Midi-Loop Test

The Italian manufacturer Ciro Mazzoni is successfully operating on the international market for many years with its magnetic loop antennas. The company supplies both radio amateurs and commercial users. The high quality of workmanship of the antennas and the way in which they tune themselves automatically at the touch of a button represent an innovation in the field of magnetic loops. The maximum transmission power allowed for the Loop antennas is also exceptional and triggered the interest of the author. According to the manufacturer, the MIDI loop can be operated between 3.5 MHz and 8 MHz with a maximum power of 300W and above that up to 14.5MHZ with 800W. After a short consideration, the antenna was ordered for a test and delivered by Ciro Mazzoni via freight forwarding in October 2019.

The antenna has a diameter of 2m and is made of aluminium tube with a wall thickness of 2mm and a diameter of 75mm and weighs around 20kg. Due to the dimensions, it is disassembled into four parts and mounted on a wooden construction, which in turn is optimally protected by a huge cardboard box for transport by forwarding agent.

Fig. 1) Transport packaging of the Midi-Loop


Since the construction was decided on very short notice in mid-November, the author had no time to organize a team of helpers and the question arose whether it was possible to erect such a large antenna single-handedly. Four benches were used to comfortably lay down and assemble the individual parts of the loop (see pictures 2 and 3).


Fig. 2)

Fig. 3)

The four-part loop has two flanges to which it is screwed with seven M8 screws with self-locking nuts. Ciro Mazzoni supplies a special conductive copper paste that is applied to the contact surfaces before joining the flanges. This paste increases the electrical conductivity and prevents oxide formation at the contact points.

The drive for tuning the capacitor of the loop is done via a 24V push rod motor. These push rod motors are tuned very precisely in pulsed operation, they are weatherproof and they are excellently suited for retuning the antenna via the control unit during QSY operation. The control line of the motor is connected via a cord already prepared by the manufacturer. The end of the cable is attached and can be pulled through the tube to a small weatherproof junction box. The 2-wiree cable required for the connection between the control unit and the antenna is also fed into this box and connected to the drive cable via a toroidal choke supplied.

The choke prevents too much HF from reaching the control unit on the control line during operation. After mounting the mast holder, together with an approx. 50 cm long mast tube, the help of the XYL was required very briefly to erect the antenna and place it in the prepared ground sleeve for the first function test. After that, the connection (motor control line and coaxial feed line) to the control unit was completed. After switching on the control unit, the desired operating frequency was entered via a small USB numeric keypad connected to the control unit and the capacitor began to move.

Fig.  4) First test 0.5m above ground


To the astonishment of the author, the antenna worked right away and despite the low installation height with an acceptable VSWR of 1.8:1 in the 80m band. One after the other, the higher bands up to 20m were tested with success.

Now came the somewhat more strenuous part, namely to bring the antenna alone (the author's XYL had left the house in the meantime) into its final position. Ciro Mazzoni recommends an installation height of 2.5 to 3.5m, for this height the adjustment of the MIDI loop is optimised. By the way, Ciro Mazzoni specifies a frontal area of 0.5m2 for the MIDI Loop, so the author decided to choose the recommended minimum set-up height of 2.5m. We had an older portable mast by Flammex Pull in the garden, which now acted as an auxiliary mast to lift the Loop into its final position. The auxiliary mast was positioned as close as possible to the mast base (ground sleeve) of the Loop and a pulley for the pull rope was attached to the mast.


Fig. 5) Auxiliary mast for raising the loop


A centrally mounted loop to the left and right of the loop condenser served as an attachment point and was connected to the hauling rope.

A small rotor is sufficient to rotate the MIDI Loop, which according to the manufacturer has two preferred directions with zero points offset by 90°. Due to the considerable bending moment of 1680N/m, however, a structure with an upper bearing is indispensable. The author's loop is located in the garden and is protected from the frequent westerly wind by a hedge and the neighbouring house.  For this reason, "only" a 1.2m long ground socket with an outer diameter of 75mm was used to anchor the mast, which, however, could only be screwed into the very hard ground with great effort.

Fig.  6) It is done

Fig. 7) The automatic tuning unit of the Midi Loop on the IC-7610


Shortly before dusk, after almost 7 hours, the installation was completed and the first tests could begin. The VSWR was immediately <1.6:1 over the entire operating range and the first tests with WSPR on 80m could begin. According to the manufacturer, the antennas have a slight preferred direction towards the Gamma Match feed and were pointed towards North America in the hope of reaching a few stations on the East Coast.

The next morning (it was Saturday 16 November) blew away all expectations.  In addition to numerous stations from Europe, other stations from Newfoundland to Texas were added to the WSPR log, as the map section shows. The transmitting power was 5W. Another test with WSPR was then started on 40m.

Fig. 7) Yield after one night in WSPR on 80m


On Tuesday morning, after a total of 24 hours, this picture emerged 

All continents could be reached on 40m with 5W.

