"Top Hat" GH with NARODs for VHF-HI: Major GH Improvement - Page 5 - Canadian TV, Computing and Home Theatre Forums

nikiml · 2010-06-10, 05:29 PM

I did it with the model from post 54:

Code:

Run     SWR     Gain    F/B     F/R     R-a     X-a     Eff.    Segs

1-0     1.4754  13.13   6.12    6.12    429.21  55.034  99.87   328     
1-1     1.458   13.1    6.11    6.11    432.63  32.872  99.87   426     
1-2     1.4458  13.05   6.09    6.09    433.7   2.9925  99.87   570     
1-3     1.4452  13.02   6.09    6.09    432.2   -17.46  99.87   754     
1-4     1.4526  12.99   6.09    6.09    428.61  -39.94  99.87   1004    
1-5     1.4662  12.98   6.1     6.1     424.21  -58.6   99.87   1310    
1-6     1.8526  13.16   6.1     6.1     510.02  -126.2  99.87   1752    
1-7     1.7968  12.89   6.13    6.13    454.7   -155.8  99.88   2318    
1-8     1.8304  12.62   6.2     6.2     406.3   -186.1  99.88   3066    
1-9     2.0167  12.15   6.35    6.35    333.99  -224.1  99.89   4074    
1-10    2.6439  10.07   7.72    7.72    115.99  -41.32  99.9    5418    
1-11    3.8206  5.9     10.98   10.98   121.58  -210.1  99.92   7198    
1-12    1981.3  8.27    10.45   10.45   0.1527  -27.33  98.71   9570

The last line should be ignored since that is where the test was interrupted by an error.

Autofils · 2010-06-10, 09:23 PM

Nikiml,

Re your conv-test results in post #61, I sent you an email with some tips and suggested investigations you could do to obtain a better understanding of these results. I don't have the time to dig into now.

I would simply point out that runs 1-0 to 1-3 are for segmentation at 10 segs per 1/2 wave to ~23 segs per 1/2 wave and the gain varies from 13.13 to 13.02. The next run 1-4 has a total seg count of 1004, which is ~31 segs per 1/2 wave.

I suspect that for run1-4 (and 1-5 to 1-12), the normal 4nec2 segment check would report errors. If this turns out to be the case, conv-test results for these runs are not valid.

If you have the time, it would be very interesting to investigate a manual nec file based on run 1-4 at ~ 30 segs per 1/2 wave. Does it pass the normal 4nec2 segment check with no errors at 585mHz? What about runs 1-5 to 1-12 ?

nikiml · 2010-06-11, 01:14 AM

You were exactly right.
1-0...1-4 are warning free, but the rest dont pass the segment checks

mclapp · 2010-06-11, 10:21 AM

I was going to run that nec model (post 54) in EZNEC and run the auto seg on it but I couldn't import it in because EZNEC likes to import actual x,y,z cooridinates and those are not. EZNEC gives 2 options for segmentation conservative or minimum recommended for the frequency you select. I was curious what it would do with that model but it didn't happen. I know with other models I've adjusted the segments to 1 and then let EZNEC auto segment and it did a great job.

Many times after auto segmentating certain models I had to adjust the source wire diameter very close to the size of the wires it was attached to to get the AGT back in line. I would think that is a good thing because I'm still on the fence about adjusting the source wire much to get a good AGT.

I've also found that some things that EZNEC doesn't protest about 4NEC2 does like certain dia/seg ratios and dia/length ratios. I'm not very good with 4NEC2 I really only use it to get some of the nice graphs it generates.

Autofils · 2010-06-11, 12:26 PM

Quote:

Originally Posted by mclapp

..I was going to run that nec model (post 54) in EZNEC and run the auto seg on it but I couldn't import it in because EZNEC likes to import actual x,y,z cooridinates and those are not.

mclapp,

You can easily convert the 4nec2 nec file into compatible (x,y.z) co-ordinates for EZNEC, by resaving the nec file in 4nec2's geometry editor, under the settings tab in 4nec2's main window. That save will strip out all the SY cards.

re your other comment:

Quote:

Many times after auto segmentating certain models I had to adjust the source wire diameter very close to the size of the wires it was attached to to get the AGT back in line. I would think that is a good thing because I'm still on the fence about adjusting the source wire much to get a good AGT.

I have been absent from the forum for the past 15 months or so, but when I started my trolling on the Antenna R&D forums to get caught up on past events, the comments on AGT by j3d really caught my attention. The public domain nec2 core has well documented limitations, one of which is connection of wire elements having different diameters. If the AGT test is adjusted for 1.0 by changing the wire diameter of the Vsource, then we are "adding" to this nec2 core limitation in our model, rather than always trying to minimize the nec2 core limitations to get the "best" model for representing the antenna design. This is the point that j3d was making.
Reference post by j3d 5 May 2009 --> http://www.digitalhome.ca/forum/show...4&postcount=45

From my reading in this thread, I don't see that much discussion took place on this subject. For those that have a desire to better understand the limitations of the nec2 core and how best to deal with nec2 limitations in their models, I see the question to be answered as this:

If the adjustment to the wire diameter of the Vsource is only a very small percentage, does this "procedure" have merit or should it be abandoned?
In post 49 of this thread, 300Ohm mentioned the following:

Quote:

Originally Posted by 300Ohn

...And maybe Arie Voors too. Hes posted on this website in your absence, and likes us too. So has Dr. Natan Cohen, the Fractal Guy, the originator of the fractal antenna.

I would suggest that this subject should be discussed in more detail.
I'm sure that all, would like to get the views from Arie and Dr Cohen.

Cheers

300ohm · 2010-06-11, 01:17 PM

Quote:

If the adjustment to the wire diameter of the Vsource is only a very small percentage, does this "procedure" have merit or should it be abandoned?

Ive always looked at it, because the Vsource wire isnt a real wire in the build anyway, that its sole purpose is for AGT adjustment. (or maybe possibly as part of a balun wire)

Quote:

I've also found that some things that EZNEC doesn't protest about 4NEC2 does like certain dia/seg ratios and dia/length ratios.

Heh, yeah. I think some of those warnings may be built in cost overrun warnings. Like, hey, youre using wire too thick and pricey for the purpose, heh.

Although, really really fat dipoles start to behave like bowties.

