View Single Post
Old 29-Apr-2016, 5:13 PM   #101
tripelo
Senior Member
 
Join Date: Jan 2010
Location: Dallas, TX
Posts: 123
The HDTVi Antenna as a Reference

(continued from previous post)

Part 4. The HDTVi Antenna as a Reference

Another way of comparing gain of the antennas is to directly compare received signal levels to that of the HDTVi antenna. The HDTVi antenna was selected as a reference antenna mainly because it was thought that itís gain would be relatively flat across the UHF band. Also it was thought that the SWR of the HDTVi would be relatively good (close to 1). Part of the reason for the above assumptions is that Log Periodic antennas can be designed to meet the criterion: flat gain and favorable SWR. The physical appearance of the HDTVI suggest as much.

A 4NEC2 analysis of the HDTVi tends to support the assumptions of flat gain and favorable SWR.

Below are images of the results of 4NEC2 analysis




Top Image shows Gain in dBi (recall dBi=dBd+2.15), Lower image shows SWR


The above simulation SWR results agree favorably with the actual measurements. If the HDTVi gain were to be measured, likely it would also agree. If that were the case, than a direct comparison of the higher gain antennas with the HDTVi might provide insight.

Below is an image that shows the other four antennas compared directly with the HDTVi antenna:






The above image illustrates the tendency of increasing gain of the four higher-gain antennas with increasingly higher frequencies.

As can be seen in this image and a preceding image, there is a discontinuity at channel 20 and channel 31. These two channels have lower signal strength than most of the others and also have been observed to exhibit multipath effects. The effects that were related to multipath were observed as part of a previous field test performed in May 2012, some results shown in a previous post (Ch 20 & 31 data were not presented).

UHF Signal Strength vs Antenna Height (AGL)

Omitting those points (Ch 20 & 31), the overall data might look smoother, but they were included for completeness.

The following image represents a best-fit linear approximation to the measured data and could illustrate some trends. Caution when interpreting this linear approximation; for a specific channel, it might be better to refer back to the raw data averages (shown earlier).




Both Yagis show increasing gain with increasing frequency (channel number). The 8-bay bow tie, or panel array, antennas show a flatter response with some increased gain with increasing frequency.

CM4228 and DB8

Notable is the approximately 2 dB difference in gain between the CM4228 and the DB8. This difference could partially be explained by the increased SWR of the DB8 (about SWR=4 at mid band), an SWR of 4 could cause approximately 2 dB loss. Some of the signal strength difference and transmission line matching could be attributed to baluns. The DB8 balun was supplied by the manufacturer (Antennas Direct). The CM4228 balun was a homebrew hand-wound ferrite-core balun with a measured loss of less than one dB.

91XG and HDB91x

The measured signal level difference between the gains of the two Yagi style antennas is less pronounced. The HDB91x shows a clear advantage at the top end of the band, however this indication is partially an artifact of the linear approximation. Note the raw data average show signal levels from the two antennas to be fairly close near the upper end of the band. The reason for the overall differences is unknown, but could partially be related to a better impedance match to the transmission line. The S11 measurement of the HDB91x reveals a better impedance match starting at ~500 MHz (Ch 19) and upwards. It is known that the 91XG has an ~microstrip balun that in some ways resembles a usual half-wave loop (but is markedly different). The factory-supplied balun of the HDB91x has not been inspected.

After a first set of data on the 91XG was collected, it was thought something might be amiss. The 91XG balun housing was opened and inspected; a nut that holds the driven element to the balun seemed to not be tight. The nut was tightened and a second set of data was collected. The results of the second 91XG data set were close to that of the first set.

Overall, the appearance of the 91XG and the HDB91x seems similar. But, close inspection shows most component parts of the HDB91x are in some way different than the 91XG. There are measurable physical and electrical differences that may account for some of the difference in performance.

Test Location and Data Variability

Even though the test site is LOS to the DFW stations, some multipath has been observed with received signal on some channels. This is probably a result of the first Fresnel zone not being clear.

A repeat of each measurement probably helped reduce effects of brief transitory multipath effects. The individual data sets were fairly consistent. Measurements using the HDTVi at both beginning-of-test and end-of-test somewhat verified data consistency. The HDTVi final-measurements were about 0.2 dB lower in signal strength than the beginning-measurements. A scaled correction factor could have been incorporated into the data but was deemed unnecessary.

.
Attached Images
File Type: jpg HDTVi 4NEC2 Gain SWR.jpg (81.0 KB, 2355 views)
File Type: gif 4 Antenna Compare HDTVi.gif (16.2 KB, 2366 views)
File Type: gif 4 Antenna Compare - Linear.gif (11.6 KB, 2421 views)
tripelo is offline   Reply With Quote