Figure 8) WSPR yield after 24h on 40m

On the following weekend, the MIDI loop was operated exclusively on 40m and 80m in the CQWWDX - CW contest. In approx. 5 hours of comfortable search and pounce, 70 multipliers could be reached with 100W transmitting power. On 80m, numerous DX stations from the USA East Coast and the Caribbean found their way into the log, 90% of them on the 1st call, which certainly resulted from a passable signal.

Metrological evaluation of the MIDI loop

To measure antenna performance more accurately, the MIDI loop and an Inverted V dipole were each operated simultaneously with a 200mW WSPR transmitter (WSPRLITE Flexi). The Inverted V has a length of about 2 x 19m, the feed point is at about 17m above ground level and the ends are about 8m above ground. This inverted V dipole is fed via an approx. 8m long twin wire with a symmetrical tuner (Christian coupler by DL3LAC). The frequency spacing of the WSPR transmitters is only a few Hz (e.g. approx. 30Hz during the investigation in the 40m band). This means that frequency-selective fading can be virtually excluded and, due to the simultaneity of the transmission, also temporal fluctuations of the signal, as is usual on shortwave. However, the comparison of two antennas at the same location also has its pitfalls due to the coupling of the antennas.


Although the MIDI loop is 12-13m away from the feed point of the dipole and approx. 8m from the dipole ends, with the same orientation (both antennas radiate east/west) only a decoupling of 18dB can be achieved in the 40m band. If the loop is rotated by 90°, the decoupling is already approx. 40- 50 dB, but this does not allow a meaningful comparison of the performance in the main beam direction.


Nevertheless, the decoupling should be sufficient for our test, because only about 1/50 of the power of the other antenna is coupled in and radiated. In the 30m band the decoupling is already approx. 23dB.


Decoupling MIDI loop to ref. antenna


  Band Decoupling/dB
   80m    17 dB
   40m    18 dB
   30m    23 dB
   20m    28 dB


Measurement with WSPR on 80m

The DXplorer offers the possibility to select only the common spots from a definable minimum distance. In the following histogram, the minimum distance was limited to 1000km.

It can be clearly seen that the differences between dipole and MIDI loop become smaller at stations of 1000km and more.

The dipole has an advantage of only 2.18dB with a standard deviation of 3.1dB.


In the following histogram, to clarify how well the antenna works for DX traffic on 80m, only common spots from 5000km distance have been evaluated.

The difference to the dipole for these distances is only 1.3dB in favour of the dipole with a standard deviation of 1.3dB. By the way, the most distant station K9AN (6991km) from the DX10 list could only receive the signal of the MIDI loop, which clearly speaks for its suitability as a DX antenna.

With the DXplorer you can also make comparisons with other stations.

This histogram shows a comparison of the MIDI loop over 645 common spots with my radio friend Ewald DK2DB in Karlsruhe. Ewald operates a shortened rotary dipole with 2x6.55m element length, which is adapted via a very low-loss network and achieves an efficiency of about 60% in the 80m band. In this comparison of SNR, the MIDI loop is -5.5dB, but the difference in transmit power must be included here, as Ewald operates his WSPR transmitter with +37dBm (5W) transmit power. The difference to +23dBm at the loop is 14dB and leads to a signal advantage (-5.5+14dB=8.5dB) for the loop.

Of course, it is difficult to compare antennas over this long distance, especially since there are different propagations and different conditions of the QTHs (topography, ground conditions, buildings, etc.).

Measurement with WSPR on 40m

On the 40m band, the MIDI loop drops significantly compared to the author's dipole.

In the DX10 performance, the dipole reaches an average distance of 9868km and the MIDI loop 4845km. This is the evaluation over one week. Worth mentioning are 94 spots for the dipole of DP0GVN, the loop managed one spot in this time.

In total, almost 3400 common spots were evaluated on 40m over a whole week and here the dipole has the advantage with 7.1dB with a standard deviation of 6.15dB.

If we take only the common connections off 2000, even the loop is ahead by 0.17dB with a standard deviation of 6.2dB. This difference can probably be explained by the stronger directivity of the Loop.

Measurement with WSPR on 30m

On the 30m band, too, the comparative dipole can clearly score.

The MIDI Loop is 7dB below the comparison dipole in the 30m band with a total of 536 common spots, with a standard deviation of around 4.5dB.

The dipole managed 24 spots from DB0GVN over 1.5 days, the MIDI loop unfortunately none. Nevertheless, the loop got 2 spots from W3HH over 7500km distance as ODX.

Overall, the picture is similar to that in the neighbouring 40m band

Measurement with WSPR on 20m

Unfortunately, the conditions on the 20m band were quite poor on 28/29 November. Overnight the band was completely closed (at least for 200mW transmitting power) and so only in the late morning of 29.11 the first spots came. In total, after filtering out some local spots, only 38 common spots remained. The MIDI Loop was 3.89 dB below the comparison antenna. The standard deviation is 3.2dB.

