become good enough to really analyze these signals in detail.

BTW, I use F(99,99) modeling parameters instead of F(50,90) because it's generally known that the FCC parameters are a bit too generous (by roughly 10 dB, it seems). By design, my modeling results will show more pessimistic numbers than the FCC model, but in my experience, this is a better match for real-world viewing experience.

Best regards,
Andy

I think I will add the option to enter a user specified title to the top of the plot. I purposely left out any user-identifiable information in the plots because most people would prefer to keep their address and/or coordinates private. I think a short text field for a title will let people label each plot however they please (e.g., "Dave's House 30ft Mast") without giving away any secrets.




I'm not entirely clear on what you mean by this, but I THINK you're saying that if there's an error in the submitted form and the form re-appears partially filled-in, then the value in the recieve height field is filled in with the meters value instead of the originally entered feet.

If I'm reading this right, then I suspect you're exactly right. I had not accounted for this case. I'll need to clean this up in a better way the makes sure the metric stuff doesn't get intermingled with the user entered data. I should have this fixed by tomorrow.

Nice catch, thanks!


Best regards,
Andy

Comments are always welcome!

The radar plot serves a few different purposes, actually. The main purpose is for visualization of transmitter directions. This makes it easier to spot the need for omni antennas, directional antennas, rotators, or some combination of the above.

If the antenna radiation pattern (polar plot) is known, we can imagine that pattern overlaid on the radar plot to roughly know whether the beam width is too wide/narrow to capture signals from multiple tower clusters.

For anyone installing an antenna, it also doesn't hurt to have the radar plot on hand to make sure everything is aimed correctly. I know that sometimes I've mis-calculated my azumuths and ended up installing antenna "trees" that didn't match my plot. Once I see that I've made a mistake, I can go through my caclulations again and see what I did wrong.



I agree that the radar plot isn't super important relative to the screen real-estate that it takes up, so I do plan on shrinking its size. That should create a lot more space to make the table bigger.




Both the radar plot and the spectrum graph have limits on their scale. The bulls eye of the radar plot is currently at -110 dBm, so it won't show any signals below that. The bottom edge of the spectrum graph is -120 dBm, so it will also cut off some transmitters. These cutoff values are generally very low, so the majority of people will not be able to receive signals below these levels anyway. The table tries to list as much information as possible within the space allowed, down to about -130 dBm. This goes well beyond the reach of most people, so the weakest transmitters shown will not have much value except maybe for DXers.




I will be removing these columns soon since they don't help much and they created a lot of confusion. However, since you asked...

The LOS_h is the height where you'd start getting a line-of-sight path to the transmitter. That's the height where the path is just at the grazing edge of clearing the obstruction.

The Min_h is the height necessary to get a signal level of at least -100 dBm. Because of the way diffraction works, the signal strength at the grazing edge path is actually lower than the signal at slightly higher or lower elevations. It is actually possible to get a stronger signal below the grazing edge path than at the grazing edge path. The search algorithm to find this minimum height probably homed in on the local-maxima just above the grazing edge path to reach the -100 dBm cutoff. If this transmitter was any further away, it probably would have found no height sufficient to reach -100 dBm.

It's just one of those interesting "spooky" thinks about EM waves, but this information is hardly any use to people using the tools. The original idea for this column was to help people decide how high they should build a mast to get a channel that's just slighly out of reach. It will be removed soon.

I might want to put in a column for distance to nearest terrain obstacle or angle/severity of diffraction, but I'll have to think about this some more.




Yes. If you let me know what the real transmitter is broadcasting, I will add that info to my collection of FCC database "overrides". The FCC database has all kinds of errors, missing data, and superfluous data. My software does its best to filter the database to find the records that best match reality, but there's still the need to manually put in patches for some of the transmitters. The good news is that once I enter the "override" in my database, it fixes all future generated output (including coverage maps, radar plots, and the transmitter icons collection). The overrides stay in effect even after I download new copies of the FCC database.

I don't have an automated way to do this yet. I think the only way to protect the database from abuse would be to have a community voting mechanism or a network of moderators, but that's not something I'm prepared to set up right now.




Those indicators show the dominant path of the signal between the transmitter and your location. 1- and 2- Edge stand for single- or double- edge diffraction, which means that the transmission path is most like the bending of waves over one or two edges (terrain high-points, in this case). Diffraction is very common and a lot of people get strong TV signals even if they're slighly blocked by 1 or 2 ridges. Line-of-sight is not necessary for good TV reception, but diffraction can weaken signals significantly, depending on how sharp the angle of diffraction is.

Increasing antenna height for a diffracted signal can improve things significantly (depends on distance to obstruction and wavelength of signal), whereas a LOS signal will hardly improve at all with added height.



These are all great topics, and I hope my answers helped! Please keep all the good ideas and feedback coming.



Best regards,
Andy

Well, it looks like those last two columns are popular after all!

I will hold off making any changes to those columns for now. I will first experiment with shrinking the radar plot and increasing the table size. I'll try to clean up the table formatting, but will keep the columns.

Thanks everyone for the comments!



Best regards,
Andy