Canadian TV, Computing and Home Theatre Forums banner

581 - 583 of 583 Posts

·
Registered
Joined
·
94 Posts
My claim is that this galvanic insulation has no measurable effect.
Last weakend I started drawing 91XG in HFSS (from 4NEC2 model), and next weakend plan to simulate it.
When finished, I can provide surface current s animation and E/H fields animation on any cutover or surface
Yurii. дякуємо за весь ваш інтерес та допомогу щодо цього Repack 91 XG

I'm looking forward to seeing your modeling results and your very interesting animations.
 

·
Registered
Joined
·
94 Posts
Before I go into possible ways to increase Raw Gain on the highest channels (I'm going to compare readily available 698 MHz data points), I have to explain that there are several Antenna Modeling results for the A-D 91XG, which have minor differences, meaning we have to be careful to compare Apples-to-Apples.

Antennas Direct did not post the modeling results for the 91XG (dtg 2009) until much later.....I show a Dec2011 date for when it first appeared for download....long after the below modeling results were posted. It is important to note that we KNOW that there have been 75-ohm versions of the 91XG as well as the current 300-ohm version. Furthermore, the 91XG version described below, as measured by Ken Nist, might have had the SAME or DIFFERENT dimensions than the model analyzed by A-D using the Remcom, Inc. X-FDTD 7.0 Antenna Modeling program...and of course it was a different software package....and apparently does NOT use Method-Of-Moments technique as used in NEC2. So don't try to draw too many conclusions from A-D Claim of 17.44 dBi at 698 MHz. Interesting, but we can't do accurate Apples-to-Apples comparisons with that number, since we don't have access to both A-D's model and (pricey) X-FDTD Software:
https://www.antennasdirect.com/cmss_files/attachmentlibrary/Technical Data PDF's/91XG-TDS.pdf

Ken Nist (HDTVPrimer website owner) posted an EZNEC model of the 91XG (some time prior to Dec 2008 when it was no longer available), presumably based on his own measurements. User 300ohm downloaded the EZNEC file back when it was still available, imported it into 4nec2 and did some tweaks normally encountered when importing a non-native file format.....this is called "XG91a.nec", which I uploaded results to my website, reporting 16.9 dBi at 698 MHz...but this was based on applying Auto Gain Test Correction ONLY at 584 MHz. Note that I subsequently went from RevA up thru RevF, trying to minimize the dozens and dozens of Warnings found by 4nec2...so versions may have up to a few tenths of a dB difference between them, which are NORMAL FLUCTUATIONS in Antenna Modeling programs:
UHF 91XG Corner-FD-Yagi

User 300ohm also posted a modified file, adding SIX more Bowtie Directors...which are ALL THE SAME SHAPE and SIZE throughout the Antenna, called "XG91a_Plus6Directors.nec"....but since he simply took the front 6 Bowties from another Antenna, the Bowtie-to-Bowtie Spacings are NOT Optimized...they're a simple repeat of the preceding six Bowtie Spacings (no they are NOT equally spaced...they're ALL DIFFERENT Spacings). The links to 300ohm's files can be found in my analysis summary. I did not upload results for this file into my webpage (probably because I was waiting for RevA-RevF to sort itself out). I just now ran an Apples-to-Apples comparison between this file and the above file, this time using nikiml's EVAL function which recalculated AGT Correction at each test frequency (hence 0.1 dB difference).

Raw Gain at 698 MHz increased from 16.8 to 17.9 dBi when adding SIX Directors, or about 1 dB, after increasing Boom Length from 90.2-in to 122.2-in. This is about what we should expect, since "Rule-Of-Thumb" for Yagis's is that the Boom Length has to DOUBLE to increase Gain by about 3 dB. [I haven't calculated results if add perhaps 24 Bowties since it would be mechanically unstable without a much stiffer Boom System.]

=============================================
I also tried Adjusting the Size of the Corner Reflector. A WIDER Corner Reflector didn't help high frequency response....but a TALLER Reflector provided a VERY MINOR IMPROVEMENT.

