MFJ-1026 Noise & Interference Canceler
Electrical and electronic noise are the bane of a radio amateur's life. And for those of us who live in densely populated urban areas it is increasingly hard to avoid. It can come from a variety of sources - overhead power cables, electric fences, electrical equipment, switched-mode power supplies, TV and computer equipment- and if you are not the owner of the interfering equipment it can be difficult to do anything about it. If you are operating a stealth amateur radio station and the noise is coming from a neighbour's property then it is impossible even to say anything about the problem to the neighbour without breaking your cover and revealing what you are up to.
A couple of years ago I first noticed an unpleasant noise rather like an arcing thermostat that went on constantly all day long, and reached S7 on some bands. Although it spoiled the pleasure of listening, the noise didn't actually stop me from operating because I mostly use narrowband modes and a narrow filter eliminated most of it. However, a few weeks ago the noise level jumped up to reach S9+10dB, making 20m and 80m unusable and 40m very unpleasant. I was faced with the very real possibility of giving up the hobby if I couldn't do get rid of it.
In an online discussion of a problem similar to mine, someone mentioned the MFJ-1026 Noise Canceler. I checked out the reviews on eHam.net and the vast majority were favourable. A friend of mine, Dave G4AON, had got one and not had much success. In fact, when he heard about my problem he offered to send me his unit to try. But I was so desperate to see a solution to the problem that I had already ordered one from Waters and Stanton. It was a gamble that had my future in the hobby at stake.
Theory of operation
Although the idea of a device that can cancel out noise seems too good to be true, the principle of operation is actually quite simple. However, it's important to understand how it works, and that it is not 'magic', in order to appreciate what is can do and what it can't do.
The MFJ-1026 is not a noise blanker. A noise blanker is a circuit that looks for sharp pulses, such as may be caused by car ignition interference, and mutes the receiver for their duration so that they are literally blanked out.
Nor is the MFJ-1026 a noise reducer, a feature offered in many DSP-based radios and even available as an external accessory for other receivers. Noise reducers are supposed to work by filtering out parts of the received audio that are more prevalent in noise than in the wanted signal, but in many people's experience they usually just make the signal sound horrible without necessarily making it any more readable. That's certainly my experience of the K3 DSP NR, anyway.
The MFJ-1026 is a noise canceler. In order to work, it requires two antennas. One, called the noise sensing antenna, should ideally receive the noise well but wanted signals badly. The other is your existing antenna. By adjusting the gain so that the noise is equally strong on both antennas, and then adjusting the phase of the signal so that the noise on one antenna is 180 degrees out of phase with the noise on the other, the noise is cancelled out. It's simple, at least in theory!
Practical implications
Although noise cancelling is a simple idea - and if you look at the schematic of the MFJ-1026 or take a look inside you'll see that it doesn't contain many components - it is harder to achieve in practise. In most cases, a noise sensing antenna that is effective at receiving the noise will also be quite good at receiving radio signals. Therefore, surely, signals will be cancelled out as well and you'll be left with nothing?
The speed at which radio waves propagate is finite. Signals will not arrive at one antenna at exactly the same time as they arrive at the other. If the phase of the two signals is adjusted so that the noise is canceled out, only signals arriving from the same direction as the noise will cancel out. Signals coming from the opposite direction may be enhanced. This allows MFJ to claim that the MFJ-1026 can be used to create an electronically steerable antenna. Whilst probably true, without any directional calibration it is hard to imagine anyone using this product for that purpose.
The fact that only signals from one specific direction at a time can be nulled out has implications for the potential user. The MFJ-1026 can only eliminate noise from a single source. So if you are suffering interference from multiple locations, this device may not give enough of an improvement to be helpful.
Even if you only have one type of noise to eliminate, but it arrives from a wide angle of directions, then you will not be able to completely eliminate it. An obvious example of this is power line noise, where the power line is radiating noise along its whole length. Understanding this will help you appreciate what is possible using the MFJ-1026 noise canceling unit.
Opening the box
The MFJ-1026 has several sockets on the back panel for interfacing to your transceiver. There is an input for the main antenna and an output to the transceiver. Because the unit is installed between the antenna and transceiver, an RF sensing TX/RX switching circuit is incorporated. However, the instructions warn that this may not work fast enough to avoid damage to the unit when high power is used, and recommends using hard TX/RX switching from a PTT output such as would be used to control a linear amplifier. An RCA jack is provided for this PTT input.
The MFJ-1026 may use either an external noise sensing antenna or an internal telescopic whip with built-in preamp. Both SO-239 and RCA sockets are provided for the noise antenna. There is also a socket for an external 13.8V DC supply. A power cable is provided, but no mains adapter.
Some advertisements for the MFJ-1026 describe it as an active antenna. However the amount of signal received using the internal whip is pathetically small. It is barely enough to enable noise canceling to be achieved on 20m, and completely useless on lower frequencies.
