A question of often asked is whether the ‘best’ balun on a T match ATU is:
- a Ruthroff 4:1 voltage balun (as it often fitted internally);
- a Guanella 1:1 current balun; or
- a Guanella 4:1 current balun.
The subtext of the question is often that the OM has switched from (a) to (b) in pursuit of improved common mode current rejection, and found they cannot match their existing antenna which matches ‘just fine’ using the (a).
This article explores (a) and (b) using a configuration similar to the popular MFJ-949E @ 1.8MHz, and comments on (c).
Let’s declare that although the MFJ-949E is specified for 1.8MHz, it is disadvantaged… which will exaggerate the problem to some extent… but in so doing make the problem easier to see.
We will use a SimNEC model of the ATU and a purely resistive load, as large as the ATU will match. The Smith charts shown are normalised to Zref=400+j0Ω as it is a better context for visualising loads on 400Ω transmission line typical of the use with two wire lines.
1. The bare MFJ-949E T match to a coax output jack
Above is the model. The coloured squares a sample points in a sweep of values of the matching components C1, L1, and C2. Green is loss less than 3dB, Yellow 3-6dB, and red greater than 3dB.
The ATU will not match Rl>200 when Xl=0. This might seem rather limited, it is due to the choice of maximum values of C1, C2 and L.
No dots in a region signal that that area is out of range of matching, sparse dots might usually suggest that the controls will become touchy in that region.
So, note the large circular white area to the right of the chart, and the crescent shaped white region at the lower part of the chart.
Save the small area to the far right where there are some yellow and red dots, where match is possible, it is likely to have loss less than a few dB.
So, the basic coax in to coax out ATU is quite limited in its matching range @ 1.8MHz.
2. The MFJ-949E T match using the balanced line terminals
This connection uses an internal Ruthroff 4:1 balun.
Above is the ‘match map’.
Compared to (1), the ATU can now match over most of the chart, save a circular white are around the ‘splat’ mark and again a crescent shaped white area near the lower circumference… but, only a small part of the ‘match map’ is green (<3dB loss), and much of the chart is red (>6dB loss).
You might observe that the match map is not simply 4 times (1), and ask how is it able to match beyond 4 x 200Ω, the limit in case (1)?
This is due to the non-ideal nature of the internal balun, its magnetising impedance and leakage inductance is included in the model and they alter the impedance transformation… rotating the circular white patch clockwise a little and 5000+j0Ω now comes within matching range. It is no longer simply a T match circuit with three components, the magnetising impedance and leakage reactance of the internal balun become significant.
So this configuration shifts and rotates the usable matching range. If one had an antenna that could not be matched, common advice is to add or subtract some feed line, so rotating the load out of the no-match zone.
One might regard that since most of the map that is red, it is effectively a radiating dummy load… hey, but the VSWR is perfect… who is going to know?
Note that a voltage balun is not a good solution for reducing common mode current for antennas with significant asymmetry.
3. The MFJ-949E T match using the coax output and a directly attached Guanella 1:1 current balun
This connection uses an external Guanella 1:1 balun with zero length coax interconnect.
Above is the ‘match map’.
Unsurprisingly, it is quite similar to (1.)
The fundamental T match circuit of the ATU will not match Rl>200 when Xl=0.
So, note the large circular white area to the right of the chart, and the crescent shaped white region at the lower part of the chart.
Save the small area to the far right where there are some yellow and red dots, where match is possible, it is likely to have loss less than a few dB.
So, the coax in to coax out to directly attached Guanella 1:1 balun is quite limited in its matching range @ 1.8MHz.
Still, you might argue that coax connected antennas would tend to be lower impedance, to the left of the chart. Well, not necessarily, eg end fed wires (in the ‘wire’ output configuration), or external baluns for example.
4. 4:1 current balun
A good 4:1 current balun may provide a useful change to the match range (at the expense of matching and efficiency of low Z loads).
But, (there is always a but), MOST 4:1 current baluns that one sees on Youtube, in ham pages, and in commercial listings online, on eBay etc are deeply flawed.
Note that the common mode impedance of a Guanella 4:1 balun is approximately half that of the component baluns, so they are a little less effective in common mode current reduction than a 1:1 using a component balun.
So this could be a solution, but it will be quite different to the internal voltage balun (see the mention of the effect of magnetising impedance and leakage reactance).
Badlands
The article Avoiding flashover in baluns and ATUs explains one of the badlands on the Smith chart. The article suggests you avoid operating where G<0.0005S to avoid extreme voltages.
On some of the charts above a crescent shaped region to the right of the chart was low efficiency, they are in the badlands where G<0.0005S (the crescent outside the G circle scaled 500u on the Smith charts given above).
This speaks to the problem that an ATU and balun do not universally solve problems of a fundamentally bad antenna, though is lots of cases they will ‘paper over’ the problems, hiding risk of component failure and lost power.
So, one should ask “is there something sub optimal in the rest of the antenna system that gives rise to observed problems with the ATU?”
A possible solution with this ATU and ones like it
Internal components impose limits on the range of matching. The internal voltage balun is not necessarily a bad choice, but it is less effective in reducing common mode current.
Where measured common mode current is high (or is suspected high… hams don’t tend to measure and fix), an external good Guanella current balun connected between the balanced output terminals and the feedline may be effective in reducing common mode current to an acceptable level. This effectively implements a Roos Balun.