2008-12-20
MFJ-269 Review
After my review of my MFJ-207 Peter VK2TPM offered me his MFJ-269 to play with for comparison. Naturally I jumped at the chance. The MFJ-269 is the fully-featured descendent of the earlier and simpler antenna bridges like the MFJ-207. The manual can be found on MFJ's website.
The device is microprocessor based with an LCD interface and two main control buttons "Gate" and "Mode". There is also a Frequency band switch, a reduction geared tuning capacitor, power and UHF mode switches. Two analogue meters display SWR and impedance magnitude continuously, offering a trending display which is more user-friendly than the figures on the LCD display. In UHF mode there is a "bargraph" of sorts on the 2nd line of the LCD for trending SWR.
The unit tunes about 1.76 MHz to 172.7 MHz in six overlapping bands. There is also a "UHF" switch that triples the oscillator in the 114-170 band giving 415-470 MHz coverage the edges of which are enforced in software, the display telling you to increase or decrease frequency until within that band. UHF supports only loss and SWR features, not reactive measurements - likely because bridge performance issues at UHF would have required extensive calibrations to have any hope of accuracy.
The coaxial connector is a N-type and is surrounded by a generous rubber gasket. The frequency counter input is a BNC. The DC plug is a conventional unit, tip positive, I am unsure what internal over-voltage and polarity protection it has. There is a large grounding lug also provided.
Features
There are 4 "Main" analysis features and 3 additional menus of Advanced ones. The advanced ones are mostly just different presentations of the same data from the bridge voltages (impedance in polar form, reflection coefficient, return loss, etc), but there is the option to select a different "Zo" for the calculations and also a two-frequency point semi-automatic calculating feature for line length, distance to fault, etc. The line length and fault distance calculations are pretty good, resolving several short (~2 metre) lines and offering compensation for velocity factor.
By pressing the "Mode" button you can cycle through the different modes, both main and advanced (once entered). Holding both the "Mode" and "Gate" buttons gives access to the advanced menus. The "Gate" button is used in frequency counting mode and as an enter-key of sorts in the calculating modes. Holding both buttons down on startup then alternatively pressing them as it boots enters the "Test" mode which gives a display of the raw ADC data for calibration of the unit with a set of calibrated dummy loads. The button mashing required to enter "Test" takes some practice, thankfully it isn't something you'd be doing often. The software version is displayed at boot, in this case 4.46, copyright 2004.
Impedance Analysis
The default "main" mode is the "Impedance R + X" mode and displays frequency, VSWR, and impedance as Resistive and Reactive magnitude (Rs +/- jXs) which update continuously as you tune around. The sign of the reactance is not resolved and displayed, which is pretty typical of the bridge system the device uses. The frequency counting is probably the "killer feature" of the device as well as the reactive magnitude display which allows searching for true resonances not just minimum reflection coefficient magnitude.
The impedance mode does seem to work pretty well, especially below 50 MHz, but is limited by the device's calibration accuracy. I am not very confident the unit I was playing with is calibrated properly, resistive magnitudes reads high, sometimes more than double - and above 300 Ohms resistive the reactive magnitude rises very quickly too, even with a "pure" resistance in the lowest frequency band. This and other behaviour I noticed suggests the "Vz" channel gain is a bit high. Upon checking the calibration pots under the battery holder I noticed that R72 was bottomed out, offering insufficient range to properly calibrate the unit as described in the calibration procedure. This is suggestive of mismatched or damaged diodes in the bridge or perhaps an incorrect resistor value or tolerance catastrophe. As the unit is not mine I didn't pull it apart to inspect the rest of the circuit and attempt to diagnose it further.
Coax Loss
The second "main" mode is "Coax Loss". It appears to simply display half the return loss measured by the bridge. When an unterminated coax line is attached the bridge sees the through-and-back (returned) loss of the line, which is twice the line loss for that particular length. Similarly you can attach attenuators and other lossy devices to the unit and measure their loss this way. The loss maximum appears to be 24 dB and is likely limited by the best bridge directivity over its frequency range (ie 48 dB, pretty good). As the unit is measuring with a 12 bit ADC a linear signal (not a log detected one) the dynamic range is pretty limited and the loss quantisation is fairly coarse. For "good" vrs "completely waterlogged" measurements of coax it is likely sufficient.