Autofils · 2010-06-11, 02:48 PM

Quote:

Originally Posted by 300Ohm

Ive always looked at it, because the Vsource wire isnt a real wire in the build anyway, that its sole purpose is for AGT adjustment. (or maybe possibly as part of a balun wire)
... Heh, yeah. I think some of those warnings may be built in cost overrun warnings.

300Ohm,

Re your comment "the Vsource wire isnt a real wire in the build anyway": Are you stating an opinion that you believe the "AGT wire adjustment procedure" is valid or it is not valid, since the Vsource wire element is not a part of the antenna design?

I'm also not exactly clear on what you mean by "those warnings may be built in cost overrun". By cost overrun, are you implying that 4nec2 warnings are not really all that important and thus can be ignored? Only errors are really important?

I'm not being critical, I just feel that clear understandings are needed to have a good discussion.

Cheers

300ohm · 2010-06-11, 04:15 PM

Quote:

Are you stating an opinion that you believe the "AGT wire adjustment procedure" is valid or it is not valid, since the Vsource wire element is not a part of the antenna design?

That it is valid for adjusting the antenna to get the Average Gain Test value to 0 db, so that the model would be reliable. (other than adding or subtracting the AGT value, which on a sweep becomes unmanagable)

Quote:

By cost overrun, are you implying that 4nec2 warnings are not really all that important and thus can be ignored? Only errors are really important?

Yeah, the errors are the big no-no's.

Warnings on the otherhand,
The warning for the wire diameter size for reflectors, doesnt seem to make a difference in real life builds, from what Ive observed, other than fatter/wider the better.
And the angle too sharp warnings, for example on bowties, dont seem to amount to anything either.

I havent hit upon every warning yet, so there may be some that should cause serious concern. But for the most part it seems the warnings are something that should be acknowledged, ie yes its a fat reflector or yes thats a sharp angle, but then move on, heh.

300ohm · 2010-06-11, 06:33 PM

Since the reflector-less SBGH with NARODs is a pretty popular antenna to build, (not ANY bi-directional vhf-hi/uhf antennas are commercially available AFAIK) nikimls top hat NAROD is a good overall compromise over the vhf-hi band and it minimizes the channel 33 -38 gain dip. For channel ranges 7 - 9 or 11 - 13, and if channel 33 - 38 arent a concern, the specific vhf-hi channel range vhf-hi straight NAROD would be the ticket.

NEC file with gen1 reflector-less SBGH with nikimls optimized top hat NAROD.

Code:

CM model segmentation is derived from autosegmentation 21 at 698Mhz
CM AGT = 1.0 (0db) at 585 mhz
CM Based on Nikiml top hat gen1 SBGH with 89mm feed point 
CM and using 6 gauge wire for driven elements and NARODs.
CE
GW	1	21	0	-1.7519685	0	0	-6.7629921	5.01101575	0.08101138
GW	2	21	0	-6.7629921	5.01101575	0	-1.7519685	10.0220315	0.08101138
GW	3	21	0	-1.7519685	10.0220315	0	-6.7629921	15.0330472	0.08101138
GW	4	17	0	-6.7629921	15.0330472	0	-12.353543	15.0330472	0.08101138
GW	11	7	0	-8.5472441	17.4304724	0	-8.5472441	15.4934646	0.08101138
GW	14	15	0	-8.5472441	15.4934646	0	-13.429134	15.4934646	0.08101138
GW	21	21	0	1.7519685	0	0	6.76299213	5.01101575	0.08101138
GW	22	21	0	6.76299213	5.01101575	0	1.7519685	10.0220315	0.08101138
GW	23	21	0	1.7519685	10.0220315	0	6.76299213	15.0330472	0.08101138
GW	24	17	0	6.76299213	15.0330472	0	12.3535433	15.0330472	0.08101138
GW	31	7	0	8.54724409	17.4304724	0	8.54724409	15.4934646	0.08101138
GW	34	15	0	8.54724409	15.4934646	0	13.4291339	15.4934646	0.08101138
GW	10	51	0	8.54724409	17.4304724	0	-8.5472441	17.4304724	0.08101138
GW	41	21	0	-1.7519685	0	0	-6.7629921	-5.0110157	0.08101138
GW	42	21	0	-6.7629921	-5.0110157	0	-1.7519685	-10.022031	0.08101138
GW	43	21	0	-1.7519685	-10.022031	0	-6.7629921	-15.033047	0.08101138
GW	44	17	0	-6.7629921	-15.033047	0	-12.353543	-15.033047	0.08101138
GW	51	7	0	-8.5472441	-17.430472	0	-8.5472441	-15.493465	0.08101138
GW	54	15	0	-8.5472441	-15.493465	0	-13.429134	-15.493465	0.08101138
GW	61	21	0	1.7519685	0	0	6.76299213	-5.0110157	0.08101138
GW	62	21	0	6.76299213	-5.0110157	0	1.7519685	-10.022031	0.08101138
GW	63	21	0	1.7519685	-10.022031	0	6.76299213	-15.033047	0.08101138
GW	64	17	0	6.76299213	-15.033047	0	12.3535433	-15.033047	0.08101138
GW	71	7	0	8.54724409	-17.430472	0	8.54724409	-15.493465	0.08101138
GW	74	15	0	8.54724409	-15.493465	0	13.4291339	-15.493465	0.08101138
GW	50	51	0	8.54724409	-17.430472	0	-8.5472441	-17.430472	0.08101138
GW	100	11	0	-1.7519685	0	0	1.7519685	0	0.06424494
GS	0	0	0.0254		' All in in.
GE	0
EK
LD	5	0	0	0	2.49e7	0
EX	0	100	6	0	1	0
GN	-1
FR	0	1	0	0	698	0
RP 0 60 73 1001 0 0 3 5

mclapp · 2010-06-12, 11:09 AM

Adjusting the source wire can also be accomplished by changing it's length and get the same result.

In many of my models I use wires running away from the feed point like balun wires would be and then attach the source to that. By doing that it allows me to change the source wire length and it has the same effect as changing the diameter. I don't think it's any better than changing the source wire diameter except maybe it doesn't fall into some of NEC limitations but I don't think that a very small increase or decrease in wire diameter does either. I've tried it both ways and got the same results.

If your segmentation is way off and you have to adjust your source wire a lot to get a good AGT then probably your convergence is bad due to the poor segmentation. In that case I would say adjusting the source wire diameter is just fooling yourself.

