The small SWL beam (1...30 MHz)

Use this forum to discuss any radio-related topic outside of asking for advice (check out our Questions and Advice forum for those topics). As with all of our forums, treat others with kindness and respect.
13dka
Posts: 98
Joined: Sat May 28, 2022 4:27 am
Location: On or near a dike

Re: The small SWL beam (1...30 MHz)

Post by 13dka »

That's shortwave for ya: Conditions started out disturbed but OK-ish yesterday night until local midnight (UTC+2) and then went downhill so much that even the FT8 signals almost disappeared on 17m (rarely seen that happening) and the entire US slowly faded out on 20 towards sunrise, instead of the West Coast fading in, not to mention VK and ZL in the morning. The MUF must have dropped really low. Meanwhile, the solar indices started looking normal again this morning but shortwave was kaput anyway, I guess because it sensed that I was really exited about trying something. :mrgreen:

Results:

Besides hoping for the SDL being useful for me personally, even in the ideal environment at "the dike" I see it as a more than interesting alternative to the active small magnetic loops the SDL could be confused with. That's why I really wanted to verify/document/demonstrate this specifically, because the SML is certainly one of the most used receiving antennas of them all and people could just put an SDL where their SMLs live now, and hope for some noticeable improvement (to put this very careful).

Like I wrote above, I fully expected the outcome but ended up a little bit shocked about the difference at times. First off, the YouLoop with its 2 turns runs circles around and then wipes the floor with the SDL on MW. It's known to be good on MW, even in its passive form. But then again, the 2 sharp nulls of the YL often don't help that much disentangling multiple stations/countries using the same frequencies, and the SDL seems to have an edge there, at least when you sit right between the wanted and the unwanted station.



The sensitivity break-even point seems to be around 2 MHz and from 5 MHz upwards it goes downhill for the YL. AFAIK this is true for so many SMLs, some Wellbrook models are said to be better tuned for shortwave. But still, they're SMLs sharing the same properties the SDL is trying to improve upon. Here is Singapore VOLMET 6676 kHz:



Moving on up to 7 MHz, here's a Romanian station in QSO with a Japanese one on 41m:



Gander Radio on 13270 kHz. This is one of the examples I found rather shocking because Gander Radio is not anything extremely weak or distant. A likely explanation could be that far above 10 MHz it's not only the higher takeoff angle but also a plain lack of sensitivity putting the SML so far behind:



While we're at it, here's the pretty faint Tokyo VOLMET next door:



Next in line is a Japanese ham on 20m:



In lieu of too weak SSB stations that would just show a simple "on/off-effect" when switching antennas anyway, heres the FT8-spot on 15m:



I save you the 11m highbanders recorded around 27,555 kHz as you certainly got the idea by now. :) Also I almost forgot that I actually got a serviceable signal from RNZ in the evening. What I find interesting is how the signal appears much more unstable and fluttery on the YouLoop. Assuming the signal has wildly scattered components with a varying number of hops and incidence angles, I imagine it misses the main component arriving flat and mostly picks up the scattered leftovers, while the SDL is just confirming the simulation and is sensitive at angles where the signal has its most energy:



As mentioned in the introduction (and also that's why I mentioned it), the condx went downhill from there -- they always do to some degree after midnight and RNZ becomes really weak after midnight but will return with a vengeance after sunrise (likely on the long path then), but not today, this morning RNZ was a faint line on the waterfall for the most part. The more surprising I found that I heard ZKAK's (New Zealand VOLMET) monotonous voice again, and not the weakest signal from that station ever:



As you can see, the improvement over the amplified YouLoop is pretty considerable, I may repeat that test with the ML-200 at some point to be absolutely sure I didn't miss anything here. For the duration of the comparison (IOW until shortly after midnoght when the stupid preamp broke) I used the SDL with collapsed mast to get it rougly as low as the YouLoop, and swept both around where necessary.

It could be questioned if the YL+LANA HF combination is really equivalent to a 400 bucks SML with a bigger loop. Alas I never tried a Wellbrook but I chose the YL+LANA because it's so lightweight and portable while performing unsurprisingly similar to my 400 bucks Bonito ML-200 with a rigid 80cm loop. However, there's a broad agreement that SMLs all perform very similar. Even if I mentally interpolate an SML with noticeably better sensitivity on shortwave, there would still be the issue with the comparatively low level from signals with a low angle of incidence and therefore lower SNR.

