Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
- Loopantennas
- Messages
Search
|
Re: HYBRID QUADRATURE JH-6-4 or equivalent
Hi Hisami,
?
Thanks for the report!? Your results look great.? Here is a screen shot of my nanoVNA sweep from 2 MHz to 32 MHz -- your results and mine look very similar:
?
?
A couple of suggestions which may or may not be of help:
1)? When I made my 50 Ohm to 25 Ohm transformer, I pushed all the turns close together and wound the last 3 turns on top of the first 7 turns.? Both of these actions were indented to reduce leakage inductance in this auto-transformer.? Also, as you mentioned, using a core material with a lower permeability than 75 material (like 43 for example) might help.
2)? Low leakage inductance (again) in the two common mode chokes used inside the all-pass filter networks seems to be important based on my LTSPICE analysis.? I'm not sure if the chokes you are using are good in this respect or not.? I found that the Murata parts I used (part number 50513AC ) had very low leakage inductance.? Perhaps when the parts suppliers in Japan reopen you might consider ordering these and giving them a try.
?
The OSA103 Mini LCmeter looks awesome.? I'd like to have one on my bench but I don't think we can order these in the USA.
?
Obtaining the 14.7 uH inductor by unwinding turns from a 33 uH drum core was a clever idea.? I wish I'd thought of that.
?
Thanks again.
?
Mark
AG5RT |
|
Re: HYBRID QUADRATURE JH-6-4 or equivalent
Hi Mark,
?
Thank you very much for the ideal values.
It was very helpful. I made one immediately. uploaded ?some pictures on my blog. ?
I wound the coil, measuring with a cute OSA103 Mini LCmeter.
and L3 rewound from 33uH choke.
?
The final result was well close.
L1 = 3.49 uH <-- 3.45 uH L2 = 0.269 uH <-- 0.281 uH L3 = 14.7 uH <-- 14.9 uH L4 = 1.13 uH <-- 1.14 uH ?
Capacitors were checked while soldered to the PCB.
Individual differences were large and troubling. SMDs are difficult to measure. hi C1 = 340 pF <-- 342 pF C2 = 26.9 pF <-- 28.9 pF C3 = 1470 pF <-- 1480 pF C4 = 113.4 pF <-- 122 pF ?
The high band loss and low band bump seems to be the wrong core.
50 Ohms to 25 Ohms transformer seems to be the wrong material. It seems to have been 75. ?
Still, overall satisfactory and adequate performance.
Isolation between ports is also good. ?
I will make another one after I have corrected and finished the transformer.
?
tnx & 73, Hisami 7L4IOU
|
|
Re: HYBRID QUADRATURE JH-6-4 or equivalent
It takes a separate phase coherent and stable feedline for each "loop". . . even then, it will require creating a frequency vs phase "map" file and that can be an intersting pursuit. I am pretty sure this has been done before, maybe by Appleton and Barnet in 1925. Might be worth creating an AI agent to hunt through the web accessible literature. David Martyne, about 1935, was working on wave polarization/propagation. Jim/VEZ |
|
Re: HYBRID QUADRATURE JH-6-4 or equivalent
Hi Hisami,
?
Thanks for the response.
?
On Mon, Apr 28, 2025 at 12:58 AM, hisami dejima wrote:
Yes. If the coupler were completely lossless, insertion gain would be -3.0 dB to the 0 degree and 90 degree ports (another words, power split equally to each output port). This coupler only adds 0.5 dB and 0.8 dB of loss beyond that.
I think you could implement the four inductors using T37-2 cores. You would need to match the inductances as close as possible to the ideal values. Here are the ideal values: ? L1 = 3.49 uH L2 = 0.269 uH L3 = 14.7 uH L4 = 1.13 uH ? C1 = 340 pF C2 = 26.9 pF C3 = 1470 pF C4 = 113.4 pF ? The 14.7uH inductor would require 60 turns which seems like a lot. Two inductors could be placed in series here if 60 turns is a problem.
Matching the frequency basically sets the phase shift through a section.? The gain through a section is controlled by the ratio of L to C. Hence, tuning a particular part will effect both gain and phase. I would start by getting component values as close as possible to the ideal values given above. Note that you can tweak the inductance of an inductor implemented on a T37-2 core by pushing the windings closer together or further apart. ? Mark
AG5RT
|
|
Re: HYBRID QUADRATURE JH-6-4 or equivalent
Hi Mark
?
Thank you very much for the information.
Please tell me. > Insertion Loss: 3.5 to 3.8 dB
Does the 3 in 3.5 to 3.8 dB mean 'diveded into two port'?
?
Japan will be on holiday for about a week so difficult to buy parts.
I want to build it on a T37-2 core. ?
For fine tuning, do I just match the frequencies?
example, L1A 3.3uH, L1B 330nH and C1 360pF are 4.4 MHz Or is there another guide?
?
73, Hisami 7L4IOU
|
|
Re: HYBRID QUADRATURE JH-6-4 or equivalent
Hi Piotr and all,
?
Thanks for posting the Chris Smith article and the LTSPICE analysis.? Back in the 1990s I developed a circuit very similar to that described by Chris but with one major simplification: each of the four filter sections is reduced from four reactive components to two. Utilizing this simplification, I just completed and tested the prototype circuit described here:
?
/g/loopantennas/files/AG5RT/90%20Degree%20Hybrid%20Prototype.pdf
?