Autofils · 2010-06-13, 11:25 AM

My thanks to 300Ohm and mclapp for their comments:
Reference links:
300Ohm (post 68) --> http://www.digitalhome.ca/forum/show...1&postcount=68
mclapp (post 70) --> http://www.digitalhome.ca/forum/show...2&postcount=70

Like 300Ohm, I have used the AGT adjust "procedure" in all my past posted 4nec2 nec files. I never really questioned the effect of different wire diameter, on this AGT adjustment, because the results never indicated any significant wrong results.

After recently reading j3d's post on the subject, and with a re-read Cebik's modeling guide, part 4 ( the last 2 pages under the heading "Testing Models"), I thought it would provide a better understanding of 4nec2 limitations, if jed's question could be answered with some detail and if possible some reference links.
Reference link to j3d's post --> http://www.digitalhome.ca/forum/show...4&postcount=45

For me, the questions that need answers is simply this: ... in the case of the AGT adjust procedure, do two wrongs actually may a right" and along with that, do other modelers, besides DigitalHome.ca forum members, use this AGT adjust procedure?

The recent posting (post 61) by Nikiml re Convergence Tests on his model, provides some "food -for-thought" and I invite comments. If the majority of readers on this subject feel, that this is somewhat of a waste of time, I will stop posting on this subject, as I have plenty of other activities to keep me busy. The only way I can "measure" the interest in this subject is from the posts that are made.

Food - for - Thought Comments
=========================
Nikiml's post 61, Convergence-Tests showed, that up until segment-check errors occurred, the raw gain varied from 13.13 dBi (total_segs=328) to 12.99 (total segs=1004). Reference link (post 61): --> http://www.digitalhome.ca/forum/show...9&postcount=61

By extrapolation of Nikiml's data, 21 segments per 1/2 wave raw gain would be 13.03 dBi. For this specific model, the Convergence Test reveals the "most accurate" raw gain convergence is 12.99dBi. So does the recommended model at 21 segments per 1/2 wave using the AGT adjust procedure, give a more accurate result, not only for this model, but for any model, regardless of the complexity? The beauty of the Non-adjusted AGT Test and the Convergence Test is: you can depend on the results, regardless of antenna model differences (GH vs Yagi or BowTie etc)and the antenna's complexity.

In an effort to continue this discussion, here are some more "food-for-thought" links.

1. Arie's 4nec2 Forum
The AGT is off - Why? A bug?
http://fornectoo.freeforums.org/the-...-bug-t175.html

Auto Segment
http://fornectoo.freeforums.org/auto-segment-t189.html

2. Practical Antenna Modeling - WX7S
http://www.wx7s.com/wordpress/?m=200812

3. Antenna Modeling for Radio Amateurs
Here is an excellent overview about some of the history of antenna modeling with references to mathematical models. There is a brief discussion about limitations of the various modeling approaches and a summary of EZNEC and 4nec2 is included. This pdf doc was written by Steve Stearns, K6OIK.

http://www.fars.k6ya.org/docs/Antenn...o-Amateurs.pdf

...Cheers

nikiml · 2010-06-14, 02:44 AM

I decided to cleanup the models I posted earlier, to remove the segmentation warnings, to fix the spacing between the NARODs and the driven elements to 12mm and remove few inaccuracies.
Bellow are the dimensions for the "cleaned-up" jeds GH mods with Top hat NARODs and 3 narod reflectors.

I propose to call them GH10n3,..., GH4n3 - n for NARODS and 3 for the narod reflectors.

For all models use the dimensions from the JED's website plus the additional dimensions from below.

The X dimension is the distance between the driven elements plane and the NAROD reflectors plane.
s is the distance (center to center) between the horizontal legs of the NARODS and the driven elements.
The b dimension from the JED's models is also the length of the horizontal legs of the narods.

Code:

	    GH10n3	       GH8n3	      GH6n3	      GH4n3	
--------+-----------------+--------------+---------------+---------------+
Dim	| mm	inch	  | mm	inch	 | mm 	inch	 | mm 	inch	 |
--------------------------------------------------------------------------
s	| 12	1/2	  | 12	1/2	 | 12	1/2	 | 12	1/2	 |
a	| 427	16 13/16  | 428	16 7/8	 | 399	15 13/16 | 405	16    	 |
h	| 49	1 15/16	  | 48	1 7/8	 | 60	2 3/8	 | 57	2 1/4	 |
L1	| 907	35 11/16  | 960	37 13/16 | 995	39 3/16	 | 961	37 13/16 |
L2	| 804	31 11/16  | 813	32    	 | 813	32    	 | 810	31 15/16 |
Z	| 360	14 3/16	  | 433	17    	 | 448	17 5/8	 | 470	18 1/2	 |
X	| 240	9 7/16	  | 300	11 13/16 | 314	12 3/8	 | 316	12 7/16	 |

Gain values (scroll for all models):

Code:

	GH10n3					
=======================			
Ch.	Freq 	R-Gain	R-in 	X-in	SWR	NetGain
--------------------------------------------------------		
	171	8.08	256.81	-499.31	5.06	5.49
	174	9.11	231.29	-171.44	1.99	8.61
7	177	9.46	266.23	12.45	1.14	9.44
	180	9.58	336.79	130.76	1.53	9.39
8	183	9.6	438.92	194.54	1.91	9.15
	186	9.57	557.45	189.96	2.15	8.95
9	189	9.52	649.62	108.94	2.24	8.83
	192	9.45	666.92	-12.84	2.22	8.77
10	195	9.36	609.35	-111.47	2.12	8.76
	198	9.27	520.17	-155.26	1.95	8.79
11	201	9.15	435.78	-152.02	1.74	8.82
	204	9.01	370.32	-119.47	1.51	8.83
12	207	8.83	325.97	-69.96	1.27	8.77
	210	8.59	302.42	-9.27	1.03	8.59
13	213	8.24	302.07	60.93	1.22	8.2
	216	7.74	334.63	139.65	1.57	7.52
	219	7.01	424.63	216.01	1.98	6.51
.......................................................			
	467	13.37	289.63	-32.3	1.12	13.36
14	473	13.75	295.63	-57.54	1.21	13.71
15	479	14.08	282.73	-67.07	1.27	14.02
16	485	14.34	266.78	-62.95	1.29	14.27
17	491	14.5	249.68	-49.03	1.29	14.43
18	497	14.58	234.07	-20	1.3	14.51
19	503	14.65	237.36	24.94	1.29	14.58
20	509	14.76	267.13	59.15	1.27	14.7
21	515	14.82	298.55	74.65	1.28	14.75
22	521	14.82	327.56	85.09	1.33	14.73
23	527	14.8	359.82	90.91	1.39	14.68
24	533	14.77	394.49	87.42	1.45	14.62
25	539	14.73	426.14	73.11	1.5	14.55
26	545	14.7	450.03	51.13	1.53	14.5
27	551	14.68	464.88	26.17	1.56	14.47
28	557	14.67	471.88	1.89	1.57	14.45
29	563	14.67	473.1	-19.67	1.58	14.44
30	569	14.67	470.65	-37.62	1.59	14.44
31	575	14.66	466.41	-51.81	1.59	14.43
32	581	14.64	462.05	-62.58	1.59	14.41
33	587	14.59	459.22	-70.79	1.59	14.36
34	593	14.49	459.48	-78.39	1.61	14.25
35	599	14.37	462.97	-89.1	1.64	14.11
36	605	14.31	465.57	-106.52	1.68	14.02
37	611	14.37	461.02	-128.46	1.73	14.05
38	617	14.52	448.43	-148.78	1.76	14.18
39	623	14.69	431.09	-164.65	1.78	14.33
40	629	14.85	411.32	-176.26	1.8	14.48
41	635	14.99	389.92	-184.18	1.81	14.62
42	641	15.12	367.14	-188.48	1.81	14.74
43	647	15.24	343.31	-188.89	1.81	14.86
44	653	15.35	318.95	-185.04	1.81	14.97
45	659	15.46	294.81	-176.7	1.8	15.09
46	665	15.56	271.74	-163.79	1.78	15.21
47	671	15.65	250.62	-146.55	1.74	15.32
48	677	15.71	232.22	-125.47	1.71	15.4
49	683	15.72	216.93	-101.27	1.66	15.44
50	689	15.64	204.7	-74.53	1.62	15.39
51	695	15.44	195.21	-45.27	1.6	15.2
52	701	15.11	188.52	-12.94	1.6	14.87
	707	14.61	185.4	23.11	1.63	14.35



	GH8n3					
=======================			
		
Ch.	Freq 	R-Gain	R-in 	X-in	SWR	NetGain
--------------------------------------------------------		
	171	9.23	430.41	-412.05	3.13	7.89
	174	9.71	388.99	-155.22	1.68	9.42
7	177	9.81	413.22	-18.91	1.38	9.7
	180	9.77	458.11	48.68	1.56	9.56
8	183	9.67	502.68	70.44	1.73	9.35
	186	9.56	534.54	63.59	1.82	9.18
9	189	9.44	549.3	43.61	1.85	9.04
	192	9.33	549.25	22.2	1.84	8.94
10	195	9.23	539.92	5.78	1.8	8.86
	198	9.13	526.84	-3.56	1.76	8.79
11	201	9.05	514.16	-6.27	1.71	8.74
	204	8.97	504.58	-3.8	1.68	8.68
12	207	8.91	499.79	2.14	1.67	8.63
	210	8.85	500.93	9.81	1.67	8.57
13	213	8.79	508.92	17.33	1.7	8.49
	216	8.74	524.63	22.46	1.75	8.4
	219	8.69	548.74	22.09	1.83	8.3
.......................................................			
	467	13.74	185.86	-80.76	1.79	13.38
14	473	13.87	202.7	-47.29	1.55	13.67
15	479	13.97	223.22	-19.89	1.36	13.87
16	485	14.05	246.06	1	1.22	14.01
17	491	14.12	269.6	15.44	1.13	14.1
18	497	14.19	292.22	24.11	1.09	14.18
19	503	14.25	312.7	28.26	1.11	14.24
20	509	14.3	330.57	29.46	1.14	14.28
21	515	14.36	346.14	29.1	1.18	14.33
22	521	14.39	360.26	27.91	1.22	14.35
23	527	14.39	373.69	25.78	1.26	14.33
24	533	14.35	386.45	22.25	1.3	14.28
25	539	14.29	398.06	17.33	1.33	14.2
26	545	14.24	408.34	11.54	1.36	14.14
27	551	14.19	417.56	5.27	1.39	14.07
28	557	14.17	426.01	-1.42	1.42	14.04
29	563	14.15	433.86	-8.62	1.45	14
30	569	14.15	441.13	-16.42	1.47	13.99
31	575	14.15	447.81	-24.86	1.5	13.97
32	581	14.16	453.9	-33.93	1.53	13.97
33	587	14.17	459.47	-43.7	1.56	13.96
34	593	14.18	464.62	-54.32	1.58	13.95
35	599	14.18	469.4	-66.18	1.62	13.93
36	605	14.19	473.57	-79.92	1.65	13.92
37	611	14.2	476.3	-96.25	1.69	13.9
38	617	14.23	476.03	-115.37	1.73	13.91
39	623	14.29	470.97	-136.41	1.78	13.94
40	629	14.37	460.04	-157.4	1.81	13.99
41	635	14.47	443.35	-176.16	1.85	14.07
42	641	14.58	421.89	-190.98	1.87	14.16
43	647	14.7	396.98	-200.86	1.89	14.27
44	653	14.82	369.92	-205.26	1.9	14.38
45	659	14.93	341.86	-204.01	1.9	14.49
46	665	15.03	313.84	-197.08	1.88	14.6
47	671	15.11	286.81	-184.58	1.86	14.7
48	677	15.18	261.6	-166.68	1.82	14.79
49	683	15.21	239.03	-143.61	1.78	14.86
50	689	15.19	219.81	-115.68	1.72	14.88
51	695	15.1	204.61	-83.13	1.66	14.83
52	701	14.92	194.16	-46.16	1.61	14.68
	707	14.6	189.32	-4.93	1.59	14.37