Soooo....does that all look good to you? :D

Edit: Added JIA video
User avatar
Andrew (grayhat)
Posts: 210
Joined: Sat May 28, 2022 5:56 am
Location: JN63pn

Re: The small SWL beam (1...30 MHz)

Post by Andrew (grayhat) »

First, I won't thank you enough for your efforts in testing the SDL

That said, judging from the tests, the SDL seems to do what it was designed to, that is, offer good performance and directivity over the SW range, and I think this is a good thing

As for MW and lower frequencies; let's start from the NEC model of the SDL, start a sweep, set the range to 0.5 to 30.5 MHz and the step to 0.2, run the sweep and at end look at the gain curve

The point is that going down in frequency the SDL gain drops with a quite steep curve, now the first workaround which came to my mind was a higher gain preamp, but I'm not convinced it may solve the problem... the second idea I had is adding an inductor in series with the resistor , so I'll need to run some NEC simulations and if the inductor idea will show promise, I'll try finding the optimal positioning and value for the inductor

we'll see :D
13dka
Posts: 98
Joined: Sat May 28, 2022 4:27 am
Location: On or near a dike

Re: The small SWL beam (1...30 MHz)

Post by 13dka »

Andrew (grayhat) wrote: Sat Jul 30, 2022 12:51 pm The point is that going down in frequency the SDL gain drops with a quite steep curve, now the first workaround which came to my mind was a higher gain preamp, but I'm not convinced it may solve the problem
The practical behavior reflects the simulation to the i-dot in this regards and I'm entirely sure that more gain will not help, simple telling from the response to different gain from the LANA HF (remember, I can switch between full and less gain via Vss) not doing anything except shifting the same SNR (or lack thereof) to a different signal level.
Andrew (grayhat) wrote: Sat Jul 30, 2022 12:51 pm ... the second idea I had is adding an inductor in series with the resistor , so I'll need to run some NEC simulations and if the inductor idea will show promise, I'll try finding the optimal positioning and value for the inductor
we'll see :D
It would sure be cool being able to have all the goodness below 1 MHz too, even if that means having to switch something in and out but OTOH I'm already happy about that it kind of starts working around 500kHz, the upper half of MW is already pretty serviceable so I could live with that.

Edit: Hmmm or an inductor that has by accident a DC-R of 530 Ohms?
User avatar
Andrew (grayhat)
Posts: 210
Joined: Sat May 28, 2022 5:56 am
Location: JN63pn

Re: The small SWL beam (1...30 MHz)

Post by Andrew (grayhat) »

I think that a separate inductor would be better, see, didn't simulate it (yet) but if it will work, my bet is that it may change the pattern at higher frequencies, if that's the case, then an inductor which may be switched in or bypassed would allow to extend the usable range of the SDL
User avatar
Andrew (grayhat)
Posts: 210
Joined: Sat May 28, 2022 5:56 am
Location: JN63pn

Re: The small SWL beam (1...30 MHz)

Post by Andrew (grayhat) »

as for NEC inductance, I don't have NEC at the moment, but if you want, here's how to add it

http://www.antentop.org/w4rnl.001/loadtl2.html

it's a matter of adding an LD tag line similar to the one used for the resistor, a starting value of 1uH (1.00e-06) could be ok for testing, then the NEC optimizer will help optimizing the value to achieve max gain at the desired frequency
User avatar
Andrew (grayhat)
Posts: 210
Joined: Sat May 28, 2022 5:56 am
Location: JN63pn

Re: The small SWL beam (1...30 MHz)

Post by Andrew (grayhat) »

Ok, gave a spin to the inductor idea, after fiddling with values and positioning I came to the design in the following NEC model

Code: Select all


CM File: SDL.nec
CM
CM Small Directional Loop antenna
CM feed using a 9:1 balun transformer
CM place at 3m from ground to avoid
CM pattern distortion, optionally add
CM a 20...30dB preamplifier
CM
CE

' symbols definition
SY freq=1.5                 ' test frequency
SY hght=3                   ' height of bottom corner from ground
SY side=0.762               ' length of one side
SY diag=(sqr(2)*(side/2))   ' half diagonal
SY wire=0.00125             ' wire radius
SY vres=530                 ' loading resistor value
SY segm=13                  ' number of segment in wires
SY segs=5                   ' short wires segments
SY wfed=1                   ' feedpoint wire
SY sfed=segm                ' feedpoint segment
SY wres=4                   ' wire hosting the resistor
SY sres=1                   ' segment hosting the resistor
SY wind=3
SY sind=2
SY vind=900.00e-06

