MFJ-993B on my G5RV with tuned feeder

This article is an analysis of why my recently acquired MFJ-993B will not match my multiband antenna system on most bands above 20m. The MFJ-993B replaces an Ameritron ATR-30 which was capable of matching the antenna system on all HF amateur bands.

A detailed analysis is performed 18.15MHz on the first problem band.


The antenna system uses a tune feeder configuration.


The alternative tuned feeder arrangement described at (Varney 1958).

In this case, the open wire line is 9m of home made 450Ω line (2mm copper wires spaced 50mm air insulated), a 1:1 current balun and 0.5m of RG400 tail to the ATU.

A 470K 1W metal film resistor and a 2095-100 gas discharge tube in parallel are connected in parallel with the ATU antenna terminals to reduced static build up and modest transient protection.

Impedance was measured looking with a Rigexpert AA-600 into the cable end that plugs onto the ATU, at 18.15MHz is is 4.7-j69.5Ω.

MFJ-993B specifications

Screenshot - 06_09_16 , 09_22_42So, what does this mean?

It refers to impedance as a scalar quantity which strictly means that the unstated imaginary part is zero… which makes inequalities like “<50Ω” a nonsense as it is difficult to use scalar inequalities on complex quantities, so perhaps they mean |Z| (which is a scalar quantity) is less than or greater than 50Ω.

The measured load impedance is 4.7-j69.5Ω, |Z| is 69.66Ω, VSWR50 is 31.25.

On that reasonable interpretation, this load is within specified matching range.

Theoretical lossless L match solution

The MFJ-993B allows only two of the eight possible L match configurations, a series inductor and shunt C which may be on either side of the inductor.

To get a feel for the matching problem, lets explore a simple lossless L network to match the measured impedance to 50(+j0)Ω.

Screenshot - 06_09_16 , 08_57_45

Above is a solution, and it requires a series L of 737nH.

Now lets find the maximum and minimum L that give a chance of getting within the VSWR=1.5 circle (with appropriate C of course).

Screenshot - 06_09_16 , 08_59_44

Above, a minimum of 710nH (though impractical because it requires a specific value of C to just touch the VSWR=1.5 circle).

Screenshot - 06_09_16 , 09_00_20

And a maximum value of 770nH (again requiring an exact value of C to just touch the VSWR=1.5 circle).

So, L must be adjustable within the range 710-770nH, preferably say 725-755nH, a maximum increment of 30nH.

The MFJ-993B would appear from its manual to have maximum L of 24µH adjustable in 256 steps of about 94nH… three times what is needed for matching this load.

The same issue can be analysed for the capacitance element, but in this case, inability to adjust L in sufficiently fine steps is a show stopper and the principal reason that the ATU will not match better than VSWR=4 on this load.

Manually stepping L and C about the best match point confirms that the steps in L are too coarse to permit adjustment to lower VSWR.


The MFJ-993B is not a wide range tuner. Whilst the ATR-30 and MFJ-949E could achieve a match on 8/8 HF bands with this antenna system, the MFJ-993B achieves a good match on 4/8 bands and a poor match on 1/8 bands.

Exploring the 18.15MHz case reveals that the ATU is not defective, the issue relates to the design of the ATU.


  • Varney, Louis. July 1958. An effective multi-band aerial of simple construction In RSGB Bulletin July 1958.