The KA7OEI FT-817 pages
Repeater coverage analysis using the FT-817 and GPS

Front panel view of my FT-817
A Front-panel view of my FT-817. 


The FT-817 includes the capability to read the S meter via the serial port.  This article describes how an FT-817 can be used to accurately measure received signal strength over a range of more than 50 db on multiple bands and how that information may be used to characterize the coverage area and performance of a repeater's transmit antenna.

Several tasks required of the '817 and the program:

Definition of the problem:

The '76 repeater has a problem.

It didn't always have this problem:  In about 1976, when the transmit and receive antennas were originally installed coverage was excellent in all directions, limited only by geography.  Coverage to the south was a bit limited, owing to the mountain range, but that was okay because there was almost no population in that direction.  A web page about the 146.76 repeater may be found here. The page has pictures and information that can help explain the problem.

In 1997 we had to replace the antennas.  This was not because there was obviously anything wrong with the original antennas - except that they were on the wrong tower.

Let me explain:  The original tower (the one on which the antennas were installed in 1976) was a guyed tower, approximately 100 feet high, and it was scheduled to come down within a year.  It seems that it had outgrown its usefulness, not being up to the task of holding the new antennas expected to be required in the future.  Also, a high-powered UHF TV transmitter was going on the air from a nearby site:  With the over 2 megawatts of ERP, it seemed prudent to replace everything and start from a clean slate to avoid IMD problems resulting from old, corroded hardware.

So, we had to put antennas on the new tower.  We decided to replace the antennas as well - considering that they were the original ones - still in service after 20 years.  Anyway, installing new antennas/feedline before removing the old antennas allowed nearly seamless operation of the repeater during the transition.
View of the '76 transmit antenna
View of the 146.76 repeater's transmit antenna (it is the white vertical farthest to the left.)  The antenna, being amongst other antennas and dishes - as well as being on the south face of the tower - accounts for its deep null to the north.
Click on the picture for a larger version

The '76 repeater has always used separate transmit and receive antennas:  Doing so greatly reduces the isolation requirements of the duplexer.  It also goes a long way toward improving effective repeater sensitivity as not only are duplexer losses reduced, but having the receiver antenna located away from the transmit antenna (or even other users' transmit antennas for that matter) can significantly lower the receiver's noise floor by removing broadband noise sources and potential sources of intermod. 
To this end, the new receive antenna is located at the top of the tower, its base being at the 120 foot level while the transmit antenna was located amongst other transmit antennas at the 60 foot level.

Initial tests concluded that the new receive antenna/location worked at least 3-4 db better than its predecessor.  The transmit antenna seemed to work at least as well the original...  or so we thought until we started getting reports from other directions.

Toward the East the transmit signals seemed to work every bit as well as it had previously.  To the North, however, (in the precise direction of the Salt Lake valley) things seemed to be different:  Some people reported it as being "about the same" while others could hardly hear it at all.

In the years since the new antennas were installed the picture has become clearer:  In fact, to directions from the Northeast through the Southwest the signals from '76 seemed to be as good or better than before.  Directly to the North (in the western side of the Salt Lake valley) the signals were definitely pretty poor:  One could easily get into the repeater using an HT, but one couldn't hear it very well.  Even mobile stations complained a bit about the weak signals there.

It was time to measure the signals to find out exactly what was going on.

When one does measuring of signal strength from a repeater, there is the time-tested method which consists of these basic steps:

  1. Find a clear location for measurement.
  2. Find and note the measurement location on a map.
  3. Determine distance and bearing to the repeater.
  4. Measure signal strength.
  5. Put the equipment away.
  6. Repeat steps 1-5 many more times.  When you are all done, you then have a huge volume of data to sift through and make calculations.
It need not be said that making signal measurements in this manner is very time-consuming.  One must make a decision as to how many data points one might collect - a factor that not only affects now long it will take to make these measurements, but also the quality of the data.  There is also the problem that some individual data points may not produce "good data" - that is, a signal reading at a particular location may be affected by some local feature - something that may not be immediately obvious at the site.  Furthermore, one may have to choose sites based on their accessibility rather than their suitability of signal readings.

