Just to make sure: With this you mean that the primary winding should be a tap (true autotransformer-style) instead of using twisted magnet wire for the first two or three turns, don‘t you (as shown on your images)?
73 de Martin, DK3IT
PS: One more question for @HB9BCB and the others here: For a vertical (20-17-15 with traps), feedpoint ca 1-1.5 m above ground (11m mast), would you use 1:49 or 1:64 as an impedance ratio (3:21 or 3:24 turns ratio)?
One more question (apologies if this has already been answered in this now substantial thread):
As for measuring efficiency, you basically use two identical transformers connected back to back, right?
And then, do you use
a) an spectrum analyzer plus tracking generator to measure the attenuation, or
b) an RF power signal (5W or more) plus a power measurement at the other end?
My understanding is that only b) would show losses from e.g. secondary effects from heating.
Also, if I remember correctly, there are subtle limitations of the back-to-back measurement setup for EFHW transformers, but I forgot the details.
Martin,
Here my answers to your last 3 posts.
Re splitting the winding:
Yes, as an autotransformer with a tap after 3 turns.
Below an example of a (not recommended) splitted winding, meaning the first 3 common turns. Also not recommended is the Reisert winding method, which is intended, for practical reasons, for winding toroidal cores with coaxial cables.
Photo below copied from the beginning of this thread.
Re measuring transformer efficiency:
Yes, these points have also been discussed on this reflector, but don’t worry about your question.
Useful and also recommended by Owen Duffy: A transmission measurement (s21) with a single impedance transformer (terminated with the nominal resistive load, or for perfectionists: the nominal load minus 50 ohms to account for the meter input impedance), usually performed with a VNA (photos of my measurement setup below).
Back to back measurements with 2 impedance transformers are often made for practical reasons (e.g. because of the capabilities of the measuring device or the math skills of the tester) and are not good enough for certain purists, hi.
BTW, The transformer efficiency of my very first 1:64 EFHW impedance couplers with a FT-82-43 toroid and 3 primary turns was something like this:
65% at 3.6MHz
70% at 7.1MHz
72% at 10.1MHz
74% on 14.2 and 18.1 MHz
75% on 21.2, 24.9 and 28.5 MHz
The measurement of the transmission properties when the ferrite core heats up due to use and/or extreme ambient temperatures is perhaps a bit too academic for our needs. There are scientific studies to investigate the temperature-dependent behavior of the complex permeability of ferrite toroidal cores (example for NiZn ferrites below), busy readers could read chapters 1 and 4 at least …
So just complicate things as necessary - and use the time to climb mountains and make QSOs, hi.
Someone asked why the VNA graphs do not also show the loss and efficiency of the measured impedance transformer.
The requested values can be calculated in a simple way.
It should be noted that the load resistance used (3203 ohms) and the VNA input impedance (50 ohms) are also measured during the transmission measurement (s21) of the 1:64 impedance transformer.
In addition, because the impedance of the transformer input is not exactly 50 ohms, there is a (small) matching loss that must also be taken into account.
Calculation example for 7.0 MHz (Marker #2)
Impedance at transformer input 47.56-j0.94 ohms
→ Mismatch loss 0.003 dB (calculator by Owen Duffy see below)
Transformer loss=
-|s21|-load loss(load resistance+VNA input resistance)-mismatchloss=
-18.59-(-10*log50/(50+3203))-0.003=-18.59-18.13-0.003=0.457 dB
Transformer efficiency=10^(-0.457/10)=90.0 %
Calculator by Owen Duffy to
Calculate VSWR and Return Loss from Zload (orYload or S11) and Zo
https://owenduffy.net/calc/ZY2Vswr.htm
For those interested: With the ferrite core and compensation capacitor of the 1:64 impedance transformer (above), when wound as a 1:49 impedance transformer, the transmission values shown below result.
At 7.0MHz, the transformer loss is 0.437dB and the transformer efficiency is 90.43%.