SY spac=(wire*4)            ' spacing to simulate coax routing


' wires geometry
'  ID seg    x0     y0   z0         x1     y1     z1          wire rad
GW  1 segm     0     0  hght      -diag     0   hght+diag       wire
GW  2 segm     0     0  hght       diag     0   hght+diag       wire
GW  3 segm -diag     0  hght+diag     0     0   hght+(diag*2)   wire
GW  4 segm  diag     0  hght+diag     0     0   hght+(diag*2)   wire

' coax feeder "simulation"
GW 20 segs -diag  spac  hght+diag -diag  spac   hght-diag       wire*2
GW 21 segs -diag  spac  hght-diag     0  spac   hght-diag       wire*2
GW 22 segs     0  spac  hght-diag     0  spac      wire*4       wire*2

' ground parameters
GE  1
GN  2  0  0  0  13  0.005

' wires loading
LD  7    0     0    0 2.1 wire    ' insulation
LD  5    0     0    0 58000000    ' wire
LD  1 wres  sres sres vres   0    ' resistor
'LD  0 wind  sind sind    0 vind 0

' enable extended kernel for calc
EK

' feedpoint
EX  0  wfed sfed 0 1 0 0

' initial test frequency
FR  0  1  0  0  freq  0

' end
EN

in the model the inductor is commented, so the antenna is the "standard" one, willing to insert it, just uncomment the "LD 0 wind..." line and reload the model, the inductor value is 900uH, the result is below, this first image shows the "standard" antenna pattern at 1.5 MHz and the gain curve between 0.5 and 5 MHz

Image

and that's what we already saw, adding a 900uH inductor right "above" the feedpoint, will change the antenna pattern and gain as shown in the image below

Image

as you can see, there's a little bit more gain, but the pattern is heavily changed and while retaining a low angle and good F/B the "null" is almost gone, also, if we then look at the impedance matching we'll see that it's totally "off" since, adding the inductor, the reactance skyrockets, now, while the gain is SLIGHTLY increased, I'm not sure that adding an inductance may offer advantages; first of all the gain boost is minimal, next, one will need to add a switch to bypass the inductance, otherwise the pattern at higher frequencies will be totally changed

Bottom line, not a good idea imHo, but I think that all in all, we may still be happy with the performances the SDL offers on the SW bands and, partially, on the MW too
13dka
Posts: 98
Joined: Sat May 28, 2022 4:27 am
Location: On or near a dike

Re: The small SWL beam (1...30 MHz)

Post by 13dka »

Andrew (grayhat) wrote: Sat Jul 30, 2022 4:48 pm as for NEC inductance, I don't have NEC at the moment, but if you want, here's how to add it [...] it's a matter of adding an LD tag line similar to the one used for the resistor, a starting value of 1uH (1.00e-06) could be ok for testing, then the NEC optimizer will help optimizing the value to achieve max gain at the desired frequency
Unfortunately I struggle with the complexity of 4NEC2 and being a complete math failure my view gets blurry when confronted with more than 2 numbers. :)

Andrew (grayhat) wrote: Mon Aug 01, 2022 7:57 am Ok, gave a spin to the inductor idea, after fiddling with values and positioning I came to the design in the following NEC model
Phew! :)
Andrew (grayhat) wrote: Mon Aug 01, 2022 7:57 amas you can see, there's a little bit more gain, but the pattern is heavily changed and while retaining a low angle and good F/B the "null" is almost gone, also, if we then look at the impedance matching we'll see that it's totally "off" since, adding the inductor, the reactance skyrockets, now, while the gain is SLIGHTLY increased, I'm not sure that adding an inductance may offer advantages; first of all the gain boost is minimal, next, one will need to add a switch to bypass the inductance, otherwise the pattern at higher frequencies will be totally changed

Bottom line, not a good idea imHo, but I think that all in all, we may still be happy with the performances the SDL offers on the SW bands and, partially, on the MW too
Yeah, looks like this doesn't help that much, particularly down at 500kHz and below. Well, there are so many options for VLF/MW, many people already have one of them in form of an SML, but I suspect none of those could beat the SDL on shortwave. :) Now that I have a few surplus resistors in the drawer I could try terminating the 1000m fence I tried 2020/21, to make that unidirectional (it's pointing at Alaska but picked up some Midwest and East Coast just fine). :mrgreen:
User avatar
Andrew (grayhat)
Posts: 210
Joined: Sat May 28, 2022 5:56 am
Location: JN63pn