Flash Forward to the 21st century:
signal strength
signal strength
0 <-108 <-114.5
1 >-108.3 >-113.8
2 >-107.3 >-113.0
3 >-106.7 >-111.6
4 >-106.0 >-110.2
5 >-105.1 >-108.8
6 >-104.2 >-106.4
7 >-103.0 >-104.7
8 >-100.4 >-102.8
9 >-84 >-101.0
A >-74.5 >-99.5
B >-70.1 >-97.8
C >-58.9 >-96.8
D >-50.8 >-95.8
E >-40.8 >-94.6
F >-30.1 >-93.5
Table 1
Serial-port S-Meter readings versus signal input (as read via the serial port) on 2 meters for my FT-817 as shipped from the factory.

Nowadays, this sort of data collection can be nearly automatic.  Several commonly-available pieces of equipment can be utilized:

Putting all of these pieces together is probably the biggest challenge: Assembling the pieces:

For this exercise, I used a modest laptop:  A 75 MHz '486.  While this machine wasn't a speed demon, it was more than fast enough to do the task.  There was one problem:  It had only one serial port - and two were needed:  One for the GPS receiver, and another for the '817.  While this wasn't a "show stopper" it was a complication.  Since adding another serial port to this old laptop really wasn't an option, I decided to "multiplex" the two using one of the serial port's handshaking lines and a relay:  A few readings are grabbed from the FT-817 and then the relay switches over to the GPS receiver.  When complete NMEA sentences are obtained, the relay is switched back to the '817.  While this solution isn't pretty, it works.

I already had the GPS receiver:  An old Magellan GPS4000 (non-XL version) with the serial interface.  My software had to extract the NMEA sentences and parse them to get the data I needed to determine my location, the current time, quality of the data, and the bearing and distance data to the waypoint - the repeater in our case.

I also had the '817 and the official CAT-62 interface cable - so I was able to read the "S-Meter" from the '817.  I quickly wrote a simple program to read this data and sat down to make use of this data.
0 <-110.7   8 -94.2
1 -108.9   9 -91.5
2 -106.2   A -89.2
3 -104.2   B -87.2
4 -102.3   C -85.2
5 -100.6   D -82.1
6 -98.9   E -78.1
7 -96.7   F >-75.7
Table 2
Serial-port S-Meter readings versus signal input using FM mode (as read via the serial port) after the described recalibration of the FM-S1 and FM-FS parameters. 

I soon realized that I had a problem...

Making the FT-817 S-Meter data useful:

There are two completely separate S-meter circuits available on amateur bands from the '817:  The AGC-based S-meter available in the SSB/CW/AM modes and the RSSI (Received Signal Strength Indicator) based S-meter available in the FM mode.  (See Table 1).  As shipped from the factory, the S-Meter on my '817 (but not necessarily yours) was as Table 1 shows.  As is typical for RSSI circuits, there are is on-chip circuitry that performs a logarithmic conversion, so the RSSI voltage (digitized by the FT-817) is proportional to the signal strength in db - and with reasonable accuracy!
Judging by this data I could see that the FM S-Meter reading was more appropriate to weak signals and the SSB/CW/AM scale was more useful with strong signals.  I even wrote the first algorithm to switch the radio between AM (with its wider IF filter being less affected by incidental modulation on the repeater) and FM in order to get the "best of both worlds" - that is, take data from both scales.

There is a problem with this:  The "average" signal strength of the repeaters around here is between -50 dbm and -90 dbm - a range that is wholly outside that available via the FM S-Meter - where I had decent resolution.  What could I do?  Clearly, knowing the signal strength to the nearest 5-10 db is hardly adequate - and such would be the case were I to use only the SSB/CW/AM S-Meter.  I could manually switch some external attenuation in and out to be able to use the FM S-Meter, but this didn't seem to be particularly desirable.

