James,

I've just noticed this old thread and I'd like to append my own suggestions to it, hope it can still be useful. I agree that the stock Rockmite sidetone is ear-splitting. To cure the problem I added a 47uF electrolytic capacitor in parallel with the audio output jack. That works rather well and it also has the positive side-effect of increasing the receiver selectivity, much of which depends on the audio stage in a DC receiver, by cutting the annoying high-pitched hiss typical of direct conversion radios. Depending also on the actual impedance of the earphones/earbuds being used, you may want to try a lower value cap, maybe 33uF or less, until you find a value that suits your tastes. This is a trivial mod but it can improve things quite a bit.

73/72 de Carlo

I've just stumbled upon this interesting thread, so let me step-in. I agree with the two wires and tuner suggestion. I have tried all sorts of setups in the quest for the "perfect" /P antenna: resonant, non-resonant, tree-supported, kite-lifted, basically everything conceivable. Each solution has its pros and cons, of course. In the end I settled on the two-wire-and-tuner, for:

• Full-size wires: in most practical /P situations they get detuned by the surroundings: think of an hotel room, a balcony, the wire casually hoisted into a convenient tree, etc, so you'll end up needing a tuner anyway. Plus, it's not always (almost never ?) easy to accommodate the 10 meters of wire needed by a full-size dipole, neither vertically (hotel), nor horizontally (crowded beach, hotel room/balcony, etc.).
• Short whips: terribly inefficient if all you can feed them with is QRP(p). I don't like them.
• Kites: great fun, but go figure about handiness .
• Trees: rarely mother nature places a suitable one where I need it, if at all.
• Hoisting weaponry: slingshots, guns, crossbows, you name them. All but safe, clumsy to carry/use, forget them (throw-bags are better, but they still need trees: BTW, tree branches and leaves are lossy, too).

I mean, all the above can be used in specific situations where it makes sense, but none of the above can be seen as an all-purpose /P solution, if that's what the OP is asking for. Hence the two wires, both of 25-28 feet in length: one is the radiator, the other one is the (single) counterpoise. The latter doesn't even need to be laid straight, and it can usually be supplemented by a convenient metallic structure nearby, or often by nothing at all if on 20m and up, where capacitive coupling, the body of the operator and all that can provide all the "ground" that is (reasonably) needed. The radiator can be set up more-or-less vertically with a lightweight fishing pole of suitable length, so that you don't need to rely on local supports, that may not be available. The radiator does not need to be vertical, nor straight, so it can be made to lean out of an hotel balcony/window, and that. Of course you need a tuner, which needs to be small and have a built-in SWR indicator, like the fine NorCal BLT (which can handle unbalanced antennas just as well), or an even smaller homebrew one (here's an inner view of my own 20-40m L-network, which is really tiny: http://www.strozzi.it/users/carlo/hamradio/misc/epsn4945.html).

Performance-wise, I do not know of any /P antenna which is not a compromise, and resonant wires are no exception: I believe that a low dipole, albeit resonant, isn't all that more efficient than my own two wires. EFHWAs are fine, and Dale's dipoles are the best ones I know of, but try and set them up half-decently out of a hotel balcony or on a crowded beach (unless it's the 6m version, of course). Plus, I have found that over rocky soil, such as atop high mountains, you do need some sort of counterpoise, be it one (or more) separate wire or a run of feedline providing the same function.

Caveat: I'm only concerned with /P setups. Going pedestrian it's a very different problem I have no experience with.

Just my 2-cent, hope it helps.

According to Dale, the former owner of the subject End-Fedz production line, the multiband transmatch can work just as fine between abt 5 and 30 MHz, provided you fit it with the correct half-wave length of wire. The supplied trapped multiband wire is just for convenience, not to have to take the antenna down to change the band of operation. On the other hand, I have always found the bulky 40m trap to be a bit inconvenient, because it makes the wire sag under its weight, it gets caught into tree branches, and so on. So, using pre-cut wires certainly makes sense, provided you do not need to change band too quicky.

On a different note, while I'm rather happy with the performance of these End-Fedz antennas, I'm not sure they (or any other EFHWA) perform as well atop high mountains, over very rocky soil, where a dipole may be a better performer. I have no real data to back my statement, but from the very few tests that I have done I'm inclined to think that's the case.

After reading a few sources around the Net I'm not sure I'd buy the 1056. Here's one of the best references that I found (you'll have to use the google translator): http://www.stalad.it/iz4bqv/tec1056/tec1056.html. If you insist on a direct conversion receiver I would suggest you rather go for a Rockmite, which is bound to one single frequency but which may actually have got enough selectivity to prevent having 3 or 4 overlapping QSOs coming in. The Rockmite is also a transmitter, which may not interest you but it won't hurt either, in case you also want to try and join in.  An even better choice would be an SST, by Wilderness radio, which too is a transceiver, only slightly bigger than a Rockmite but with some frequency agility and a decent 2 watt RF output. These are good choices if you can limit yourself to CW, otherwise they may not suite your needs.

