Anyone who wants to listen to shortwave radio, or monitor high frequency transmissions, would have met the dilemma of trying to get a decently performing antenna. As it stands, most smaller radios with inbuilt telescopic antennas rarely reach more than 1m in length, and they tend to be inadequate in marginal to poor signal areas such as Australia. Worse still, they are more than capable of picking up local interference with little hope of grabbing enough signal to overcome the local noise – poor signal to noise ratios are the result.
A very simple and inexpensive antenna is known as a longwire, and as the name suggests, typically consists of a random length of long wire strewn about outside where the signal is better, being remote from the local interference. Getting a longwire to work is, however, not necessarily as simple as hooking it up.
For small portable receivers, some success can be had by connecting the longwire to the telescopic whip, although often this results in overload (i.e. signals which are too strong) resulting in distortion and signal images (finding signals at frequencies where they shouldn’t be). Some better portable receivers offer switchable sensitivity (DX or LOCAL switch) to cut down on the internal amplification (which amplifies noise as much as signals) to try and avoid this. Others offer a less sensitive external antenna jack which is often just a mono 3.5mm jack intended for connection to a longwire.
But better receivers generally don’t have such tacky methods of interfacing to an antenna. Proper RF equipment almost certainly uses a 50 ohm coax connector of some description, be it a PL-259, N-type, BNC or SMA connector. These connections are optimized for use with 50 ohm impedance sources. The problem is that longwire antennas are not 50 ohms impedance, and direct connection typically causes marginal performance due to signal losses in impedance mismatch as well as a secondary signal degradation by having the antenna itself consisting of all segments including those close to the receiver (where the noise generally is).
A way around this is to use an impedance transformer known as a longwire adapter or a longwire balun, which is typically a 9:1 impedance transformer (450 ohms to 50 ohms). Such an adapter connects to a longwire antenna (often just one connection, or one connection plus ground for better units) and offers a coax connector on the other end. This allows you to use a shielded coax lead-in that prevents local noise from getting into the antenna and allows for a better impedance match (although imperfect due to random wire length) to the longwire antenna, typically boosting signal by about 6dB.
Such adapters can be home made, but the hassle of getting the right sort of ferrite beads, connectors, and then soldering/wiring it up yourself without knowing if it’s quite right is really an inconvenience I’d rather do without. As a result, I ended up purchasing commercially made adapters for my own usage.
In this post, we will be comparing a low cost ~AU$20 generic unbranded Chinese longwire adapter with a five year old Australian WiNRADiO branded WR-LWA-0130 ~AU$55 adapter to see how well they are constructed and how well they perform with the aid of the Tektronix RSA306 Real-Time Spectrum Analyzer.
The Generic Adapter
Can you say cheap? This adapter cost just under AU$20 posted from China, and is as generic as can be.
It sits in a beige box with no branding whatsoever. A BNC connector adorns one side, with a binding post/banana plug combo jack on the other. That’s literally all there is to it.
It doesn’t take much to open it – just a quick pry at the slots and we’re in. Inside, there is no shielding whatsoever, nor any PCBs. Just a toroid with enameled copper wire windings soldered directly to the connectors, with the toroid left hanging unsupported. I’m surprised it got here without a wire detaching or the ferrite toroid being smashed into dust.
The banana jack wasn’t very well secured either – using the binding post feature could end up spinning the solder tag inside when tightened up, causing wire fatigue and eventual wire breakage.
Given the questionable looks, I wonder how it performs. The eBay listing itself stole images from I6IBE’s website, so I suppose the design itself is probably okay, and the performance will come down to construction and component choices.
WiNRADiO WR-LWA-0130 Longwire Adapter
The other contender in the head-to-head is an old reliable. This adapter has been with me for five years now, and has had a hard life. At one stage, it spent over a year outside getting alternately cooked and rained on, and it does show. But it’s always been there for me, and worked relatively well, and at a price of AU$55, it was acceptably priced.
When you look at the unit, you know it’s something proper. It has a cast metal exterior with a logo unmistakably moulded on it. The all metal exterior acts as an excellent shield.
While there’s some signs of rust, there is clear evidence of sealing around the edges to prevent water ingress. Removing the screws allows us to look inside.
It’s clear that no water has made it inside, which is a relief. The adapter itself has a PCB and seems decently supported. A soldering iron was required to de-solder the tags on the wire side, and then the BNC connector retention nut undone, to allow the PCB to be moved out of the enclosure.
The single sided PCB has two toroids on formers on the top, and thus is likely to use a different, seemingly more refined design. The manufacturing looks spot-on, with no rough edges. A very professional product.
Even the inside of the casing had silicone applied around the BNC plug entrance to ensure water would not seep past the edges. This attention to detail is what you get from a proper product from a company that knows what they’re doing.
A head to head comparison was done by using my outdoor 15m longwire, connected to one adapter at a time for a minute, with the peak hold spectra saved. The red trace indicates the result from the generic, with the blue trace indicating the result from the WiNRADiO. Of course, these results only hold true under my testing conditions – in the presence of strong RF signals, other effects from saturation/non-linearities may show up, or at different wire lengths, performance may differ.
Throughout most of the spectrum, the two adapters are very close in readings, with the signal strengths of the WiNRADiO seems to edge out the generic by about 3dB in absolute magnitude. Above 21Mhz, it seems the generic adapter seems to put through more signal by almost 5dB, but that could just be noise. In all, both adapters seem pretty much neck and neck, although I would have to give the edge to the WiNRADiO for performance and design reasons, although I probably wouldn’t be disappointed if the generic was all I had.
Interestingly, this might go to prove one of the counter-intuitive things with RF. Sometimes, a design might seem better as “on paper” it seems to solve the problem better in theory. But because of the “gap” when it comes to implementation, sometimes adding more complexity to try and improve the result actually ends up introducing more imperfections which work against you. It could be the case here that the two-toroid design with smaller toroids might have more losses than a loosely and messily wound single-toroid design causing the performance of the two designs to “converge”.
Bonus Teardown: Mini-Circuits ZFSC-2-1-75
A while back, I purchased a hybrid type power splitter with the intention of using a single antenna amongst several receivers. Sadly, the only unit available was a 75 ohm impedance unit, which isn’t ideal for 50 ohm systems (mismatch causes loss), but I decided to get it anyway. The antenna output signal wasn’t sufficient to overcome the -3dB splitting loss, insertion loss and impedance mismatch loss, so I counted this as a write-off, but it’s worth looking at how a reputable company constructs its splitters as a key to how RF products should be manufactured (contrasted with mass-market units, like this).
The unit itself is made completely with a metallic body, forming a low impedance ground path and shield across the active internals. The unit is held together with two screws, which are easily undone to reveal the insides.
The inside shows the three inputs neatly lined up on a small PCB with two transformers and some passives as well. The wiring appears straightforward, and neatly done.
The rear of the PCB is left un-etched and instead is used as a ground plane with plenty of solder to the shells of the connectors to ensure a low impedance ground path and consistent RF performance across the PCB.
This contrasts strongly with the low cost Chinese splitter linked earlier with no ground plane, only relying on the unit’s metal casing as a shield and ground, and hand-wound transformers with messy wiring.
In RF, the performance of devices often comes down to minute details in the design and construction of devices and their consistency. It is clear that properly designed equipment tends to be well shielded, rigidly mounted, and constructed in a way which is consistent (e.g. by use of formers, PCBs, etc).
Despite the obvious shabbiness of the generic adapter, its performance was not objectionable – which turned out to be a big surprise to me. This is likely because the unit wasn’t completely designed without thought, but also because HF frequencies are relatively low and the demands it places on dimensions are much less severe than in UHF/microwave work where hand-made often isn’t good enough.
Despite this, it is clear that the lack of shielding is a concern in RF harsh conditions, and it only really gets away with it because of the low flux leakage properties of toroidal transformers. The lack of weatherproofing and ground attachment point is an inconvenience as well, with the lack of rigid mounting making the core and enameled copper wire vulnerable to damage from mechanical shock. That being said, it’s nothing a few modifications can’t help or fix, but I suppose it wouldn’t hurt to spend a little more for convenience.