	GH6n3					
=======================			
		
Ch.	Freq 	R-Gain	R-in 	X-in	SWR	NetGain
--------------------------------------------------------		
	171	9.38	582.83	-616.95	4.41	7.18
	174	9.74	461.54	-292.53	2.39	8.94
7	177	9.79	445.98	-133.7	1.71	9.48
	180	9.72	456.82	-54.85	1.56	9.51
8	183	9.6	468.65	-19.24	1.57	9.38
	186	9.47	472.55	-5.13	1.58	9.25
9	189	9.34	467.95	1.26	1.56	9.13
	192	9.22	457.92	7.52	1.53	9.03
10	195	9.11	446.1	16.9	1.49	8.94
	198	9.02	435.39	30.2	1.46	8.86
11	201	8.93	427.76	47.18	1.46	8.78
	204	8.85	424.54	67.31	1.48	8.68
12	207	8.78	426.77	90.04	1.54	8.58
	210	8.72	435.61	114.9	1.63	8.47
13	213	8.66	452.55	141.38	1.75	8.33
	216	8.6	479.81	168.67	1.89	8.16
	219	8.55	520.71	195.14	2.07	7.99
.......................................................			
	467	13.52	200.78	-56.34	1.59	13.29
14	473	13.61	229.49	-24.95	1.33	13.52
15	479	13.67	260.24	-3.96	1.15	13.65
16	485	13.71	290.04	7.25	1.04	13.71
17	491	13.74	316.27	10.59	1.07	13.74
18	497	13.76	337.53	8.85	1.13	13.74
19	503	13.78	353.89	4.73	1.18	13.75
20	509	13.8	366.43	0.14	1.22	13.76
21	515	13.8	376.55	-4.08	1.26	13.74
22	521	13.78	385.39	-7.91	1.29	13.71
23	527	13.74	393.45	-11.73	1.31	13.66
24	533	13.68	400.76	-15.69	1.34	13.59
25	539	13.61	407.43	-19.61	1.36	13.51
26	545	13.57	413.87	-23.38	1.39	13.45
27	551	13.56	420.47	-27.28	1.41	13.43
28	557	13.56	427.32	-31.71	1.44	13.42
29	563	13.58	434.31	-37.02	1.47	13.42
30	569	13.6	441.21	-43.44	1.5	13.42
31	575	13.62	447.73	-51.08	1.53	13.43
32	581	13.65	453.6	-59.98	1.56	13.44
33	587	13.68	458.55	-70.1	1.59	13.45
34	593	13.71	462.32	-81.38	1.62	13.46
35	599	13.74	464.63	-93.75	1.65	13.47
36	605	13.78	465.15	-107.07	1.68	13.49
37	611	13.84	463.5	-121.09	1.71	13.53
38	617	13.91	459.35	-135.38	1.74	13.58
39	623	13.98	452.49	-149.37	1.77	13.63
40	629	14.07	442.92	-162.43	1.8	13.7
41	635	14.16	430.81	-174	1.82	13.78
42	641	14.25	416.4	-183.6	1.83	13.86
43	647	14.35	399.98	-190.8	1.84	13.95
44	653	14.43	381.89	-195.2	1.85	14.02
45	659	14.52	362.53	-196.45	1.85	14.12
46	665	14.6	342.31	-194.2	1.84	14.2
47	671	14.67	321.76	-188.12	1.82	14.28
48	677	14.73	301.45	-177.91	1.79	14.37
49	683	14.76	282.07	-163.35	1.75	14.43
50	689	14.76	264.43	-144.25	1.68	14.47
51	695	14.72	249.42	-120.57	1.61	14.48
52	701	14.6	238.08	-92.39	1.51	14.42
	707	14.39	231.5	-59.99	1.41	14.26



	GH4n3					
=======================			
		
Ch.	Freq 	R-Gain	R-in 	X-in	SWR	NetGain
--------------------------------------------------------		
	171	9.25	480.26	-600.55	4.51	6.99
	174	9.73	400.55	-300.01	2.42	8.91
7	177	9.83	398.74	-144.31	1.65	9.56
	180	9.77	417.54	-61.69	1.45	9.62
8	183	9.66	436.98	-20.34	1.46	9.5
	186	9.53	449.47	-1.17	1.5	9.35
9	189	9.41	454.02	8.26	1.51	9.22
	192	9.29	452.94	15.31	1.51	9.11
10	195	9.18	449.35	23.56	1.51	9
	198	9.08	445.98	34.25	1.5	8.9
11	201	8.99	444.89	47.45	1.51	8.8
	204	8.91	447.57	62.71	1.54	8.71
12	207	8.84	455.24	79.38	1.6	8.6
	210	8.77	469.1	96.66	1.67	8.49
13	213	8.71	490.56	113.45	1.77	8.36
	216	8.66	521.36	128.06	1.89	8.23
	219	8.61	563.66	137.63	2.03	8.08
.......................................................			
	467	13.21	203.37	-73	1.63	12.96
14	473	13.3	240.41	-44.8	1.32	13.22
15	479	13.33	277.38	-34.12	1.15	13.31
16	485	13.32	306.54	-37.95	1.14	13.3
17	491	13.28	323.56	-48.99	1.19	13.25
18	497	13.24	329.51	-59.83	1.24	13.19
19	503	13.21	328.53	-66.33	1.26	13.15
20	509	13.18	324.75	-67.65	1.26	13.12
21	515	13.17	320.96	-64.66	1.24	13.12
22	521	13.16	318.62	-58.71	1.22	13.12
23	527	13.17	318.32	-51.05	1.19	13.14
24	533	13.19	320.23	-42.63	1.16	13.16
25	539	13.22	324.24	-34.16	1.14	13.2
26	545	13.25	330.18	-26.17	1.14	13.23
27	551	13.27	337.86	-19.04	1.14	13.25
28	557	13.29	347.04	-13.08	1.16	13.27
29	563	13.3	357.51	-8.54	1.19	13.27
30	569	13.3	369.04	-5.63	1.23	13.25
31	575	13.31	381.44	-4.52	1.27	13.25
32	581	13.3	394.48	-5.43	1.32	13.22
33	587	13.3	407.94	-8.58	1.36	13.2
34	593	13.3	421.5	-14.24	1.41	13.17
35	599	13.29	434.72	-22.69	1.46	13.14
36	605	13.29	447.01	-34.2	1.51	13.11
37	611	13.3	457.55	-48.9	1.56	13.09
38	617	13.32	465.3	-66.66	1.6	13.08
39	623	13.36	469.19	-86.93	1.65	13.09
40	629	13.4	468.29	-108.72	1.7	13.1
41	635	13.45	462.03	-130.71	1.74	13.12
42	641	13.5	450.28	-151.43	1.77	13.15
43	647	13.56	433.33	-169.43	1.8	13.19
44	653	13.61	411.91	-183.37	1.83	13.22
45	659	13.67	387.01	-192.19	1.84	13.27
46	665	13.72	359.88	-195.12	1.84	13.32
47	671	13.76	331.82	-191.75	1.83	13.37
48	677	13.79	304.15	-181.92	1.81	13.41
49	683	13.8	278.09	-165.74	1.77	13.45
50	689	13.76	254.8	-143.45	1.71	13.45
51	695	13.65	235.32	-115.4	1.64	13.39
52	701	13.46	220.71	-81.98	1.55	13.25
	707	13.12	212.03	-43.65	1.47	12.96