Re: The small SWL beam (1...30 MHz)

Post by Andrew (grayhat) »

13dka wrote: Mon Aug 01, 2022 12:26 pm Yeah, looks like this doesn't help that much, particularly down at 500kHz and below. Well, there are so many options for VLF/MW, many people already have one of them in form of an SML, but I suspect none of those could beat the SDL on shortwave. :) Now that I have a few surplus resistors in the drawer I could try terminating the 1000m fence I tried 2020/21, to make that unidirectional (it's pointing at Alaska but picked up some Midwest and East Coast just fine). :mrgreen:
check this one

http://www.kk5jy.net/lla/

:D

as for the fence, assuming it's iron or the like, given the length you should consider that it will already have quite some resistance; it would be interesting to try turning it into a beverage antenna, that would require a resistor termination and quite a bunch of radials, but may offer interesting results
13dka
Posts: 98
Joined: Sat May 28, 2022 4:27 am
Location: On or near a dike

Re: The small SWL beam (1...30 MHz)

Post by 13dka »

Andrew (grayhat) wrote: Mon Aug 01, 2022 12:32 pm check this one

http://www.kk5jy.net/lla/

:D

as for the fence, assuming it's iron or the like, given the length you should consider that it will already have quite some resistance; it would be interesting to try turning it into a beverage antenna, that would require a resistor termination and quite a bunch of radials, but may offer interesting results
The fence is a wire mesh fence with 2 fat wires running on the top, I have tried it semi-successfully 3 times, faintly picked up Faroer Islands 100kW MW station (1700 something kilometers) during daytime, RNZI on 6170 kHz on another day (no idea if I had heard that better with some other antenna but NZL on 49m was a tough customer here) and used it for MW DX one night. It didn't feel better than the ML-200, I forgot to bring a balun but it generally worked and brought in a good recording of a 1kW station in Cedar Falls, IA (KCNZ). :)

I have a recording of the RNZI reception (wire with alligator clip to the fence wire, the other end stuffed into the thread where the car antenna goes, yes I have no mercy):

https://www.youtube.com/watch?v=Fi6DiJjs_VA

This is a LPAM station in in the Midlands, received with another (shorter) sheep fence at the dike that's oriented broadside (huh?) to England:

https://www.youtube.com/watch?v=1UvbIfAHTRw

We know that you can turn everything into an antenna. But without local QRM and the help of an ocean you can obviously turn everything into a good antenna***. :) But I digress:

I'm just back home from testing the SDL against the LoG, deliberately at a spot that's a few km off the coast, not free of local QRM but providing enough free space to not disturb the antennas. I aborted the test after the NE short path DX faded out because the main QRM sources are NW of the antennas (so not much fun turn the antenna towards the US) and also it was getting f...ing cold after sunset, not to mention the A index at 10 being a bit discouraging.

I wanted to do this test because what I wrote and marked with *** above - this location is reflecting the reality of listeners a bit more, no perfect ground, QRM sources, ocean too far away to help with gain. If you reckon that this is a disadvantage for the LoG, that's probably both right and wrong - it loses up to 10dB of level there, but so does the SDL, which admittedly has more spare gain than the LoG. That difference was expected and what I wanted to verify. But we know that the (big) LoG is the SNR king and it shows! It was oriented N-S and had the QRM source in its lobe as well and it had a better SNR whenever the SDL was pointing too much towards the QRM. If it wasn't, the SDL unambiguously won though, and had the QRM not been there it would have won in pretty much all situations... because the LoG needs the beach more than the SDL to perform great.

Here's an example with the same Japanese station from the last test: First with the SDL pointing due north like the LoG (the wide horizontal lobe of the SDL should not get much less level if it's not directly pointing towards Japan, but it gets a lot from the QRM from the NW):



This is after turning the SDL to the NE, with the left side of the lobe not covering the QRM source that much - you can tell that immediately by the darker waterfall background:



I'd love to post more here right now but unfortunately the Telekom / T-Mobile has currently a major slowdown+outage problem here and uploading videos takes like an hour each, if it even happens at all. Grrr.

Anyway, with the results from this evening I feel confident enough to say that this antenna delivers what the simulation promises, it's a (very) worthwhile project to build,and - here's the big point - (much) more than a yet another "compromise" antenna, and this is likely a hefty understatement. This is a pretty big achievement, Andrew! Now I wish there would be more people building this now to add their own data points to the discussion!