One annoying thing is the lack of auto-zeroing, this particular unit would always read about 0.9 - 1.6 dB return loss at HF. It would be nice to have software zeroing. In fact I can't see why the entire device could not have been designed to ship with a set of calibration loads and full software fine calibration. I do understand that some of the gain settings are quite critical, especially for reactance zeroing with the limited dynamic range offered by the 12 bit DACs fed by linear detectors. Still it would be nice to have "tare"-style zeroing for some measurements.
Capacitance and Inductance Measurement
The Capacitance and Inductance measurements simply compute the reactor value that matches the current reactive magnitude reading at the current frequency. It has no way of telling if the load is actually inductive or capacitive, and once outside the few ohms to 1.5 kohm range of reactance magnitude the software stops giving you a solution. This makes perfect sense, you just need to tune the VFO until the frequency offers a reactive magnitude in-range of the unit. The computed values are most accurate when the reactance magnitude is close to the bridge reference resistance (50 Ohms). This particular unit read high, about 19 pF high (about 6 pF of which was the test fixture) on average and again there is no way to null it. Inductance was probably worse but I didn't test it extensively, only a few trial inductances were compared to values measured by my LC tester. The LC measurement is nice to have, but I wouldn't trust it - in fact I'd go as far as to suggest using the frequency counter feature along with an external jig like the LC tester instead.
Correctly calibrated the general measurement method is probably fine and quite useful as it tells you the true impedance of the reactive component at radio frequencies (unfortunately only those that present a 7-1500 Ohm impedance). For testing a component for spurious resonances this is very useful, but in practice I couldn't seem to easily find misbehaviour in capacitors. Inductors on the other hand were fairly easily tested for self resonance, but being placed in the tester environment pulled their self-resonances quite a lot. Dipping them instead is probably more reliable.
Frequency Counting
The frequency counting feature works as advertised. It can accept signals up to 5 volts and has an high-impedance input. I could easily get a stable measurement with just a few turn coil plugged into the input using my dip meter as a signal source from several inches. The counter counts to beyond 170 MHz, I didn't test it higher. I assume the "UHF" feature actually counts the VHF generator and triples the value in software? I can't see a counting range spec in the manual. The gating works as expected, long gating times give more resolution figures. The reference seems stable and accurate enough for the resolution offered.
Advanced Features
Of the "Advanced" modes the stub length and resonance searching ones are probably the most useful. The stub length helpers are great for cutting matching and phasing stubs. Accuracy for 1/4 lines seems pretty good when compared to dipping a line with other instruments. It is unfortunate that the physical lengths are only given in feet. A metric physical measures software option would be nice, but considering this is a US product this is perhaps not surprising. The resonance searching mode makes the Impedance magnitude meter show Reactive magnitude while looking for a null. Oddly it does not use the LCD bargraph like SWR in UHF mode, instead showing Xs numerically.
The match efficiency mode is arguably the most useless feature. I'm not really sure what the point of it is. I guess there was extra room left in the MCU code space, so yet another "feature" was invented to fill it. Personally I'd prefer metric measures or software nulling instead.
Use as a Dipper
I understand there is an optional dipper coil set for the MFJ-269? - I just plugged my few turns of wire on a BNC into it and tried it out on some inductor self-resonances. It works quite well on HF and VHF. I'm not too confident about it on UHF, but the UHF range is quite narrow which makes it fairly useless for dipping anyway.
Power Consumption
The device does guzzle power. On HF-VHF in the default mode it pulls 140-170 mA (rising with frequency). On UHF it pulls over 350 mA! In counter-only mode it still pulls 90 mA. The AA batteries won't last long, and it takes 10 of them. Unlike the MFJ-207 at least you only need to remove two short machine-threaded screws to access the batteries. The external power option is almost mandatory (use at least a 500 mA plug-pack), but it does offer the ability to charge rechargeable batteries inside the unit by changing a jumper on the PCB. I have no idea how well it manages them if you choose this option.
By default the unit enters a sleep mode to reduce power consumption after a period of inactivity. This would help extend battery life, but it still pulls more than 50 mA. The sleeping feature can also be disabled by holding down buttons while the unit is powered on and remains disabled until the power is cycled again, much like Test mode.