Model (again scroll for each model):

Code:

CM GH10n3
CE
SY radius=0.003175
SY l1=0.381
SY g1=0.019
SY z1=0.037
SY l2=0.280
SY g2=0.009
SY z2=0.137
SY l3=0.556
SY g3=0.0055
SY z3=0.262
SY l4=0.511
SY g4=4e-3
SY z4=0.414
SY l5=0.270
SY g5=4e-3
SY z5=0.592
SY x=-0.081
SY a1=0.196
SY a2=0.192
SY a3=0.190
SY b=0.120
SY feed=0.0685
SY a1ma2=a1-a2
SY a1pa2=a1+a2
SY a1ma2pa3=a1ma2+a3
SY a1pa2pa3=a1pa2+a3
SY n_top_l=0.213588
SY n_b_l=b
SY n_h=0.049287
SY n_b_z=a1pa2pa3*.7071+0.012
SY n_top_z=n_b_z+n_h
SY n_back=-0.24027
SY n_refl_size=0.402098
SY n_refl_z=0.359698
SY n_refl_size1=0.453799
GW	1	23	0	-feed	0	0	-feed-0.70711*a1	0.70711*a1	radius
GW	2	23	0	-feed-0.70711*a1	0.70711*a1	0	-feed-0.70711*a1ma2	0.70711*a1pa2	radius
GW	3	23	0	-feed-0.70711*a1ma2	0.70711*a1pa2	0	-feed-0.70711*a1ma2pa3	0.70711*a1pa2pa3	radius
GW	4	15	0	-feed-0.70711*a1ma2pa3	0.70711*a1pa2pa3	0	-feed-0.70711*a1ma2pa3-b	0.70711*a1pa2pa3	radius
GW	5	45	x	g1	z1	x	l1+g1	z1	radius
GW	6	33	x	g2	z2	x	l2+g2	z2	radius
GW	7	65	x	g3	z3	x	l3+g3	z3	radius
GW	8	61	x	g4	z4	x	l4+g4	z4	radius
GW	8	33	x	g5	z5	x	l5+g5	z5	radius
GW	13	7	0	-n_top_l	n_top_z	0	-n_top_l	n_b_z		radius
GW	14	15	0	-n_top_l	n_b_z	0	-n_top_l-n_b_l	n_b_z		radius
GX	20	010
GW	10	51	0	n_top_l	n_top_z	0	-n_top_l	n_top_z	radius
GW	15	95	n_back	-n_refl_size	n_refl_z	n_back	n_refl_size	n_refl_z	radius
GX	40	001
GW	100	17	0	-feed	0	0	feed	0	radius*.8675
GW	16	107	n_back	-n_refl_size1	0	n_back	n_refl_size1	0	radius
GE	0
LD	5	0	0	0	24900000
GN	-1
EK
EX	0	100	9	0	1			
FR	0	0	0	0	800	0
RP	0	60	73	1001	0	0	3	5		
EN

******************

CM GH8n3
CM 
CE
SY radius=0.003175
SY l1=0.314
SY g1=0.018
SY z1=0.083
SY l2=0.340
SY g2=0.0305
SY z2=0.258
SY l3=0.417
SY g3=0.0425
SY z3=0.439
SY l4=0.302
SY g4=.054
SY z4=0.620
SY x=-0.081
SY a1=0.193
SY a2=0.193
SY a3=0.183
SY b=0.121
SY feed=0.0725
SY a1ma2=a1-a2
SY a1pa2=a1+a2
SY a1ma2pa3=a1ma2+a3
SY a1pa2pa3=a1pa2+a3
SY n_top_l=0.214173
SY n_b_l=b
SY n_h=0.048478
SY n_b_z=a1pa2pa3*.7071+0.012
SY n_top_z=n_b_z+n_h
SY n_b_ystart=-n_top_l	'-feed-0.70711*a1ma2pa3-n_b_ysp
SY n_b_yend=n_b_ystart-n_b_l
SY n_back=-0.29961
SY n_refl_size=0.406309
SY n_refl_z=0.43301
SY n_refl_size1=0.479869
GW	1	23	0	-feed	0	0	-feed-0.70711*a1	0.70711*a1	radius
GW	2	23	0	-feed-0.70711*a1	0.70711*a1	0	-feed-0.70711*a1ma2	0.70711*a1pa2	radius
GW	3	21	0	-feed-0.70711*a1ma2	0.70711*a1pa2	0	-feed-0.70711*a1ma2pa3	0.70711*a1pa2pa3	radius
GW	4	15	0	-feed-0.70711*a1ma2pa3	0.70711*a1pa2pa3	0	-feed-0.70711*a1ma2pa3-b	0.70711*a1pa2pa3	radius
GW	5	37	x	g1	z1	x	l1+g1	z1	radius
GW	6	41	x	g2	z2	x	l2+g2	z2	radius
GW	7	49	x	g3	z3	x	l3+g3	z3	radius
GW	8	35	x	g4	z4	x	l4+g4	z4	radius
GW	11	7	0	-n_top_l	n_top_z	0	n_b_ystart	n_b_z	radius
GW	14	15	0	n_b_ystart	n_b_z	0	n_b_yend	n_b_z	radius
GX	20	010
GW	10	51	0	n_top_l	n_top_z	0	-n_top_l	n_top_z	radius
GW	15	95	n_back	-n_refl_size	n_refl_z	n_back	n_refl_size	n_refl_z	radius
GX	40	001
GW	100	17	0	-feed	0	0	feed	0	radius*.8675
GW	16	113	n_back	-n_refl_size1	0	n_back	n_refl_size1	0	radius
GE	0
LD	5	0	0	0	24900000
GN	-1
EK
EX	0	100	9	0	1			
FR	0	0	0	0	800	0
RP	0	60	73	1001	0	0	3	5		
EN