Using Same Rod-Rod Separation, adding SIX more Reflector Rods (3 each Top & Bottom) improved Raw Gain at 698 MHz by a mere 0.1 dB and adding TWELVE more Reflector Rods improved Raw Gain by only 0.2 dB. This is hardly worth the extra work, cost and weight increase/imbalance...and would be too small a difference to be noticed:
UHF 91XG - Taller Corner Reflector
Note that I used 91XG RevF version and to ensure I was comparing Apples-to-Apples, I replaced the 91XG's complicated "Grid" Reflector (where vertical wires are pretty much inconsequential) with a simplified 20 Reflector Rod version, as is common in many other Corner Reflector Yagi's.

=====================================================
To truly improve the EXISTING 91XG without completely re-manufacturing the Antenna, the only other HALF-WAY GOOD idea I have is to take the EXISTING set of matching Bowtie Directors and recalculate an OPTIMIZED set of Bowtie-to-Bowtie Separations to Re-Populate a NEW STRONGER BOOM SYSTEM THAT IS TWICE AS LONG...and hence MIGHT provide up to a 3 dB increase....but modeling would be needed to determine whether that could be achieved by using the Existing number of Bowtie Directors....or whether a second 91XG would need to be cannibalized for extra Bowtie's....

I'm NOT volunteering to run THAT Optimization....two to four dozen SYmbol Variables in an Antenna with 4617 Segments....that's going to take many DAYS....or WEEKS for each run....and it usually takes AT LEAST half a dozen runs to arrive at an "acceptable" solution for a new design...

=====================================================
What CAN improve on the performance of a 91XG is to upgrade to a PAIR of MATCHED 91XG's (gee....I wonder how many different "versions" of the 91XG there are....). See the Stacked Antenna Thread...a Vertical or Horizontal Stack in a "Idealized World" with no Fading and Gaussian Noise (etc) will provide UP TO 2.5 dB of Gain, using a conventional RF Combiner.....but measurements of RF Combiner "efficiency" posted at www.antennahacks.com illustrate that even this is wishful thinking....esp. on higher frequencies...due to Amplitude and Phase Mismatch INSIDE the RF Combiner (Hybrid in RF Combiner only Sums signals when they are MATCHED).

HOWEVER, the BIGGEST improvement using Stacked Antennas is the Spaced Diversity Combining Gain since it is unlikely for BOTH Antennas to experience a Multipath Null at the same time. Typically there is 10-20 dB of Diversity Combining Gain in a Multipath Rich environment....and not so much for non-Faing, Line-Of-Sight paths.

BTW: Vertically Stacked 91XG's should have a Vertical Height Separation of AT LEAST 5-ft to minimize interaction degradations:
UHF Vertical Stacked A-D 91XG
I have NOT (yet) run equivalent simulations for Horizontally Stacked 91XG's....
We're still working on this project trying to figure out the exact measurements for the director extensions. I read back a little bit in this thread and found this. It's an excellent recap on Holland's work even touches on stacking and ganging 91 XGs this is a long thread, and I still keep finding things that give me hope. I do have some great news. I called Antennas Direct yesterday. I inquired about getting some new X directors for my storm damaged 91XGs which I considered a total loss. Instead of me buying some new X directors they decided to warranty my two antennas! I am still in shock. This is amazing, now I'm going to really have some spare parts ! :cool:

12012


I appreciate everybody's help so far! I really want to make this work.

If anybody has any more input on director spacings, director extensions etc... please chime in. I've got some momentum here.
 

·
Premium Member
Joined
·
1,159 Posts
We're still working on this project trying to figure out the exact measurements for the director extensions. I read back a little bit in this thread and found this. It's an excellent recap on Holland's work even touches on stacking and ganging 91 XGs this is a long thread, and I still keep finding things that give me hope.
Thank you for the link to holl_ands post in Dec 2014 that gives the background history. His model with the wider Bowties goes a little further to increase the gain for the lower channels with the MAX at 626 MHz.
UHF 91XG - Wider Bowties, Better Ch14-51

I used this NEC file on that page
UHF 91XG - Wider Bowties, Better Ch14-51

to show on the NEC Viewer by nikiml. The director extensions show, but he didn't increase the director spacing because the boom length is the same as the stock 91XG.

91XG nikiml NEC Viewer wider directors side view_1.jpg


91XG nikiml NEC Viewer wider directors front view_1.jpg


That's pretty close to the estimate made by jop.
Here is the director's modification to a 117% rescaling in the reference model he uses.
The extension that is added is approximately 1 inch, the simulation is like this
But I still want to find a way to confirm that extension length with a measurement method.
 
581 - 583 of 583 Posts
Top