Reducing Local Noise Interference for LF & MF Reception

by Julian Hardstone - Feb.2005

The purchase of a new PC here at Omega House resulted in a desperate increase in noise on the MF bands, limiting MW listening and completely precluding DXing below 7MHz. The noise was present as long as the New PC (NPC) was plugged-in, whether or not switched 'on'. Indeed, it was clear that the NPC's PSU is always on, as evidenced by the optical mouse always showing its sensing light. The resultant noise appears to be broadband and significant up to at least 20MHz.

 For the first few days, the NPC dominated my attention, but after a week or so it was obvious that something would have to be done for the radio. The HF Receiver was set up adjacent to the PC with a remote ærial connected by a co-ax feeder. The signal levels due to the noise in a 4kHz bandwidth were measured on two normally quiet frequencies, 1611 and 3925kHz. When the Vatican appeared to be testing DRM on 1611, 1621 was used instead.

With the NPC off and unplugged (really 'off') the noise was at a reasonably low level. With it plugged in or really on, the noise was 9 + 6dB on 3915 and S6 on 1611. It was found that the CRT monitor does not contribute anything on these frequencies (well done that emc engineer!).

A large professional RF mains filter was to hand, and was set up in the NPC's mains feed. The peripherals were tried both through and not through the filter, and it was found that there was a useful reduction in received noise, the noise being slightly lower with the monitor and other peripherals NOT fed via the filter.

















Various simple filters were then tried, starting with the mains cable wound round various ferrite toroids, then with the mains supply taken through various ferrite inductors, with the earth cable connection made directly or via the inductors, and it was observed that, generally, the more the L the less the Pn, and a direct earth was better than one via an inductor. The optimum required a small earth inductance, obtained by winding a few turns on a toroid.

Rather than commit the rather fine mains filter to serving the NPC, I wanted to make one specifically for this use. It then occurred to me that I had a suitable mains-filter type of common-mode dual choke available on an old PC monitor board that I'd recently scrapped. Closer inspection showed a number of suitably rated Y-type filter capacitors arraigned around the choke, so it was decided to cut out the section of the board which carried the mains filter. An unbroken earth connection was made through the assembly, which was fitted in a suitable Maplin plastic box.

The effect was impressive, with the noise levels from a remote æerial no longer raised from the 'off' condition. I suspect that most recent CRT monitors include effective mains filtering, so it would be well worth salvaging the components or even the relevant section of the board from any scrap monitors you find. With the whole world rushing to upgrade to TFT monitors, small CRT monitors are ten a penny, even when they are working. Those made in the past eight years or so should be emc compliant, implying effective filtering. 

Additional benefit was obtained by winding the power and audio leads of the receiver through ferrite toroids for 10 – 12 turns to form toroidal chokes, as in the final picture below.

Some measurements:


Filter                             1611kHz           3925kHz

No filtering                     S6                    S9+6

Belling-Lee unit              S1                    S8

4turns 1" toroid              S6                    S7.5

2×330µH, earth thro'       S1                    S8

3×330µH         S5                    S9

'Monitor' mains filter        S2                    S5.5




Vertical Rod Aerial

Having seen some success with the mains filter, it was felt that it would be opportune to put up a modest vertical ærial in the garden, suitable for use from LF to HF and most particularly for LF, MF and the 'tropical bands'.

If your living space allows you the room to put a small vertical aerial in a garden, with a distance of at least 20' from the nearest occupied rooms in your house and your neighbours', then you may be able to achieve a reasonable freedom from all those contemporary sources of noise - QRN - such as TV's, PC's, economy light bulbs, house alarms and the rest. The best rejection of local noise in an urban environment will probably be achieved only with a loop antenna. If a loop is to be mounted remote from the receiver, it really needs to be an untuned loop with a matching amplifier, such as one of the Wellbrook products, or to a design such as Marco Bruno's:  But that's a whole different project from mine.

My vertical is a vertical rod of height 20ft, using a basic ½-wave CB antenna. In place of the original resonant tuned matching transformer in the base, a broadband ferrite transformer is used to give a compromise match to 50 ohms at LF and MF, and also gives useful performance throughout the HF spectrum.


 A transformer being wound

The impedance of a short monopole is relatively high and capacitive, and although the resistance component is low (relative to 50W) good sensitivity can be obtained with a high resistance load. Car radio installations now commonly use a very short ærial length with a MOSFET input stage to provide a high resistance load and achieve good sensitivity. The approach here is to use a much longer ærial, with a ferrite transformer to give a moderately high-resistance load, matched to 50W without any active devices.

The Matching Transformer was wound on 13.5 x 7 x 7mm two-hole core, made by Epcos (formerly Siemens) using their N30 grade ferrite, and available from R-S Components as stock number 212-0617

Testing was then carried out with a signal generator and a wideband ac millivoltmeter performing much the same function as an oscilloscope, but reading in dB relative to a datum. It was determined that 4 turns was sufficient for a 50ohm winding at 150kHz. A ‘turn’ consists of one turn around the central core, going in one hole and out of the other. 

Testing of the transformer was carried out from LF to HF

The most suitable turns ratio for the 18ft vertical ærial was chosen after measurements of the received signal strength with various ratios. It was expected that the required load resistance should be in the region of 500 to 1,000 ohms, and in the event a turns ratio of 20:4 was chosen, which implies a load of 1250 ohms.

Sensitivity down to the noise floor is all that is required from a receiving ærial, and this is easily attained with this long-ish vertical. Because the load resistance is relatively high, there is no requirement for a low-resistance earth connection.  However, where noise from the PC is a problem, it may be   worthwhile driving a standard electrical earthing rod into the ground adjacent to the ærial.



This ærial is quite portable; with a length of stout steel pipe to drive into the ground, it can be set up in a minute. However, rejection of indoor interference may benefit from a better earthing arrangement, so be prepared to experiment with one or two electricians’ earth rods.


The transformer could alternatively be used with a wire ærial plus an earth rod. It may be worthwhile trying different turns ratios for a different ærial length. A shorter ærial should need more turns, and somewhere between 15 and 30 turns should suit most installations