The telescopic whip pokes inconveniently out of a hole in the top of the case, and must be installed by the user. (A whip on a right-angled adapter that plugged into the socket on the back would have been simpler.) There are also some internal jumpers that need to be set, depending on whether the internal or external noise antennas are to be used, and whether the preamp is required for the external antenna. This necessitates removing the cover to install the whip and check the jumper settings.
Internal inspection of the unit confirms that all the comments made by eHam.net reviewers about the MFJ build quality are true. Several SMT parts are mounted crooked, there are some decidedly dodgy-looking cold solder joints, a lot of solder flux splashes, and the ferrite transformers look like they were wound in a hurry. There is no attempt to maintain a 50 ohm signal path for the received (and transmitted) signal through the unit, and the SO-239s are connected to the circuit board using some unnecessarily long and untidy wiring. The picture above was taken after I shortened the lengths of hookup wire to make the connections as direct as possible.
Fortunately, I have an Elecraft K3 which is fitted with the KXV3 transverter interface board. This provides the option of routing the receive path out the back of the transceiver and in again, enabling an external preamp to be easily installed - or a unit like this one. This means that the transmitted signal doesn't need to pass through the MFJ-1026 at all, nor does the receive signal on bands where the noise canceling isn't needed, like 10m. The K3 takes care of the TX/RX switching, and I can switch the 1026 in and out of circuit with the press of a front panel button. The K3 even remembers which bands I use it for.
Noise antenna
I can receive my interfering noise quite strongly in the shack on my portable receivers, so I had hoped to be able to use the MFJ-1026 internal telescopic whip for the noise sensing antenna. However, as mentioned earlier, it turned out to be pretty ineffective. Although it was possible to null out the noise on the 20m band, the gain on the main antenna signal had to be reduced to a very low level in order to do so, and I was probably losing weak signals in the process. The internal whip was no use at all on 40m or 80m.
I experimented with various external antennas such as random wires, but found that the signals they received were just as poor as the telescopic whip because they were not matched to the RF amplifier. An ATU was required to extract a workable signal from these non-resonant wires. In the end I hit on the idea of using the Wonder Wand portable QRP antenna, which is essentially a long telescopic whip with built-in ATU. You can see this in the picture at the start of this review. A Miracle Whip would probably work just as well.
Subsequently, I tried various different noise antennas including a 20m Hamstick, a SuperAntennas MP-1 and an ATX Walkabout. The ATX proved to be the best choice overall and was easy to retune when changing bands - just move the jumper lead.
Adjustment
The MFJ-1026 does need different settings for each band, and finding the correct settings that result in an effective null of the noise can be difficult. First of all, there is a FREQ button on the front panel that needs to be in or out depending on whether the frequency is above 10MHz or below it. Forgetting to check can result in failure to find a noise null, and is a frequent annoyance!
Next, you need to adjust the RF gain of the main and auxiliary antennas so that the noise level on both is approximately equal. It's best to use a clear frequency where only noise can be heard for this. Normally (but not always) the main antenna receives the strongest noise signal, so it must be operated at reduced gain while the auxiliary noise antenna gain is at maximum. I turn the main antenna gain down to zero and the auxiliary antenna gain to max, note the S-meter reading, then turn the auxiliary gain to zero and increase the main gain until the same level is reached. Then I turn the auxiliary gain back up to maximum.
Now you must adjust the phase of the two signals so that the noise component cancels out. There is a PHASE button which may be in or out. Rotate the PHASE control, and if no null is found, try the other setting of the PHASE button. You should hear the noise level go down and see a dip in the S-meter reading at this point. Try to find the point at which the noise is lowest.
Now fine-tune the adjustment of the gain control that is not set at max - usually the main antenna gain in my case. You will probably notice a further drop in the noise level. Go back and forwards between the PHASE and GAIN controls until the noise is nulled as far as possible.
Although this seems a long-winded process, once you have found the best settings they will not change much within a band, so you can note them down to get you in the ball park on future occasions. The exact null point is quite critical, though, so there is no avoiding the fine-tuning process.
Conclusion
As stated earlier in this review, whether the MFJ-1026 can eliminate the noise that is troubling you will depend on it being a single noise source, radiating from a single point. If it isn't, you might be disappointed.
In my case, it was possible to completely eliminate the arcing noise that I was troubled with. The best illustration is to listen to this sound clip made on 20m, during which I switch the noise canceler in and out. I think you'll agree that without the noise canceling, 20m is completely unusable.
The MFJ-1026 is expensive for what it contains and quite poorly made, so I don't feel any pride of ownership of it. It's also a hassle having to tweak its controls whenever I change bands. But pressing the button to take it out of circuit quickly reminds me just why I have it.
I'd probably had sold my equipment and quit the hobby if it wasn't for the MFJ-1026. This little box has kept amateur radio usable for me. So despite my criticisms, thanks MFJ for making it!