The difference is not exactly dramatic - on a real antenna, the different transformation ratios will be particularly noticeable in the lowest and highest frequency bands - also not dramatic, but possibly relevant when working without an ATU.
I am not capable of doing this so if somebody is.
I am interesting of sumerize all this articles in one short with drawings what to use, both feritte and capacitor, how to wind, what lenght of wire
As with most things, that depends on your requirements (bands, efficiency, power handling, …).
I started my experiments with the well written document of Heinz @HB9BCB (in German) and continued my experiments that resulted in the following document:
At that time, I was not aware that Heinz experimented as well with the same toroid core. My main goal was to design a 20m long portable multiband EFHW that is resonant on more bands than the usual 40-,20-,15- and 10m bands and requires a minimum intervention to change the band segments. The solution was to insert one coil with one bypass link or switch.
The majority of transformer designs that one finds on the web and even many commercial products are not very efficient. Just read Heinz’ articles in this reflector and you’ll find out.
They work, but Owen Duffy would say: “Well, in ham radio, everything works. But systems that work better increase the prospects of contacts.”
73 Stephan
Well, that is what I would like to avoid. Reading another 20, 30 or 50 pages. I would read them one day, but not today. I want direct recipe what and how to do.
So, I will start with my setup. You may tell what is good and what is bad.
I am using also another transformer 2:14 turns
I wound it on 2x 140-43 core. Choke is wound on one 140-43 core.
With both boxes I use 2x5 meters long counterpoise. It goes into green connector. I am not using capacitor parallel to primary because I use it on 40, 30 and 20m only. I think that capacitor is not needed for this three bands.
I do not know wire lenght. I broke it two days ago. After repair I will measure wire lenght and give swr curve for grey box
Hello Martin,
I’m using Wuerth toroids (material 4W620) for a while. Some weeks ago I tested a transformer (UnUn, 2 + 12 windings) with such a toroid. The loss @ 40 m … 20 m was around 0.5 dB ( - 12 %) or below.
My measuring setup was similar to what HB9BCB described.
Now I’ve build a transformer with 3 + 18 windings (autotransformer, z transformation 1:49) using the mentioned wuerth 742 701 0.
The loss is not higher than 0.4 dB (+/- 0.2 dB) @ 80…15 m. The VSWR on the 50 Ohm side is 1.15…1.56 @ 80…20 m and 2.4 @ 15 m. (+/- 0.2 dB due to accuracy of the measuring equipment)
With 120 pF parallel to 50 Ohm the loss is not higher and the VSWR 1.1…1.8 @ 80…15 m. Loss and VSWR @ 10 m are a little bit better than for 15 m.
So in my opinion you may use such a toroid. This toroid (and some more) is available from reichelt.de.
73, Ludwig
As far as I see there are 34 or 35 visible turns on the toroid on your first picture. You mentioned 27 and 3 turns. What is wrong?
73, Ludwig
I do not remember exact number of turns. From my weak memory are 27. Maybe you are right.
It is not very critical. Just wound turns and try. All you need is resonance on the band. Than change the number of turns for best swr.
I connect 3,9 kohm (it may be 4,7 or 3,3) and tune with capacitor for resonance. NanoVna is grat tool and you can watch dip moving as you turn capacitor.
Do not make my mistake and glue toroid before tuning.
Than cut litlle more than half lambda of wire, go out, erect antenna and little by little cut wire until best swr. Again, nanovna is great tool
I am tired of long articles and lot of links. I just want short know how.
Also, if one thinks that something is wrong I do not like: you are wrong and here is why … (link to 50 pages of tehnical explanation).
I would like explanation in two or three sentences and in short what to do. Pictures or drawings also.
It takes more time to drill holes in the box that to wind toroid. Experimenting with different number of turns is very easy.
EFHW antena is simple. 20 euros in toroids, magnet wire and few meters of insulated wire gives you years of experimenting and using antenna in the field
Here are few youtube videos. See the 5 band tuner. One inductor and capacitor from transistor rx
I would like something like this about transformers
Ok, very quick:
Most EFHW transformer designs on the WWW are flawed, despite their popularity.