In order to get a better data basis for the 20m evaluation, the 20m operation was repeated at a nine point in time.

As can be seen in the DX10 table, the dipole is slightly ahead of the MIDI loop.

Over 627 spots, the advantage of the dipole over the MIDI loop was 4.3dB with a standard deviation of 3.65dB.

If we only consider DX connections >5000km, the dipole has an advantage of 4dB with a standard deviation of 2dB. The standard deviation decreases because the DX connections were almost exclusively achieved in the main beam directions of the two antennas.

To get some absolute numbers for the antenna gain from the WSPR comparison, the author's reference antenna was simulated with the help of EZNEC 6.0. DX-relevant is the radiation behaviour of the antennas in a low elevation. The gain of the reference antenna was calculated for an elevation of 20° and listed in Table 1.

Table 1:

Band RefANT/dBd MIDI Loop vs. Ref.Ant/dB MIDI/dBD


specification MIDI- LOOP/dBD

80m -2,9 -2,2 -5,1 -4,0
40m -0,8 -7,9 -8,7  
30m 1,4 -7,0 -5,6  
20m 4,6 -3,9 0,7  -0,3


The midi-loop reaches the factory specifications on 80m down to -1.1dB, although at this point it is important to point out the somewhat greater measurement uncertainty of the method. We recall that the mean deviation at 80m over the spots >1000km distance had a standard deviation of 3dB. On 20m the deviation from the manufacturer's specifications is also in the range of 1dB, this time in favour of the MIDI loop, but even here we had a standard deviation of a good 3dB and due to the moderate conditions only 38 spots. On 40m and 30m, the reference dipole does quite well for DX and has less directivity than the MIDI loop with its quite pronounced nulls, which certainly puts it at a bit of a disadvantage in this test.

Matching and bandwidth 

80m Band

The antenna was tuned to 3610 kHz, resulting in a VSWR of 1.26:1 and a VSWR 3:1 bandwidth of 3.81 kHz.


40m Band

On 40m, the VSWR 3:1 bandwidth is already 14.2kHz and the VSWR is 1.04:1 in the middle of the band.


30m Band

The VSWR 3:1 bandwidth in the 30m band is 26.2kHz and the VSWR in the middle of the band is 1.24:1 20m band.


20m Band

 In the 20m band, the 3:1 VSWR bandwidth is already 76.4 kHz and the VSWR in the middle of the band is 1.3:1.


Before starting this test, the author did not expect the loop to match the performance of the Inverted V comparison dipole on 80m. Obviously, it is characterised on this band by a flatter radiation compared to the dipole, which is clearly conducive to DX traffic.

DX traffic is also possible on the higher bands, but the author's Inverted V outperforms the MIDI loop quite significantly. During the tests, a strong influence of weather and humidity on the tuning was noticed and so, for example, the loop had to be retuned after the onset of rain. Humidity also plays a role when operating with higher power levels and so, with a damp antenna, a power level of 250W in the 80m band and 500W on 20m could not be exceeded without voltage flashovers becoming noticeable at the capacitor. Basically, this antenna is very well suited for 200W transceivers.

Operation with higher power should only be ventured in dry weather. The VSWR is optimal at a mounting height of 2.5m on all bands, as the measurements show. Unfortunately, the author's antenna showed that the minimum frequency is 3510 kHz. At lower frequencies, the controller was no longer able to tune the antenna, but did not stop the attempt either, so that a restart of the controller became necessary.


Since very high magnetic field strengths occur with a magnetic loop, the EMC protection distances must be checked carefully before operation (e.g. with Watt32). The MIDI Loop by Ciro Mazzoni is an interesting alternative for radio amateurs with limited space and already works very well at low ground level. It should also be noted that the amount of material required for the magnetic loop is quite low compared to the 3-mast solution for realising the author's Inverted V. The magnetic loop's extremely narrow bandwidth of 1.5 watts makes it ideal for use in the field. Due to its extremely narrow bandwidth of a few kHz, hardly any strong interference signals reach the receiver.


Simulation of the reference antenna with EZNEC


On 80m we have only just -4.3dBD in an elevation of 20°, the horizontal diagram shows as expected a slight preferential direction in the direction of East/West.


At 40m we already have -1.7dBD in the 20° elevation, the maximum of 1.8dBD is radiated at 50°.

The horizontal diagram shows a maximum 70° offset from the presumed main beam direction, but only 0.2dB higher.


And on 30m we achieve a slight gain of 1.67 dBD for 20° elevation and 2.70dB at 30° The slope is - 6dB related to the maximum at 30°.

In the direction of the west (0° in the plot) only -1.7dB is reached, the maximum is 85° offset in the direction of the north or south with 1.67dBD.


In the 20m band, 4.8 dBD is already reached in the main radiation direction at 20° elevation.