I took another look at the '817 and noticed that there were two items in the "Soft Calibration" menu referring to the FM S-Meter:  #9, FM-S1 and #10, FM-FS.  These two parameters do precisely as their labels imply:  They set the amount of signal that correlates with S1 and full-scale reading respectively.  I noticed that the FM-FS reading grossly shortchanged the dynamic range of the FT-817's RSSI circuit:  Using a service monitor I determined what the range of the '817's hardware actually was (by setting FM-FS to 255 and then noting the signal level that produced "compression" - that is, where further increases in the S-Meter reading no longer correlated with increases in signal strength) and setting FM-FS just below this.  The result of this "recalibration" may be found in Table 2.
0 <-110.7 <-98.5 <-88.1 <-78.7
1 -108.9 -96.8 -86.7 -77.7
9 -91.5 -79.5 -69.6 -60.4
D -82.1 -70.5 -60.2 -50.9
E -78.1 -66.3 -56.5 -46.9
F >-75.7 >-63.8 >-53.6 >-44.5
n/a 12.0 22.0 31.1
Table 3
Values of the VHFRXG parameter (soft calibration menu item #5) versus the signal input.  The bottom row shows the average difference between the "unattenuated" reading versus the reading obtain with differing amounts of "attenuation".
Note:  The above values are for my FT-817.  Every '817 will be different, requiring individual calibration to assure accuracy.

The improvement is rather dramatic:  Rather than covering only about 13 db or so (readings of 0 and F aren't useful, except for indicating too little or to much signal) the range has been increased to about 30 db.  Each step was approximately 2 db - a reasonable compromise between resolution and accuracy and considering the imprecise nature of real-world signal analysis, it was more than adequate.  But, there was still a problem:  Even having a useful range from -109 dbm to -78 dbm wasn't sufficient for field measurements without additional attenuation.

It further occurred to me that the FT-817 had yet another parameter that could be useful here:  The VHFRXG parameter (soft calibration menu #5) did precisely this.  A bit of experimentation showed that using this parameter, at least 30 db of extra "attenuation" could be added.  Further experimentation showed that once this amount of attenuation was determined precisely, one simply added it to the calculated signal reading.  In other words, only one table needs to be constructed with all 16 signal levels and fixed constants are added to the result depending on the attenuation.

The results?  Table 3 details the measurements.

As you can see, the useful signal measuring range has been extended to approximately 64 db with a resolution of roughly 2 db - a range and resolution more than adequate for our signal measurements.

I could now get down to the business of writing the software...

Because of this project I had already started determining how to read/write to/from the EEPROM in the '817 as well as what data resided at which address.  Since I could already modify the "Soft Calibration" factors via the serial port, I was well on my way toward being able to use the information I had gathered to my advantage.  One of the functions I wrote was responsible for getting the signal level from the '817 - a task somewhat complicated by the necessity to implement the "software" attenuator.  The algorithm to do this is approximately thus:

"What about Fresnel effects?"

You may be asking yourself, "Self, what about the Fresnel effects - did you take those into account?" 

The answer is:  Yes and No.  Because the elevation of the repeater is approximately 7200 feet and our receiver was at (more or less) 4400 feet most of the time, plus the fact that we kept a distance to the repeater of between 5 and 13 miles, the "Fresnel Dip" was usually well above the intervening terrain. 

Because so many pieces of data were collected - the majority of them having been taken with the vehicle in motion - it is unlikely that a "standing null" would have any significant effect on the data as a whole.  Even if the data is "skewed" by a few db either way, the data can still be used in a "relative" fashion.   Also, don't forget that any repeater user will also be subject to any fresnel-related degradation, so what we are getting are more "real world" readings, anyway. 

What are fresnel effects?  Along a signal path, if there are obstacles near the direct line-of-sight path, these objects can cause a slight change in that portion of the signal path.  Some of these changes can reinforce the signal, and other can destructively cancel it.  On 2 meters with a path of about 10 miles, an obstacle located midway between the transmitter and receiver that is a few hundred feet below the direct line-of-sight path can affect signals.

The result of all of this work?  The signal level displayed by the program precisely tracks that applied to the '817 itself as well as can be expected considering that the resolution is approximately 2 db.  I then became curious as to how well the calibration held up with a change in temperature and put the '817 in the freezer.  With the '817 at approximately 35 degrees F it read a consistent 3 db low across the entire signal range- not too bad...

Note:  While the "conversion table" was originally designed for use on 2 meters, I determined that additional constants could be used to provide just as accurate calibration for 6 meters and 70 cm - and I did so, allowing this program to be used on those bands as well.