The biggest change since the 1990s is the much improved availability of tight tolerance inductors and capacitors at a reasonable cost. All the LC components I used in the prototype are stocked at Mouser and vary in price from $0.08 to $0.41. All capacitors are 1% tolerance.? All inductors are 2% tolerance.??
?
I am very pleased with the measured results (see plots in the link).? Over the band 2 MHz to 32 MHz, I obtained the following:
?
Maximum phase deviation from 90 degrees:? ? 5.5 degrees
Maximum gain imbalance: 0.22 dB
Isolation between 0 and 90 degree ports: 21 dB minimum
? ? Insertion Loss: 3.5 to 3.8 dB
?
All measurements were taken using a nanoVNA.? I did some curve fitting in LabVIEW to obtain cleaner plots.? Note that no tweaking of component values was needed to obtain the measured phase deviation and balance.? Parts with the usual 5%, 10%, or 20% tolerance would have produced much worse results.? I think the parts mounted on an actual printed circuit board would yield even better results.
?
Mark
AG5RT
|
|
Re: HYBRID QUADRATURE JH-6-4 or equivalent
?
Jim
thanks for the inspiration, however, things are not that simple. A hybrid analog coupler can be placed directly at two orthogonal antennas (e.g. small loops). Behind the coupler is a chain: amplifier, cable, receiver (two-channel, to its second receiver is fed the signal from the vertical antenna and its amplifier).
Now let's follow your line of thinking: Feed the signal from two orthogonal small loops to two mixers. But how to ensure that these loops are broadband matched (across the HF range, and perhaps even lower to LF) to the impedance of the cable. ?This task is usually fulfilled by amplifiers such as the LZ1AQ, Wellbrook, etc. However, introducing amplifiers also means introducing a gain difference between them, and this already affects the accuracy of angle-of-arrival measurements. A simple analog solution allows accuracy of better than 5 degrees over a wide frequency range. I don't know if the inclusion of two amplifiers in the circuit won't cause more deviation. I once wrote to Andrew Ikin asking if he could produce two reasonably identical ALA100LN amplifiers. His answer was short: "forget it". These are my thoughts on a Sunday afternoon. ? |
|
Re: HYBRID QUADRATURE JH-6-4 or equivalent
or more precisely: /g/loopantennas/files/NX0E_hybrid_coupler/loss_precise.jpg? |
|
Re: Coupling loop advice
The unshielded coupling loop is likely to produce the lowest losses.
?
When using a toroid, especially on the lower frequency bands, a high permeability core material is required. These tend to be lossy, in addition, core saturation can also occur. This is because a small diameter core is required to achieve adequate coupling to the loop structure, and this in turn means that a very small core cross-section is carrying the majority of the flux.
?
Regards,
?
Martin
?
?
On Sun, Apr 27, 2025 at 09:05 AM, V¨ªctor J. S¨¢nchez wrote:
At first I'll made an unshielded coupling loop so it's the easiest method althougt the toroid is an excellent alternative too. |
|
Re: HYBRID QUADRATURE JH-6-4 or equivalent
Hi Hisami,
?
I've attached a new file: /g/loopantennas/files/NX0E_hybrid_coupler/loss_plot.jpg . V(n005) is the voltage at the source.
?
The losses are not 9 dB but about 3 dB.?
?
Kind regards
?
Piotr
? |
|
Re: Coupling loop advice
Yes, this would work quite well.? In technical terms, its called a Rogowski Coil.? Uncle and amateurs have fed the base of a water fountain using a Rogowski coil as well as feeding trees.? I'[ve never tried it, but the theory is sound! Dave - W?LEV On Sat, Apr 26, 2025 at 7:28?PM Dan Clementi via <dan=[email protected]> wrote:
--
Dave - W?LEV |
|
Re: Coupling loop advice
On Fri, Apr 25, 2025 at 07:25 PM, V¨ªctor J. S¨¢nchez wrote:
Hi, Victor.
?
Not to derail the discussion too much, but the toroid coupling method shown in figure 32 of that link works very well indeed.
Not only is it mechanically very simple and robust, but you can also infer something about the parasitic resistance of your main loop and interconnect to the capacitor.? I.E. once you have the turns ratio for best SWR at resonance, you can calculate the total resistance loading the secondary side of the transformer and then subtract out the radiation resistance.
?
Dan - K3GMQ |
|
Re: Coupling loop advice
Hi Dave
Snake oil..lol Just dont make any improvement..good selling point though! Simon g0zen, ex maker of 160/80m mag loops that can get into USA, Caribbean etc from UK ssb.. ( now live on Dartmoor.. as you know,( rsgb tech marconi T) no longer need a tx mag loop. Though I do still have it incase I end up end of life in a HOA.( put a bird box in it..its a modern art tree..!) Regards.Simon |
|
Re: Coupling loop advice
If you look at the physics of the "shielded" loop, you will rapidly realize it is NOT SHIELDED from the induced currents!? That's the whole reason for the gap at the top high-Z point, opposite the feed point and according to the reference diagram: "This will not work". The use of the term "shielded" certainly connotes the physical appearance of the loop, but from an EMAG standpoint, it is totally in error. Dave - W?LEV On Fri, Apr 25, 2025 at 11:25?PM V¨ªctor J. S¨¢nchez via <vjosesan=[email protected]> wrote:
--
Dave - W?LEV |
James Redding WA9VEZ
to navigate to use esc to dismiss