Comparisons to the MFJ-207
The 207 is very basic compared to the 269. The 269's frequency counting and reactive magnitude display are its best features. The 269 has none of the FMed oscillator problems of my particular 207 and has much improved buffering and harmonic distortion. There is a pot in the 269 for setting buffer bias up to minimise the harmonic energy. I didn't test it extensively, but there is a detailed procedure available using a coax stub instead of a spectrum analyser to ensure this is set correctly. Tests with narrow band antennas like my balcony HF vertical and bicycle loop antenna show it is very much improved over the 207. I can resolve the resonance of my balcony vertical on 80 metres with the 269 fairly easily where it is next to impossible with the 207. I even tried measuring a colourburst crystal with the 269. The xtal resonance is very steep and the analyser tries its best, but it is simply not sufficiently stable or well buffered enough to stay in the resonance. You can detect it and even get a fairly good idea of its frequency however, and see spurious resonances of the xtal as well.
The 269 covers part of VHF where the 207 stops just above 30 MHz - and the 269 also has the narrow UHF option. On UHF the 269 is essentially as limited as the 207 is on HF, measuring just return loss. I am highly suspicious of the UHF feature's accuracy and debate it actual usefulness for most HAMs.
Like the 207 the 269's SWR meter is largely just for trending. The calibration point is 2:1 (using a 100 Ohm load), above and below this the displayed figure is quite wrong. The LCD display however shows the correct figure, at least below about 5 - and of course if the unit is calibrated properly. (The impedance meter is a bit better, its calibration point is 50 Ohms using a flat load. Again the screen gives a more accurate reading.)
Summary
The good:
- Frequency display
- Reactance magnitude display
- Separate impedance and SWR analogue meters
- Mechanical reduction driven tuning capacitor
- Display options including RL and Ρ
- Fairy compact unit, self contained
- Reasonable accuracy (if properly calibrated)
- Fairly stable and pure local oscillator
- Internal battery charging option
The bad:
- Large power consumption
- No nulling options (especially for L/C measurement)
- Imperial physical length measurements
- Analogue SWR meter scale calibration is poor
- Tendency to arrive broken or poorly calibrated
On the last point the evidence is only anecdotal... Well, the unit I played with has a poor calibration and Peter experienced a lot of problems with it himself. His is not an isolated experience if you search HAM forums.
Conclusion
The MFJ-269 seems to be designed quite well. I am sure with careful calibration it is capable of quite reasonable accuracy over HF and into VHF. I strongly suspect it would be quite difficult to design something more accurate without ending up building a true vector network analyser with log detectors and synthesised local oscillators. With recent advances in technology it might be possible to build such a device for about the same asking price, but I am sure the development costs would place its initial RRP closer to $1k, making the MFJ-296 look like quite a bargin.
Is the MFJ-269 overpriced? Well perhaps. Especially in the light of the horror stories out there about some HAMs experiences with the unit. How many of them are due to "operator error" is unknown, but MFJ is very well known for having quality control issues, so it is likely that many are real hardware problems. Fortunately there is also a great deal of support available for the unit, with MFJ sending part kits out to repair blow-up or factory-buggered units. By virtue of its ubiquity (and from the earlier 259 and 259B units) the HAM community has plenty of resources for fixing busted MFJ-269s.
Would I buy one? If it was < $300 AUD - yeah I would take the risk of getting a lemon. I would feel more comfortable if there was a circuit diagram in the manual, that way at least I could fix it easily enough. There are diagrams online that profess to be of the 269 or 259/259B, they seem pretty correct but I haven't extensively studied this loaner unit for comparison. There is a calibration procedure available, and one for the 259/259B as well from the original designer, but note that the 259 has 8-bit ADCs so the hex values displayed on the 269 and the pot numbers don't match this page - still it is a useful read and the document on the MFJ site seems to match the general guidelines when you convert to 12 bit numbers.
Unlike a VNA-solution the MFJ-269 isn't tied to a PC which is important for field work on antennas, but this is becoming less of a problem with USB-based VNAs that can run from a Netbook form-factor PC, perhaps it would be best to save the money and invest in a VNA as many of us already have a netbook PC, especially for use in the lab rather than just antenna work. Still, it would be nice to have it all in the one portable box like the MFJ-269.
5 comments.