Depending on the band and design, some pass only 25% of your output power to the antenna. That is a loss of 6dB and this can mean that you can only work strong stations.
The new best practice is as described and illustrated by Heinz above:
a) Use either the recommended toroid, the alternative from Wuerth, or TWO FT-50-43 stacked and glued together with superglue.
b) one winding of 24 turns and a tap at the 3rd turn; the tap is the input.
c) styroflex or other high-quality capacitor 68-100pF in parallel to the input.
d) tight windings (see Heinz‘ perfect example), not too thick wire (0.4 or 0.5 mm); thicker wire will not lay as closely on the toroid and hence cause losses.
Simple enough? 
73 de Martin, DK3IT
PS: Two stacked FT-50-43s perform surprisingly well in case you have difficulties sourcing the other toroids.
Yes, short and simple. Thank you.
By the way, I measured the lenght of my antenna wire for 20m. It is 9,9 m. It is 20cm shorter than formula in hb9sota document for monoband efhw (142,5/f). I think it is because I use it in slope configuration and not vertical.
I also measured resonant frequency in lambda/4 configuration, hoping to get resonance in 40m band. No. Resonance is 7700 kHz
Hi Damir,
You could try adding little coil 5-8 turns on 2-3cm diameter former and 3-4 radials of 5mtr. This is what I use sometimes. Vertical wire is from trapped EFHW for 20/15/10 bands used as vertical, when I want 40 I attach coil and radials. You might need CMC on coax in such situation.
73 Marek
Thank you Marek for your answer.
I know about coil. A friend of mine is using efhw at home and that little coil helps him to get resonance in 7, 14, 21 and 28 without tuner. Without coil resonance goes out of higher bands.
I think, and maybe I am wrong, that aditional inductance with 24 turns instead of 14 turns is doing same job as that coil. Helping to lower resonance on higher harmonics.
But again, maybe, very likely, I am wrong
PS: I reread your answer Marek and I 100% miss the point. You are talking about lowering the frequency of short wire. Yes, you are right. That will help
But, little coil about 2m from 2:14 transformer helps
Hi Damir,
Sorry I wasn’t clear enough. My idea is that instead 49:1 transformer I put little coil and radials. Basically it becomes 1/4GP for 40m band.
The coil you meant is 5-6turns coil on antenna wire about 2mtr from transformer. It is still EFHW.
73 GL Marek
Hello folks,
Following recommendation about toroid winding I did a few experiments (not yet finished).
But a the beginning I want to say thank you for pointing to this ideas.
Here are what I got:
Pic. Separated windings for primary and secondary but with crossing (W1JR). Too much inductance due to fact that this crossing reduced winding capacitance. I want to try 0.4mm wire instead of 0.6mm as on picture. But can be used if you observe to many capacitance (upper graph is reactance with step 10 Ohms ). Measured capacitance for lowest SWR is 88 pF but I soldered 100 pF (don’t have 88 =) )
Pic. Common (not recommended winding) with gathered primary and secondary 3 turns and capacitor 150pF. Unfortunately forget to add reactance graph. Used wire is 0.4mm.
Pic. Separated windings for primary and secondary but WITHOUT crossing. Measured capacitance for the lowest SWR is 70 pF but I soldered 100 pF. Used wire is 0.6mm
It is important to mention that wire diameter impacts on winding capacitance and that is why care about this and don’t solder 150 pF without measurements.
I used variable capacitor 20-250 pF for adjasting for lowest SWR.
Hallo Yuri
Which toroidal material did you use?
73 Armin
Thank you for the insightful comparison.
I would now be interested in how the SWR behaves on a EFHW of 20m in length.
Because I suspect that the antenna wire at the feed point causes additional capacity and the cross-over version might be advantageous on 15m / 10m.
73 Chris