Edit: Here's another picture showing how the drawing below looks in RL:

SDL-Bazinga_total_APT.jpg
SDL-Bazinga_total_APT.jpg (72.27 KiB) Viewed 1128 times
User avatar
Andrew (grayhat)
Posts: 210
Joined: Sat May 28, 2022 5:56 am
Location: JN63pn

Re: The small SWL beam (1...30 MHz)

Post by Andrew (grayhat) »

13dka wrote: Tue Aug 02, 2022 12:50 am Anyway, with the results from this evening I feel confident enough to say that this antenna delivers what the simulation promises, it's a (very) worthwhile project to build,and - here's the big point - (much) more than a yet another "compromise" antenna, and this is likely a hefty understatement. This is a pretty big achievement, Andrew! Now I wish there would be more people building this now to add their own data points to the discussion!
At this point, given the results, I agree with you, the SDL is definitely a good "weapon"; sure, the LoG rules when it comes to cutting off noise, but then it's way bigger and I think that the small SDL defends itself pretty well, plus its reduced size would allow people with small space, or restrictions, to use it :D ! As for more people... <sigh> my wrist is still "out of service" and will probably remain so till next September, otherwise I'd go on, build an SDL and test it here; but there's still hope I think, maybe some other folks here, on the forum, will decide to give the SDL a spin and offer some reports, that would help having a full assessment, although given your results (at home, at "the dike "and at the other "less ideal" place) I'm very confident that the little SDL will deliver good results

For the ones willing to try a build, here's the overall view of the SDL setup
SDL_struct.png
SDL_struct.png (38.09 KiB) Viewed 1143 times
the easiest way to build it will be using off-the-shelf components, for example, the 9:1 can be the NooElec one, while the preamp may either be from NooElec or any other commercial preamp covering the 0.5...30MHz range, the remaining components, with the exception of the snap-on ferrites and coaxial cable, may be found at any local home improvement stores (or the like)

The current NEC model for the antenna can be found below

Code: Select all


CM ----------------------------------
CM File: SULA.nec
CM ----------------------------------
CM
CM Small Unidirectional Loop antenna
CM
CM feed using a 9:1 balun transformer
CM place at 3mtr from ground to avoid
CM pattern distortion, optionally add
CM a 20...30dB preamplifier (LNA)
CM
CE

' symbols definition
SY freq=7.100               ' test frequency
SY hght=3                   ' height of bottom corner from ground
SY side=0.762               ' length of one side
SY diag=(sqr(2)*(side/2))   ' half diagonal
SY wire=0.00125             ' wire radius
SY vres=530                 ' loading resistor value
SY segm=13                  ' number of segment in wires
SY segs=5                   ' short wires segments
SY wfed=1                   ' feedpoint wire
SY sfed=segm                ' feedpoint segment
SY wres=4                   ' wire hosting the resistor
SY sres=1                   ' segment hosting the resistor
SY drop=hght-diag           ' coax drop section length
SY spac=(wire*5)            ' spacing to simulate coax routing
SY dist=diag+spac           ' spacing for coax feedpoint
SY coax=(wire*2)            ' simulated coax radius


' wires geometry
'  ID seg    x0     y0   z0         x1     y1     z1          wire rad
GW  1 segm     0     0  hght      -diag     0   hght+diag       wire
GW  2 segm     0     0  hght       diag     0   hght+diag       wire
GW  3 segm -diag     0  hght+diag     0     0   hght+(diag*2)   wire
GW  4 segm  diag     0  hght+diag     0     0   hght+(diag*2)   wire

' coax feeder "simulation"
GW 20 segs -dist     0  hght+diag -dist     0   drop            coax
GW 21 segs -dist     0       drop     0     0   drop            coax
GW 22 segs     0     0       drop     0     0   spac            coax

' ground parameters
GE  1
GN  2  0  0  0  13  0.005

' wires loading
LD  7    0     0    0 2.1 wire    ' insulation
LD  5    0     0    0 58000000    ' wire
LD  1 wres  sres sres vres   0    ' resistor

' enable extended kernel for calc
EK

' feedpoint
EX  0  wfed sfed 0 1 0 0

' initial test frequency
FR  0  1  0  0  freq  0

' end
EN

to use it, you'll need a copy of the 4NEC2 software which will allow to load the above model and to run whatever simulation
Post Reply