Let me just throw my own experience into the pool. The other day I had a contact on 20m CW with G4UZN, some 800 mi away, with the meager 500mW of my Ramsey QRP-20 TX. The nice part of it is that we were both using wire antennas. The other side was running abt 10watts, into a low wire in the back-yard, while I was using my homebrew OCF dipole at 30', so probably a slightly better setup but certainly not a killer ant. Considering that I have 65' of feedline, an AH-4 ATU and a balun, I guess that by the time those 500mW had been across all that stuff my ERP must have been significantly less than that. I gave G4UNZ a respectable 569, while I was 559 on him. So much for those who say that running less than one KW into a beam isn't much fun 

I disagree. Completely. I do have a QRO rig but I've been using it less and less, and now I'm QRP most of the time, and often rock-bound too, or VXO at most. Nevertheless, I'm making regular contacts with other QRPer, with wire antennas at both ends. Sure not 10' whips, but neither beams of sort. Over the last few days I've made several DX contacts on 20m, in the 1000-3000 Km range, with only 500 mW on my side, and a few watts at most on the other, and only one of those guys was using a yagi: all others were wire. If I wanted a sure-fire way to communicate with someone I would rather send him an e-mail, or maybe give him a phone call. Ham radio isn't much different than fishing: if I just wanted to have fish for dinner I'd better buying it from a shop.

I don't think anyone has stated the contrary here. OTOH, with power been equal, you can increase the voltage and lower the current, and that's what end-feeding is all about. So, with a properly cut EFHWA the current at the feed point is certainly very low, and that would seem to indicate that only a marginal counterpoise is needed (at least according to many). In fact, providing too much counterpoise can even be derimental, as you say. I've been myself a fan of EFHW antennas, and I have had significant success with them with my /P QRP setup. However, all my contacts were done from the large plain area where I live, which I believe has good soil dielectric properties, so those 10-20 feet of coax laying on the ground can work FB as a counterpoise on 20m and 40m. This summer I took my 20m /P station over the mountains and I was expecting great results from there during the day, but to my surprise that was not the case. It only happened two or three times, so it may just have been because of poor condx, and more attempts are in order before drawing conclusions. However, those few failures account for 100% of the times I've tried working that way. Now, could it be that the rocky ground at those heights makes an EFHWA less suitable up there ? Just a doubt, but I would like to have the opinion of others about that.

The weather has finally improved here and I was able to do the suggested A/B comparison on RX between the half-size and the full-size wire. I did the test by ear, since my 20m QRP radio does not have an S-meter. Also, the evaluation was done based solely on the band noise, not on actual signals. So, all the rest being equal, including the transmatch  and the sloping of the two wires, I'd say that the loaded half-size wire is 1 S unit down over its full-size counterpart, that is a good 6 dB. Now, whether that can be taken also as an indication of the relative efficiency of the two solutions on TX I cannot really tell.

As anticipated, I have done preliminary tests with the 40m version of my shortened EFHWA, and the SWR 2:1 bandwidth of the antenna is 70 KHz, a value which seems consistent with the 120 KHz that I get from the 20m version. I'll post more results later.

One thing that may matter, and that I forgot to mention up to now, is that for the antenna to be really portable I used a (very sturdy) AWG #26 PVC-insulated wire for the radiator. The very thin section of the wire may account for the narrower bandwitdh. A thin radiator may be considered lossy by some, but in actual fact, at least on the lower half of HF and at QRP power levels, that would seem not to matter much, as also debated here: http://www.eham.net/articles/10959 .

Carlo

Andrew,

I also tried to model the antenna beforehand, but without much success (but I have very little experience with numerical modelling). I'm running Linux here and I can only read *.NEC files. Can you please export your model into a NEC file and either post it here or send it to me to iz4kbs(at)gmail.com ? My model was highly influenced by the n. of segments used to model the wire, with 9 segments being the one that produced the most meaningful results, but as I said it may just be me who does not know how to do proper modelling. In the end I gave up with modelling and decided to spend the same amount of time actually building and measuring the damn thing


That's the same impression that I had. Mine was set up in the open, sloping at 35-40° from a fiberglass pole. The matching network and the first 3' of feedline were also sloping along the same straight line as the radiator, with the remaining 3' of feedline more or less touching the ground (the transmatch was hanging at about 1mt from the ground). During the adjustment of the capacitor I tried several times to see if connecting a 17' counterpoise, also laying on the ground but running in the opposite direction from the antenna, would significantly affect the adjusted network, but with the transmatch adjusted for minimum SWR the additional counterpoise didn't influence it much. The adjustments were done with the help of a VK5JST antenna analyzer, and with the body of the rig attached to the system at the analyzer side, to simulate a realistic setup in the field (I'm only interested in /P use of this antenna). The relative independence from the counterpoise would seem to indicate that the short feedline, the mass of the rig and the wire of the earphones attached to it were probably providing all the capacitance to ground that was needed for the system to be independent from extra counterpoise wires.

