MFJ-207 Review

I purchased the MFJ-207 as it looked like a nice and simple instrument for quick tuning of HF antennas in the field. The basic "Oscillator + RLB + Meter" is a device I've been meaning to build for a long time. I have plenty of gear in the shack to lash-up the signal generator and a return loss bridge, but I have not yet built it into a single box that is easy to use out on the antenna farm. The manual contains the circuit diagram (unlike a lot of their other products), upon inspection it looked fairly well designed and quite similar to what I had planned to build anyway, so I placed the order. They once sold a similar unit with an adjustable reference resistor, the MFJ-207A (you can see it in the circuit diagram), I would have much preferred this unit but it seems it is no longer sold.

The MFJ-207 in action


The MFJ-207 costs about $100 USD and MFJ wanted $38 USD for shipping it out here. I ordered the unit on the 10th of November, the shipping calculation was deferred as I am international. I got the shipping price email 14 hours later... all good. I confirmed the order almost immediately. On the 12th I received an email telling me the unit was on back order, and I should allow 2-4 weeks for it to come in - not so good, but I was given the option to cancel - fine - I chose to wait for the device, I was in no rush. The device arrived on the 22nd. All in all the process was fine.


The unit tunes nominally from 1.4 - 31 MHz in 5 overlapping ranges called "A"-"E". My particular unit tuned 1.43 - 33.83 MHz. The calibration of the tuning dial is pretty abysmal on all but the "A" range and is especially bad on the "C" and "D" ranges. As the manual says, it is a general guide and you'll need the counter output or a calibrated receiver for practical use. The MFJ-2x9 and similar "LCD range" units with inbuilt frequency counters are obviously a better choice if you want it all in one box - you pay the premium of course.

The SWR meter calibration is horrendously poor. As a "tune for best match" indicator it is fine, but if you actually wanted to guesstimate the impedance of your antenna working back from the VSWR indicated - forget it! I tried loads that would give me 1.25, 1.5, 2 and 3 VSWRs - out of the box only the 1.5:1 was close - probably the factory calibration point. I tweaked the calibration and could get any one mark below 3:1 to line up, but it always read wrong on the others. Either the meter calibration wasn't matched to this particular design (a generic meter marking), or the diode rectifier isn't being driven hard enough to be reasonably linear. For the casual use of tuning up or looking for resonances it is fine - otherwise it might as well have an arbitrary better/worse calibration. The 1:1 null is correct and flat across the entire range which is a good sign. Similarly loads for specific SWRs present the same reading on the meter, right across the range of the unit. Loads either side of the reference for the same SWR (i.e. 25 and 100 Ohms) produce fairly dissimilar readings which is a concern. I left mine calibrated at 2:1 and accepted values reading high (worse than reality) elsewhere.

The oscillator buffering is inadequate and very poor for the counter output. The SWR calibration is noticeably affected by having a counter connected or not. As you can see from the circuit diagram, the bridge and counter pick off points are in parallel across the single J310 buffer and the ALC pick-off point is also there. It would have been much better to have an additional amplifier for the counter output, and probably a common-gate isolation amplifier in there too, plus a few pads (at the expense of power consumption and complexity/price I guess). Giving the ALC more headroom doesn't help much at all, and worsens the SWR calibration as the bridge drive drops. The most annoying problem however is the oscillator/ALC interaction with high-Q loads approaching resonance - in particular my HF vertical which I know tunes up quite well will pull the oscillator straight across its resonant point producing a brief flick down from "infinity" then right back up again. This problem is also compounded by moderately harmonic-rich output, especially on the lower bands (where the oscillator output waveform looks like a sawtooth and gets worse with lower impedance loads). FMing of the oscillator by the ALC loop produces FM sidebands wide enough that narrow band antennas like loops and short verticals may not see a match across the width of the oscillator output! Either way, there is frequently no way to persuade the unit to tune nicely into a very sharp match. I tried putting a 3 or 6 dB dB pad inline - which helped a bit but it still needs more isolation.

All that said, on your average dipole it works great. I had lots of fun looking at the resonances of my 30 metre inverted-V, and my 40 meter end-fed. The vertical on 30 is wide-enough to also work pretty well with the unit, but the best VSWR disagrees with what I measure with a return loss bridge, even after I compensate for the MFJ-207's poor calibration. I assume this is a harmonic radiation problem.

The FMing of the oscillator by the ALC loop is especially annoying if you wanted to use the unit as a signal generator. Although the source-coupled oscillator isn't ideal for this application; harmonic radiation is easily filtered out, the real problem is the FM modulation. The ALC is implemented using an LM324 which feeds a MOSFET in the tail of the J310 pair controlling the oscillator output level. Unfortunately this also pulls the oscillator frequency. Not a huge deal - except for stability with varying loads - as long as the ALC loop is well filtered. The LM324 output is swinging rail to rail at about 1.2 kHz, I expected it to be your average linear servo loop, but the LM324 is operating as a comparator and those 1M resistors and big caps in the MOSFET gate are the only thing averaging the DC bias. There is a small, but significant triangle wave component riding on the gate bias. You can hear the resulting FM sidebands on an SSB receiver while you tune across the oscillator output. Defeating the ALC loop removes the problem but then the RF drive to the bridge is no longer constant as you tune. The ALC loop oscillation may be a bug, or perhaps a misguided feature. I couldn't convince the loop to stabilise, and while looking I noted the PCB might need more via stitching, there is lots of RF floating around on the fragmented ground planes. The RF hygiene in places is a bit lacking, decoupling caps a long way from IC pins, etc.


The good points:

The bad (in order of importance):

  1. Insufficient oscillator buffering.
  2. ALC loop FM (oscillation and pulling).
  3. Poor SWR meter calibration.
  4. External DC socket has no polarity protection.

Other notes:

Am I sorry I bought it? No not really, it is easy to carry and useful as long as you know about its limitations - much like any instrument, commercial or homebrew. It is basically useless as-is for loop antennas or compact verticals on the lower bands - anything very narrow banded, and it can't double as a signal generator unless you can tolerate the FM noise. For dipoles it is a beaut little unit for pruning and tuning flat enough that you won't blow your radio up.

I've learnt a lot about what would make an ideal instrument by playing with this entry-level one. When building my own I'll spare little expense in buffering the oscillator, minimising the effect of the ALC loop and padding the counter output. By building it yourself you can also take more care in calibrating the meter and frequency dial which makes the unit just that little bit nicer. I'd probably add a switchable variable reference resistance too so you can measure antenna resistances once resonance is found or reactances tuned out.


2008-12-23: Fixed the MFJ-207 FM Problem
A two-component fix for the rail-to-rail oscillation of the ALC loop op-amp.