******************

CM GH6n3
CE
SY radius=0.003175
SY l1=0.322
SY g1=0.022
SY z1=0.085
SY l2=0.328
SY g2=0.036
SY z2=0.299
SY l3=0.312
SY g3=0.0495
SY z3=0.590
SY x=-0.075
SY a1=0.184
SY a2=0.205
SY a3=0.171
SY b=0.124
SY feed=0.0755
SY a1ma2=a1-a2
SY a1pa2=a1+a2
SY a1ma2pa3=a1ma2+a3
SY a1pa2pa3=a1pa2+a3
SY n_top_l=0.19945
SY n_b_l=b
SY n_h=0.059786
SY n_b_z=a1pa2pa3*.7071+.012
SY n_top_z=n_b_z+n_h
SY n_b_ystart=-n_top_l	'-feed-0.70711*a1ma2pa3-n_b_ysp
SY n_b_yend=n_b_ystart-n_b_l
SY n_back=-0.31458
SY n_refl_size=0.406713
SY n_refl_z=0.447715
SY n_refl_size1=0.497582
GW	1	23	0	-feed	0	0	-feed-0.70711*a1	0.70711*a1	radius
GW	2	25	0	-feed-0.70711*a1	0.70711*a1	0	-feed-0.70711*a1ma2	0.70711*a1pa2	radius
GW	3	21	0	-feed-0.70711*a1ma2	0.70711*a1pa2	0	-feed-0.70711*a1ma2pa3	0.70711*a1pa2pa3	radius
GW	4	15	0	-feed-0.70711*a1ma2pa3	0.70711*a1pa2pa3	0	-feed-0.70711*a1ma2pa3-b	0.70711*a1pa2pa3	radius
GW	5	39	x	g1	z1	x	l1+g1	z1	radius
GW	6	39	x	g2	z2	x	l2+g2	z2	radius
GW	7	37	x	g3	z3	x	l3+g3	z3	radius
GW	11	7	0	-n_top_l	n_top_z	0	n_b_ystart	n_b_z	radius
GW	14	15	0	n_b_ystart	n_b_z	0	n_b_yend	n_b_z	radius
GX	20	010
GW	10	47	0	n_top_l	n_top_z	0	-n_top_l	n_top_z	radius
GW	15	97	n_back	-n_refl_size	n_refl_z	n_back	n_refl_size	n_refl_z	radius
GX	40	001
GW	100	17	0	-feed	0	0	feed	0	radius*.8675
GW	16	117	n_back	-n_refl_size1	0	n_back	n_refl_size1	0	radius
GE	0
LD	5	0	0	0	24900000
GN	-1
EK
EX	0	100	9	0	1			
FR	0	0	0	0	800	0
RP	0	60	73	1001	0	0	3	5		
EN

******************

CM GH4n3
CE
SY radius=0.003175
SY l1=0.300
SY g1=0.018
SY z1=0.128
SY l2=0.312
SY g2=0.045
SY z2=0.444
SY x=-0.084
SY a1=0.196
SY a2=0.202
SY a3=0.171
SY b=0.122
SY feed=0.075
SY a1ma2=a1-a2
SY a1pa2=a1+a2
SY a1ma2pa3=a1ma2+a3
SY a1pa2pa3=a1pa2+a3
SY n_top_l=0.202458
SY n_b_l=b
SY n_h=0.056971
SY n_b_z=a1pa2pa3*.7071+.012
SY n_top_z=n_b_z+n_h
SY n_b_ystart=-n_top_l	'-feed-0.70711*a1ma2pa3-n_b_ysp
SY n_b_yend=n_b_ystart-n_b_l
SY n_back=-0.31613
SY n_refl_size=0.405376
SY n_refl_z=0.470002
SY n_refl_size1=0.480585
GW	1	23	0	-feed	0	0	-feed-0.70711*a1	0.70711*a1	radius
GW	2	25	0	-feed-0.70711*a1	0.70711*a1	0	-feed-0.70711*a1ma2	0.70711*a1pa2	radius
GW	3	21	0	-feed-0.70711*a1ma2	0.70711*a1pa2	0	-feed-0.70711*a1ma2pa3	0.70711*a1pa2pa3	radius
GW	4	15	0	-feed-0.70711*a1ma2pa3	0.70711*a1pa2pa3	0	-feed-0.70711*a1ma2pa3-b	0.70711*a1pa2pa3	radius
GW	5	35	x	g1	z1	x	l1+g1	z1	radius
GW	6	37	x	g2	z2	x	l2+g2	z2	radius
GW	11	7	0	-n_top_l	n_top_z	0	n_b_ystart	n_b_z	radius
GW	14	15	0	n_b_ystart	n_b_z	0	n_b_yend	n_b_z	radius
GX	20	010
GW	10	49	0	n_top_l	n_top_z	0	-n_top_l	n_top_z	radius
GW	15	95	n_back	-n_refl_size	n_refl_z	n_back	n_refl_size	n_refl_z	radius
GX	40	001
GW	100	17	0	-feed	0	0	feed	0	radius*.8675
GW	16	113	n_back	-n_refl_size1	0	n_back	n_refl_size1	0	radius
GE	0
LD	5	0	0	0	24900000
GN	-1
EK
EX	0	100	9	0	1			
FR	0	0	0	0	800	0
RP	0	60	73	1001	0	0	3	5		
EN

balm · 2010-06-14, 08:42 AM

nikiml,

nice work, looks promising!

300ohm · 2010-06-14, 11:11 AM

nikiml, you missed the other NAROD dimension in the above drawings, previously you called C.

nikiml · 2010-06-14, 11:38 AM

No, this dimension is fixed to be equal to the jeds b dimension.
see the equal segment symbols on the "drawing".