Yes, an A/B comparison on RX with weak signals is the only significant (and doable) test that I still need to do, as soon as the whether improves here. As a reference antenna I'm going to use the PAR EF-20, to make sure the comparison is done against a well-reputed full-size EFHWA. Unfortunately my QRP rigs do not have an S-meter, but I may be able to detect the difference by ear, to some extent. I agree with you that the too-good-to-be-true properties that I experienced would seem to account for losses in the system, but one thing that makes me be relatively optimistic is that the tuning of the transmatch is extremely sharp, and if I move the capacitor almost by nothing out of the best match point, the SWR climbs up to over 10 within only a few degrees of a turn. This is to the point that in future versions I'll probably use a parallel of a fixed and a lower value variable capacitors, to make the tuning easier. Also, if a keep my body within 2-3 feet from the hi-Z part of the system, the SWR also increases significantly. But we'll see, I'll post the RX results here as soon as I have them available.

In fact, after several days of experimenting, I think I have managed to have a shortened 20m EFHWA antenna that really works, and without needing a field-adjustable tuner. It was a matter of figuring-out the correct parameters for the L-network, and the results are quite good:

• * The radiator is only about 5mt long, with a j1000 loading coil in the middle. I have not tested the antenna in the field yet, but I expect its efficiency to be not worse than the one of any other shortened EFHWA, i.e. possibly 1dB down over a full-length radiator
• * I managed to get an SWR of less than 2.0 over the 14.000-14.120 KHz range, with a minumun of 1.0 @ 14.060. These values seem to keep relatively stable even if the radiator and the transmatch are moved around, lowered, made more horizontal, more vertical, etc., so I think the antenna is well suitable for /P work, where one cannot anticipate how it will be set up.
• * Adding a 5mt counterpoise does not seem to influence the SWR, or very little, which I think its a sign that the antenna has already got enough ground of its own (the short feedline, the case of the rig, the body of the operator, etc.)
• * Touching the Hi-Z section of the wire with one hand makes the SWR jump to over 10:1, as expected, so at least the antenna does not seem to behave like a dummy load :-)

I'm very happy with the results on 20m, so I'll now try and make a similar antenna also for 40m, which I expect to be less critical. I'll keep the thread posted.

Andrew,

I've spent the most part of the last few days playing with both full-size and shortened EFHW antennas, and I got the impression that a shortened end-fed radiator turns the feedline into the missing part of the antenna, and the higher the frequency the more it happens, no matter what I did. Namely:

• * The L-Network isn't the one by PAR, it's homemade but all measurements would seem to indicate that it is built correctly, and it was adjusted beforehand on a resistive load, as per AA5TB suggestions.
• * The center loading coil is almost exactly j1000, as per your suggestion.
• * The radiator was properly resonated, as per Allison's suggestion, with the help of a grid-dip meter loosely coupled to the center loading coil and a monitor receiver placed a few meters away from the setup.
• * The feedline is intentionally short: 6 feet on 20m and 10 feet on 40m, that is 1.5-2 times the minimum of 0.05 WL indicated by AA5TB.
• * The radiators were set up as slopers, hanging from a 31ft Jackite fiberglass pole.
• * My Antenna analyzer is a VK5JST, which has a good reputation but is limited to 500+/-j500 (it displays ">500" for values higher than that).
• * Replacing my L-Network with the one by PAR gave similar results on 20m, while I haven't got a PAR thansmatch for 40m.

All that being said, the best SWR I get is 3:1 with the 40m transmatch+antenna, and 6:1 with the 20m transmatch+antenna. Adding some 17 feet of extra counterpoise lowers the SWR to 4:1 on 20m, and does not seem to have a major influence on 40m (maybe a longer counterpoise is needed on this band to get some effect, I've not tried). With the full-length radiators I get a perfect 1:1 match on both bands, not much influenced by the presence/absence of the extra counterpoise. 


I think you should try and measure things with only a short run of feedline. Then you can add a counterpoise and see if that increases the sensitivity of the antenna on receive (increased band noise). If that happens to you, and indeed it happens to me, that's a sign that your shortened radiator only works because of the presence of the feedline, while a full length radiator, fed with a properly designed/adjusted transmatch, tends to decouple itself from the feedline and be relatively more independent from it.

Carlo

Allison,

indeed your guess is correct. I tested the wire with a grid-dip meter and radiator was not resonant. Today I'll dig further into that.


That's for sure, and all my tests seem to confirm that. As I said, today I'll proceed to more adjusting/trimming/fiddling, etc. To try and bring the radiator to resonance would you suggest that I do that with the wire connected to the L-network or not ? Currently I'm doing it with the wire connected to the L-network, and with the feedline side of the latter closed on a 50 ohm resistor, so that the wire is cut to freq in realistic conditions, but what you think ? Should I rather tune the wire stand-alone ? The L-network has already been adjusted stand-alone with the method recommeded by AA5TB at http://www.aa5tb.com/efha_wrk.html#Method, and I can confirm that that method produces repeatable results.

Thanks,

Carlo