Jan Jozef Oksiuta
Brisbane QG62lk
Australian Amateur Radio Station

DC to light, homebrewing, minimalist antennas and projects, QRSS, QRPp and less
WIA_logo FISTS_logo CWQRP_logo
Home Projects - VLF-to-MF Noise Tracker

VLF-to-MF Noise Tracker, July 2014
Switchmode supply noise seen on the 630m band
Image on left - an SDR radio spectrogram of 630m band showing persistent interference throughout most of the band and extending well into the the NDB band below. Similar pattern of interference lines repeats itself throughout the spectrum from around 100kHz up to HF bands. This screenshot was acquired with a sharply tuned magnetic loop antenna, which accounts for the marked increase in gain around the frequencies of interest.

Recent introduction of the new MF amateur radio band of 630m (475 kHz) promised a lot new challenges and projects. My excitement in monitoring the early activity on the band was however somewhat marred by the usual scourge of any suburban neighbourhood - a plethora
of unwanted man-made noise. My receiving setup consists of a variety of active tuned loops and an SDR-IQ receiver. I was able to get some improvement with a mast-head (or "loop-head") preamp and balanced feed to the shack (Cat.6 cable). Some interference sources however still persisted, with appearance on SDR spectrograms suggesting some switch-mode power supplies being responsible.

Time has come to pin-point the sources, and to deal with them. An idea was born for a weekend project to build a simple "sniffing" receiver. Some of the design points are listed below:
  •  a direct-conversion architecture, allowing superior aural perception of broadband noise when compared with simple AM detection, and easily obtained selectivity.  Single-sideband selectivity is not required for this purpose thus further simplifying the project.
  •  broadest tuning range possible, allowing its use also in VLF range, and all the way to and including AM broadcast band. For simplicity, the VFO will have interchangeable plug-in coils to cover the required bands, and tuning will be separate for the front-end and the VFO.
  • selective front-end, in order to reduce intermodulation products, and a directional pickup antenna, small in size but offering good sensitivity. All these can be obtained by the use of a small FSL (ferrite sleeve antenna). Tuning ranges will be selectable by jumpers connecting sections of the FSL coil in series or parallel, with additional fixed capacitors as required.
  • no automatic gain control - further simplifies the project, and in any case the AGC would be counterproductive in a device designed to detect changes in signal intensity "by ear". A visual RSSI would be nice but I considered it an overkill for what was meant to be a cheap and quick project.
  •  last but not least - portable and comfortable to move around, with battery power, speaker and headphone output. This can double as a PC sound card source for observation of spectrograms. I ended up with a grossly oversized carry case from an old disposed medical suction device, with lots of foam filling up unused space and also cushioning up my priceless ferrite rods.

Noise tracker schematic. VFO coil/capacitor plug-in shown (130uH/680pF) tunes the circuit between 450-510 kHz.
Most of my "quick" projects are built on the fly - straight onto the circuit board. Proper documentation with computer-generated drawings etc would probably take me longer than the project itself - hence the handwritten notes. Right-click on the image to access it in full size.

Noise tracker receiver circuit diagram

Circuit decription

The ferrite sleeve loop acts as a directional pickup and a high-Q input tuning circuit. It consists of two contra-wound 49 turns coils on a PVC pipe 111 mm O.D. and using 0.43 mm diameter wire. Coils can be connected either in series or parallel and are brought to resonance by one two parallel sections of a common polyester 60 + 160 variable capacitor and an optional 220pF jumper-selectable fixed capacitance. Inside the coil there is a sleeve made of 29 ferrite rods 9 x 100mm in size, material S1. Individual coils have 1.1 mH inductance each, with 865 uH in parallel and 3.5 mH in series. With these values in their various combinations the input circuit covers the range of 130 to 960 kHz - from the 2200m, through the 630m, up to somewhere the middle of the broadcast band.

Two centre-tapped coupling turns around the main input coil feed the signal in balanced mode to the NE602 Gilbert cell mixer. The mixer also accepts signal from local oscillator. This contains two FETs, another tuning cap, and a plug-in coil for the desired band.

Audio signal from the mixer is available on pins 4 and 5. This is crudely low-pass filtered by the 10k/10nF RC network before being passed to the SSM2019 audio preamplifier. The rest is straightforward - a trimpot for initial volume adjustment, an LM380 audio power amp, headphone socket, small speaker switchable with a jumper. The lot is supplied from an SLA 12V battery switchable in and out of the circuit with another jumper. A three-pin regulator supplies 6V to the mixer chip and also a bias supply to the audio preamp.

Finished receiver. Since this is a project for temporary and infrequent use, little attention was paid to aesthetics...
The completed receiver


The sniffer was tuned to around 476 kHz - first the VFO, then the input circuit. The buzzing noise was instantly recognizable, sounding very much the same like the pesky interference on 630m. A walk around the backyard quickly revealed a quieter corner, previously missed in my trial-and-error attempts with the main receiving loop. (Doing portable tests with a laptop and SDR receiver were fruitless, as the laptop itself generated sufficient noise to mask all the others. One has to position the receiving loop, move back inside to check results, then repeat again and again.) The loop was duly repositioned and my reception of WSPR stations from down South was instantly improved.

More walking around the neighbourhood revealed that the source of noise was most likely confined to my property. It was narrowed down to the bunch of cables leaving the shack and going to the earth stakes, antennas, ELF equipment and whatever else I had connected. Separating the cables and "sniffing" them one by one revealed that the noise was carried by the earth cable. More walking was done inside the house and eventually led me to the entertainment unit in the bedroom, where the real foxhunt begun, involving moving the furniture and then individually unplugging all the appliances (and I barely use any of them myself!). That included the TV, video player, DVD player, TVSat box, cordless phone, video sender, broadcast radio... All of them were buzzing very happily at close distance. One of them however was not content with just that - it also had to broadcast it through the mains earth wiring! Finally, a trophy - a respectable brand DVD player. Interestingly, it was rather (electrically) quiet when turned on, but was coming back to its secret life as soon as turned off into standby mode. It is indeed very rarely used which would account for my very rare noise-free periods of grace.

The offender was duly eliminated from service pending further examinations, with a view to implement whatever control measures might be recommended.  I started my 630m reception with an SDR (now sooo beautifully clear), plugged the subject in, ensured it went to its standby mode, and..... nothing. Not a squeak. Then of course - the player is Class II device, i.e. double-insulated, no protective earth, and consequently the mains cable is terminated with a two-pin plug. No earth connection. The earth connection must be coming from the A/V cable that connected it to the TV.

I then connected an earth wire to one of the output RCA sockets and yes, the same wobbly bars started dancing again on the receiver screen.

Quick negotiations with the spouse and the device will now remain isolated until "absolutely" needed, in the meantime I am going to probe it further. For now, I connected a 10 ohm resistor between the device's metal case and earth and applied a portable oscilloscope across the resistor. I was able to observe over 10 mV peak of broad spectrum noise - a clear indication of 1 mA of AC RF currents. More than enough to be seen by any sensitive loop antenna anywhere in the backyard.

1 mA of noise current to earth
There is about 1 mA of RF currents flowing from the player's case to earth with multiple frequency components

Interesting that this interference signal is ultimately earthed yet still producing interference. Another observation that puzzles me - the noise current appears unchanged regardless of the mode the player is in - yet the interference shows only when in standby mode. A lot of things out there that I do not understand.

Noise eliminated at last
Interference conquered at last

back to Projects
back to Home


Created with CompoZer and BlueGriffon. Last updated 2014-07-28