DF6NM kite activity June 19

Vos écoutes - Vos contacts - Vos observations - Vos essais
Messages : 1553
Inscription : 10 mars 2008, 23:07
Localisation : Ris Orangis JN18FP

DF6NM kite activity June 19

Message par F4DTL » 20 juin 2011, 23:17

Dear LF, Paul,

here's a short summary of today's VLF kite transmission. I'm sorry for the late notice, but the experiment had not been planned ahead, and I wasn't sure if I could get on at all. Also, the kite I am using (a 0.8 m^2 delta toy bought for 4.99 Euro) is a bit small for this sort of activity, and really requires a good and steady wind.

The QTH was actually at 49.3874 N 11.1692 E, a few 100 meters from the former location reported earlier. The GPS-controlled carriers were on air
11:15 - 12:15 6470.000 Hz ~ 50 uW, and
12:30 - 13:10 8970.002 Hz ~ 200 uW (EMRP).

Getting on 6470 was easier than expected. I simply used the same "seven-bucket" coil, which resonates my 240 pF home antenna to 8.97 kHz. Connected to the kite wire, the resonance was close enough to 6.5 kHz to be tuned by the ferrite plate variometer. Then the drive for the audio amp was increased until the onset of corona sizzling at 0.25 A. At that point, the amp drew 10 A DC from the car battery and probably delivered around 50 W. After connecting mobile internet and staring at Stefan's grabber, I was positively surprised to find a faint trace there!

For the frequency change, I removed the two bottom coil buckets, which brought the resonance to ~ 8.5 kHz. Then I had to carefully pull the remaining 5 buckets apart, inserting bits of plastic foam material as spacers. Going through a few iterations took about 10 minutes, and I got to "full power" (380 mA, 100W) at 12:30. Unfortunately, the wind had started to calm by then, and there were several dips during this transmission. Finally at 13:10, the kite landed gently in the grass, and I decided to close down and pack up early.

Traces from both transmissions were visible on DK7FC's 4.5 mHz "600" windows, and the second one also produced a nice dash on the 0.45 mHz window. It is likely that these signals would be useful for a kite-to-kite QSO on either frequency, using 10 minute dashes.

The screenshot from F1AFJ has an interesting dash before 13:00. The indicated frequency seems a bit high (8970.020 instead of 8970.002), which could perhaps be due to a small samplerate calibration error. I am also not sure about the time scale - judging by the width of the noise blobs, the setting could have been 11 mHz and 10 seconds per pixel. It would be nice if Jean-Pierre could confirm the frequency calibration and scroll settings.

I also carefully inspected the OE3GHB grabber, which runs at 1.4 mHz resolution and seems to be very sensitive. It showed an unusually dark period of very low noise between 8 and 11 UT today. But no luck for my transmission there yet.

Well, in addition to the technical aspects, being outdoors in the warm wind was again a rewarding and enjoyable experience.

Best 73, and thanks again for the interest.

Markus (DF6NM)

Dear Dreamers,

thanks very much to everyone who commented on my signal yesterday night, and all who are running their VLF grabbers.


My signal was on air on 8970.002 Hz from 18:52 to 22:01, with a long break from 19:26 to 19:52 due to a prolonged rain period, and many shorter interruptions caused by precipitation, low wind, a coil falling over, or a breaking antenna connection.

I'm aware that an advanced notice would have been valuable for potential receiving stations. But I find it difficult to plan ahead this type of activity. Yesterday even while setting things up on the site, I was not at all sure whether I would be able to actually get on air for any significant time.


The location (49.3874 N 11.1692 E, "JN59OJ02HX" for those who prefer cryptic notations ;-) was the same as in the experiment on Nov 13 last year (see email copy at bottom). While Google Maps still shows large patches of unused land around the former US airfield, "my" field is now one of very few undeveloped plots in the middle of an industrial area, inhabited mostly by large logistics companies. Though I had originally not intended to work into darkness, I found there was still plenty of light from the surrounding truck parking and unloading areas.

Kite antenna and ground:

This time I used a 4 sqm lifter kite (HQ Power Sled L3.0, similar to Stefan's smaller kite). It is held by a 150 m line. This may be stretching the limit a bit, but due to the slope and sag of the line I am sure that the peak altitude always stayed below 100 m.