2010-06-10, 09:23 PM	#62
Autofils Join Date: Mar 2006 Location: Ottawa ON Posts: 391	Nikiml, Re your conv-test results in post #61, I sent you an email with some tips and suggested investigations you could do to obtain a better understanding of these results. I don't have the time to dig into now. I would simply point out that runs 1-0 to 1-3 are for segmentation at 10 segs per 1/2 wave to ~23 segs per 1/2 wave and the gain varies from 13.13 to 13.02. The next run 1-4 has a total seg count of 1004, which is ~31 segs per 1/2 wave. I suspect that for run1-4 (and 1-5 to 1-12), the normal 4nec2 segment check would report errors. If this turns out to be the case, conv-test results for these runs are not valid. If you have the time, it would be very interesting to investigate a manual nec file based on run 1-4 at ~ 30 segs per 1/2 wave. Does it pass the normal 4nec2 segment check with no errors at 585mHz? What about runs 1-5 to 1-12 ? Last edited by Autofils; 2010-06-10 at 09:35 PM. Reason: corrected spelling

2010-06-13, 11:25 AM	#71
Autofils Join Date: Mar 2006 Location: Ottawa ON Posts: 391	More "Food for Thought" re ConvergenceTest & AGT Adjust My thanks to 300Ohm and mclapp for their comments: Reference links: 300Ohm (post 68) --> http://www.digitalhome.ca/forum/show...1&postcount=68 mclapp (post 70) --> http://www.digitalhome.ca/forum/show...2&postcount=70 Like 300Ohm, I have used the AGT adjust "procedure" in all my past posted 4nec2 nec files. I never really questioned the effect of different wire diameter, on this AGT adjustment, because the results never indicated any significant wrong results. After recently reading j3d's post on the subject, and with a re-read Cebik's modeling guide, part 4 ( the last 2 pages under the heading "Testing Models"), I thought it would provide a better understanding of 4nec2 limitations, if jed's question could be answered with some detail and if possible some reference links. Reference link to j3d's post --> http://www.digitalhome.ca/forum/show...4&postcount=45 For me, the questions that need answers is simply this: ... in the case of the AGT adjust procedure, do two wrongs actually may a right" and along with that, do other modelers, besides DigitalHome.ca forum members, use this AGT adjust procedure? The recent posting (post 61) by Nikiml re Convergence Tests on his model, provides some "food -for-thought" and I invite comments. If the majority of readers on this subject feel, that this is somewhat of a waste of time, I will stop posting on this subject, as I have plenty of other activities to keep me busy. The only way I can "measure" the interest in this subject is from the posts that are made. Food - for - Thought Comments ========================= Nikiml's post 61, Convergence-Tests showed, that up until segment-check errors occurred, the raw gain varied from 13.13 dBi (total_segs=328) to 12.99 (total segs=1004). Reference link (post 61): --> http://www.digitalhome.ca/forum/show...9&postcount=61 By extrapolation of Nikiml's data, 21 segments per 1/2 wave raw gain would be 13.03 dBi. For this specific model, the Convergence Test reveals the "most accurate" raw gain convergence is 12.99dBi. So does the recommended model at 21 segments per 1/2 wave using the AGT adjust procedure, give a more accurate result, not only for this model, but for any model, regardless of the complexity? The beauty of the Non-adjusted AGT Test and the Convergence Test is: you can depend on the results, *regardless of antenna model differences (GH vs Yagi or BowTie etc)and the antenna's complexity.* In an effort to continue this discussion, here are some more "food-for-thought" links. 1. Arie's 4nec2 Forum The AGT is off - Why? A bug? http://fornectoo.freeforums.org/the-...-bug-t175.html Auto Segment http://fornectoo.freeforums.org/auto-segment-t189.html 2. Practical Antenna Modeling - WX7S http://www.wx7s.com/wordpress/?m=200812 3. Antenna Modeling for Radio Amateurs Here is an excellent overview about some of the history of antenna modeling with references to mathematical models. There is a brief discussion about limitations of the various modeling approaches and a summary of EZNEC and 4nec2 is included. This pdf doc was written by Steve Stearns, K6OIK. http://www.fars.k6ya.org/docs/Antenn...o-Amateurs.pdf ...Cheers

2010-06-14, 11:11 AM	#74
300ohm Veteran Join Date: Mar 2008 Location: Somewhere in Delaware on the flat side Posts: 7,012	nikiml, you missed the other NAROD dimension in the above drawings, previously you called C. __________________ My builds/plans (not the latest models) are located here.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

2010-06-11, 01:14 AM	#63
nikiml Join Date: Dec 2008 Location: mississauga, ON Posts: 721	You were exactly right. 1-0...1-4 are warning free, but the rest dont pass the segment checks

2010-06-11, 10:21 AM	#64
mclapp Join Date: Aug 2008 Location: Upstate NY Posts: 597	I was going to run that nec model (post 54) in EZNEC and run the auto seg on it but I couldn't import it in because EZNEC likes to import actual x,y,z cooridinates and those are not. EZNEC gives 2 options for segmentation conservative or minimum recommended for the frequency you select. I was curious what it would do with that model but it didn't happen. I know with other models I've adjusted the segments to 1 and then let EZNEC auto segment and it did a great job. Many times after auto segmentating certain models I had to adjust the source wire diameter very close to the size of the wires it was attached to to get the AGT back in line. I would think that is a good thing because I'm still on the fence about adjusting the source wire much to get a good AGT. I've also found that some things that EZNEC doesn't protest about 4NEC2 does like certain dia/seg ratios and dia/length ratios. I'm not very good with 4NEC2 I really only use it to get some of the nice graphs it generates.

2010-06-12, 11:09 AM	#70
mclapp Join Date: Aug 2008 Location: Upstate NY Posts: 597	Adjusting the source wire can also be accomplished by changing it's length and get the same result. In many of my models I use wires running away from the feed point like balun wires would be and then attach the source to that. By doing that it allows me to change the source wire length and it has the same effect as changing the diameter. I don't think it's any better than changing the source wire diameter except maybe it doesn't fall into some of NEC limitations but I don't think that a very small increase or decrease in wire diameter does either. I've tried it both ways and got the same results. If your segmentation is way off and you have to adjust your source wire a lot to get a good AGT then probably your convergence is bad due to the poor segmentation. In that case I would say adjusting the source wire diameter is just fooling yourself.

2010-06-14, 08:42 AM	#73
balm Veteran Join Date: Jan 2009 Location: Montreal Posts: 2,248	nikiml, nice work, looks promising!

2010-06-14, 11:38 AM	#75
nikiml Join Date: Dec 2008 Location: mississauga, ON Posts: 721	No, this dimension is fixed to be equal to the jeds b dimension. see the equal segment symbols on the "drawing".