The rest of the program:

The remainder of the program was rather straightforward.  The additional program functions perform the following tasks:

The results:
Combined coverage and antenna pattern map of the 146.76 repeater
This is a coverage map (with superimposed antenna pattern) of the '76 repeater, based on the measured signal strength, predicted losses due to terrain, and calculated antenna pattern.  Note the narrow, deep null due north of the repeater site.
This image was produced with the help of VE2DBE's RadioMobile program.  For more information about this map, go to the UARC 146.76 repeater page.
Click on the image to get a larger version.

After writing this program (in April of 2001) I was pleased and gratified that the program worked exactly as expected:  The computer, radio, and GPS were connected and we simply "circumnavigated" the repeater site - a process that took several hours, some of it over very rough, 4-wheel drive high-clearance terrain.   When the program was used on repeaters whose antenna system were known to work properly and whose transmit system properties were known, the display values agreed with the calculated values to within 2-3 db or so.

As for gathering the data, all that was required was to drive around and take readings.  Obviously, we chose our route specifically to maintain line-of-sight with the repeater as much as possible as well as avoiding areas cluttered with building, trees, and other objects.  Being able to take measurements "on the fly" was very useful - we were able to take readings while cruising down the Interstate.  If we'd had to stop to take readings, we probably would not have done it alongside a freeway...  Fortunately, we were able to use an auxiliary/control link to enable us to keep the repeater's transmitter keyed continuously - something that greatly helps in the collection of signal strength.

The analysis of the '76 repeater showed that there was, indeed, a very deep (25-30 db) deep null beamed precisely at the western portion of the Salt Lake valley.  Furthermore, the deepest part of the null was very narrow (only about 3 degrees or so at the 10db points) - a width that could have resulted in the true depth of the null being missed entirely had the "old fashioned" signal measuring method (i.e. stopping the car, measuring the signal, etc.) used.  While this null was expected, the measured magnitude was a surprise.  Clearly, we'll have to look into re-mounting the antenna (or adding a "fill" antenna) to minimize this null.

Note:  The described program was written in C and operates under DOS.  Because it was written for "one-off" use, only enough effort was put into it to make it stable and useable:  Further enhancements aren't contemplated at this point.  Note that it produces raw, human-readable ASCII data that is intended to be imported into another program for analysis, or simply "looked at."  If you have serious interest in this program, look at the Repeater Monitoring Software page.

A coverage map was created showing predicted signal strength of the repeater.  An actual antenna pattern was derived by taking the empirical antenna gain measurements derived from mobile signal strength readings and adding to those the predicted signal loss for that location due to terrain.  This derived pattern was then used to create the plot.  It has "general" agreement with the field-measured data.  (The map was produced with the help of VE2DBE's RadioMobile program.)

Work continues on this page - please revisit soon!

Note:  CAT (with respect to this interface) is a trademark of Yaesu/Vertex Standard Co. Ltd.

Notice:  The information contained on this and related pages is believed to be accurate, but no guarantees are expressed or implied.  The information on this and related pages should be considered to be "as-is" and the user is completely responsible for the way this information is used.  If you have questions or if you find information that you believe to be incorrect, please report it via email.

The KA7OEI FT-817 "Front Page" - This is, well, the "front" page of the '817 pages here...
The attenuation of some common objects and obstacles on 2 meters

Having real-time readings while driving around (the passenger was looking at the computer - not the driver!) made it possible to answer a few questions that we'd had concerning the effects of obstacles on the signal strength of 2 meter signals: 

  • A windbreak of large Cottonwood trees - green, with leaves:  10-15 db.
  • Going under an overpass:  15 db or so - with the repeater off to the side.
  • Being in the lane next to a large metal Semi-trailer that is blocking your view of the repeater:  Between 10 and 20 db.
  • Being behind a large, earthen embankment alongside the freeway:  10-15 db.
  • Having a low, flat mountain range (or hills) blocking your view of the repeater:  15-20 db, once you are a short distance away from it, 20-25 db if you are right next to it.
Remember:  These values are for 2 meters and are, at best, "ballpark" figures.  Your mileage will vary.


Any comments or questions?  Send an email!

This page maintained by Clint Turner, KA7OEI and was last updated on 20091130.  (Copyright 2001-2009
 by Clint Turner)

Since 12/2010: