With my new QMX the antenna question arises.
My Elecraft KX2 atu adapts my 20m long EFHW with 1:49 transformer on all bands. A suitable z-match tuner can also be used.
But the QMX switches off when tuning with high SWR. The well-known SWR bridge with LED is a solution because the SWR never gets higher than 2 when tuning.
Here is the circuit diagram and the ugly prototype setup.
73 Chris
That’s what I am using, roughly. Works brilliantly with the QMX. Works also 60 to 15m with 1:9 Unun and 16m wire. No batteries, fast to tune, and lightweight, about 100g.
On my EFHW the SWR is ok even without a tuner on the 40m and 20m. Should I therefore use the QMX without the z-match tuner on these bands?
The measurement shows that the tuned z-match only has an insignificant attenuation in this case. So it can stay connected.
The updated circuit diagram with a small simplification. The antenna is coupled directly to the tuning coil.
73 Chris
I don’t know how familiar you are with how the Z-Match works and whether you are aware that for decades test reports have been published from time to time which generally “certify” that the Z-Match has an efficiency of only <50%.
Such general test results based on a too simple black box approach and an insufficient understanding of how it works not only fuel the ham myths, they are unfortunately also misleading for many people.
It could be therefore worthwhile for your project to take a closer look at the expected antenna impedances in the intended frequency range and to check your Z-Match for its matching load range and efficiency.
This topic was also discussed in these publications here
https://www.qsl.net/vk5br/ZMatch/ZEfficiency.htm
https://www.qsl.net/vk5br/ZMatch/ZDrop.htm
It would of course be elegant if in a QRP Z-Match not only the SWR bridge but also the Z-Match itself could be bypassed with a switch.
It’s perhaps worth mentioning that QRP tuners often have additional losses due to their construction. Thin wire, lossy cores and tiny variable capacitors with film dielectrics instead of air will all contribute. Of course these losses sometimes make it appear that it’s easier to get a good match.
Maybe I didn’t express myself clearly enough. I only measured the efficiency of my tuned z-match with input and output Z = (50+j0) Ohm. The attenuation was less than 0.3db, efficiency > 93%.
73 Chris
Ok Chris,
That’s interesting to know. So, the system impedance at the Z-Match output did not correspond to the complex impedance of the real antenna. The measured value should therefore be taken with some caution.
The effort required to become an expert in efficiency comparison measurements on antenna impedance matching devices (including transmission line and EFHW transformers) is not as great as one might think. All it takes is a bit of the right knowledge and a little work to set up a simple clamp-on RF current meter.
Both can be found, in an easy-to-understand manner, in the following publication by Owen Duffy (with chapter “Applying the meter” and corresponding “Online calculator”).
Measuring common mode current
The following publication by Dave, G4HUP (sk), also describes the construction of a simple clamp-on RF current meter and, particularly interesting, its application to coaxial cables in chapter “Test cable”. In the above-mentioned publication by Owen Duffy, this is no longer explicitly described because it is the same as for the 2 conductors of a 2 wire open wire feeder.
RF Current Meter by G4HUP
73 gl, Heinz
Chris, you posted two versions, where the output was taken differently. Which version did you use for your insertion loss measurements, and did the other version perform poorer?
Z-match is not my favorite antenna tuner partly because of the loss. Generally, a transformer wound like shown in v 0.20 suffers greater loss in higher bands, while the v 0.30 I expect improvements on 15, 12 and 10m.
Another thing about Z-match is that the inductor is facing the load of variable impedance, therefore the loaded Q of the inductor varies widely depending on the antenna impedance and the frequency. That is, the insertion loss varies widely depending on the condition.
One tuner matching network topology I’m familiar with and alleviates that problem is pi-C match. The two inductors are facing the 50ohm transmitter side, while two variable capacitors are facing the variable load antenna side. The insertion loss is pretty much flat regardless of the load impedance as long as the variable capacitors are of decent quality (usually Q > 1000). The inductors needed for the pi-C network is also pretty small, so air wound is often enough for all but 160 and 80m. (I built one that covers from 60 to 10m and I used T130-0 core only to make it easy (I get unloaded Q well above 200 with 18AWG enameled wire)… but my next version may be a smaller version for just a few bands… the inductor’s Q would be even higher) This type of network was used in Drake MN-4 series and Kenwood AT-100, 130, 200 and 230, but that topology didn’t become popular probably because it’s more difficult to design (not really! once you figure it out) and expensive to build than other networks (particularly T-match is easy and cheap to build).
Another good thing to measure for any antenna tuner is the range of complex impedance that it is capable of matching. You can terminate the transmitter side with a 50ohm and connect a VNA to the output port of the tuner. Keep measuring while turning the knobs. The complex conjugate of the read impedance is the antenna impedance it would match, so just read the Smith chart upside down (repeat the process for each band, of course). I know what it looks like for pi-C network but I’m curious how Z-match’s match space looks like.
I think the easiest way to measure the loss at high-ish Z load situation is to insert a well constructed 9:1 Ruthroff transformer at the tuner output to take it at 450 ohm (then the 50 ohm goes to your VNA port 2 and measure S21). I would expect the loss to be greater than when the tuner output is taken at 50 ohm, because the inductor operates at a higher loaded Q (sharper resonance) and greater loss.
Before anyone asks, if one wants the lowest loss tuner, the answer is L-match. But L-match requires a very wide range of inductor and capacitor values, so the only practical multi-band option is to drive the network elements with a bunch of relays just like how it’s done in modern ATUs. For portable operations, that is not always desirable (more battery to charge, weight, etc.), so I’m interested in efficient manual tuners… and I suppose people who commented above are in the similar situations.
Dear Ryuji,
thank you very much for your detailed and competent explanations. I’ll get in touch as soon as I have time. 73 Chris
Heinz HB9BCB: thank you for the links to VK5BR’s analysis and measurements of Z-match.
In his loss analysis, focusing on the region where Z > 200 ohm, the inductors that performed well on 20m did poorly on 40m, and vice-versa. That is not surprising because the operational reactance and therefore loaded Q vary widely in multi-band tuners without a band selector switch. But Z-match with a band selector loses the circuit topology’s elegance, I think, so making a balanced compromise is necessary.
As a benchmark, when we tolerate SWR of 1.5, we have 4% of the transmitted power lost as the reflected power. At SWR of 2.0, we lose 11%. So, I think it is a reasonable target to aim an insertion loss less than 0.5 dB at all useful load impedances. It appears to me based on VK5BR articles that a carefully designed and built Z-match could deliver that level of performance over one octave, but going higher or lower frequency, the loss will increase. (In other words, it is probably possible to build a Z-match with reasonable performance on 40/30/20m but trying to use it on other bands will suffer from significantly greater losses.)
In another article, he described that there is a hole in the match space, where the tuner is unable to match a capacitive load. He suggested adding a series inductor to move that void subspace so that the tuner can cover the entire space. That is also concerning.
Yes, it can be summed up well that way.
It is of course also possible to provide different switchable transformer windings for the secondary side, as was done e.g. by Steven Weber, KD1JV, in the internal Z-Match of the 40/30/20 m transceiver PFR-3 (kit retired) or by Peter Zenker, DL2FI (sk), in the ZM-4 usable on 160-10 m (QRPproject has since been closed forever).
Thanks for additional links.
I see the tuner in PFR-3 to be more of a generalized L-match with a transformer rather than a Z-match. But both that and ZM-4 have high-Z secondary windings. One recurring theme in RF design is that that kind of conventional transformer, where the primary and secondary windings are coupled via magnetic field through the core, perform well on 80, ok on 40, and increasingly lossy at higher frequencies (detail varies, of course). Chris’s version 0.30 somewhat alleviates that problem.
In the last 20 years there have been more discussions on relay-driven L-match ATUs but I think it’s a good time to review major matching networks, and reevaluate the practical options for portable operation. I might consider writing something up.
Much less precise, but a quick way to assess the relative loss of the tuner is to glue a temperature probe on the toroid and measure the temperature rise while continuously transmitting 5W into dummy loads of various impedances. This measurement is quite sensitive to wind (just the observer breathing nearby can make a difference) so I’d keep the enclosure closed. Of course the probe needs to be RFI hardened and with RF choke to minimize the impact to the tuner operation.
Incidentally, when I build my feedline transformers or transmitter final stage transformers, I usually put a piece of wax or hot melt glue on the core and make sure it is not melt after vigorous 100W POTA activation to ensure the transformer loss is acceptable level and within the expected range of the design.