The previously used 0.4 mm magnet wire had a voltage limit of about 12 kV due to corona appearing along the whole length. It has now been replaced by insulated wire (0.25 sqmm, outer diameter 1.3 mm). The total length is 180 m, with the last 30 m dangling down from the kite, intended to act as a sort of extra top load. Effective height would be around 50 m, giving 3.5 milliohm radiation resistance, or 1.3 mW EMRP at 0.6 A.

At the western end of the unused site, there is a long mesh wire fence running north-south, held by solid metal posts which served very well both for anchoring and grounding. Total resistance of the antenna circuit was around 280 ohms (including 140 ohms for the coil). This is about half of that when I used the same kite antenna in Bamberg with only a few short ground pegs.

The coil was placed below the kite line at about 3 m from the anchor point, hoping to keep the wire well above ground. However this turned out to be problematic due to the horizontal and vertical movement of the connection point. When the kite was low and due west, the wire was too long and touched the wet grass, exhibiting pretty little fireworks. When it went up or drifted aside, the wire became too short, and either plugged itself off, or tipped over the whole coil. You guess how much I like having to reconnect the ungrounded wire, with dark clouds overhead and a slight drizzle going down... I twice found myself holding down the kite line with my foot, then slowly stepping forward until the open end at least touched the earth. Next time I will definitely place the coil much closer to the anchor, and try to hold the connection up by means of some elastic plastic tube.


From the previous experiment I knew that the wire capacitance was very close to 900 pF. During the afternoon, I experimented with different coil configurations, and found that I could either use four buckets (184 ohms) with a large 10 cm pitch (ie. 9 Lego blocks), giving about 8 mm radial clearance between one coil layer and the inner radius of the next bucket. The other option was three buckets (138 ohms), stuffed into one another 3 cm pitch and 2 mm clearance (see picture ).

Mainly for mechanical reasons, I opted for the more compact three-bucket solution, but this turned out to be the wrong choice. With power applied, I heard corona sizzles around the coil at about 0.55 A (11 kV). In darkness I could see violet seams near the bottom end of the upper two windings. This means that the maximum usable layer-to-layer voltage for this design is about 3.7 kV. Next time I will definitely go for the longer version, which also has more surface exposed to the wind for heat removal.

BTW all seven buckets connected to 900 pF resonate at about 4.6 kHz - Stefan, how about a 65 km band ;-)

Amplifier and power supply:

Like before, I used one channel of a class-B car-audio amp (Nitro 3000), feeding a 12 : 80 turns ferrite transformer originally wound by DJ2LF. It delivered 0.6 A antenna current (~ 100 W) from 12V x 13 A DC input. It was possible to turn the volume up to 0.7A, but I did not dare to run that level of corona permanently. The amp was fed by a standalone 35 Ah lead-acid battery, which was repeatedly recharged via jumper wires from the car engine during the latter half of the experiment.

Signal generation:

SpecLab's samplerate was locked to DHO, which is very slightly (3 ppb) above its nominal frequency 23.4 kHz. To keep the average phase steady, I took care to never interrupt the software generator whenever I had to take the carrier off air. Anyway the time series recorded by Paul Nicholson shows two major phase jumps around 19:40 (while the TX was off) and 21:20, perhaps due to soundcard buffering issues. One problem I did not see immediately was that the DHO ferrite antenna was too close to one of the jumper wires, temporarily picking up interference when the car alternator was running - once found this one was easily fixed.

Though the traces at TF3HZ (2547 km) and YO/4X1RF (xxxx) are close to the noise, they are at the right time and frequency, and I do not think that they are random pixels. Halldor's trace even seems to show the outage period after the first bright dot.

On my own grabber, I attempted to compare signal levels from DK7FC (-71 dB, 180 km) and myself (-43 dB, 4.5 km). Correcting this difference (28 dB) by the EMRP ratio (26 mW / 1.2 mW = 13 dB), distance ratio (1/r = 32 dB) plus an additional nearfield correction (1/r³ instead of 1/r from 4.5 to 5.3 km = 3 dB), we find that Stefan was 28+13-35 dB = 6 dB weaker than we would expect by pure groundwave. This corresponds very well to the earlier experience that there is a midday skywave cancellation effect for about 200 km.


Revenir à « Trafic sur 33333 mètres »