Roof HF Attenuation

Attenuation topics

-- Return to Main Index. --

Roof Attenuation

Most of the RF penetration data on the internet was focussed on the frequency range of 1GHz to 5GHz, very little could be found that was relevant for 10-30 MHz. The North Cheshire Radio Club therefore decided to carry out a series of experiments to determine the attenuation at HF frequencies caused by the transmission passing through typical roofing material.

Roofs usually have a water proof membrane protected on the outside with tiles. Old types of Tiler's roofing felt was based on bitumen. Newer plastic types are made in three layers - a layer to withstand ultra violet light while waiting to be clad, a middle porous vapour barrier, and a bottom layer for abrasive strength. The overall thickness is less than 1mm.

Image - Eaves
Roof construction.
Image - Concrete tiles Image - Slate tiles
Concrete tiles                                             Slate tiles            
Roof Tiles Ingredients
Concrete - Sand, concrete, pigments, recycled aggregates and binders.
Clay - Clay, sand, pigments, additives.
Slate - Natural slate, or slate granules, resins, pigments and additives.

The vast majority of roofing tiles are made of concrete. Experiments were therefore carried out using concrete tiles (Marley - Ludlow Plus), and also with natural slate tiles (approx. 3mm thick) for comparison. A non-resonant field strength meter was built with a small rectangular loop antenna, and placed on a wooden table in the Club car park approximately 20m from the 14MHz Cobweb transmitting antenna.

Placing a corrugated cardboard box or a 1mm thick polythene box over the field strength meter made no difference to the received signal strength. The cardboard box was then cut down to a suitable size to support the tiles used to build an enclosure for the meter. Repeated tests with and without the tiles showed that the attenuation was less than the experimental error of 0.4dB, where 1 'S' point is equivalent to 6dB.

-- Return to the top. --

Field Strength Meter

The field strength meter for the attenuation experiment needed to be broad band so that it would not be unduly effected by its surroundings. It also had to be sensitive enough to be used with a small antenna that could be enclosed within a box formed by a few tiles.

The chosen circuit was based on the RF probe described in Ref.2 but without the complication of the micro controller. The differential output of the circuit was manually set to zero before each set of readings. The dc output signal was measured by a digital voltmeter set to read 1000mV placed outside the test enclosure. The layout of the circuit is not critical and was built on a strip of veroboard almost exactly as it was draw out in the schematic. The only deviation was in the use of 74k resistors throughout and paralleling them when half the value was required. The antenna was a self supporting rectangular loop of 18 swg enamelled copper wire approximately 27 x 15 cm.

Image - FS meter

The connector on the left goes to a terminal block supporting the antenna loop. The potentiometer on the left (1k linear) adjusts the antenna gain, whilst the the potentiometer on right (10k linear) is used to zero the meter reading. The connector on the right serves as a switch for the two 1.5 V batteries joined in series in the holder.

The maximum sensitivity setting produced a differential output signal of 1033mV dc when the field strength meter loop was face on and 20m away from a horizontal dipole antenna transmitting 20watts CW on 14.1MHz. The zero setting drifted approximately +/-5mV over a 5 minute period. This was sufficient for our purposes but the drift could have been improved if the circuit board had been enclosed in a small box to screen it from any draughts. It would also have helped if the three transistors had been placed closer together and clamped to a strip of aluminium to equalise and buffer their temperatures.

-- Return to the top. --

Circuit and Simulation

Image - FSmeter
Field strength meter circuit.

The voltage picked up by the antenna is represented by V1, where L1 is the leakage inductance of the loop. The component R7 stands in for the 1k variable resistor which sets the antenna gain. The rest of the circuit is based on that of a current mirror where the transistors Q1 and Q2 are biased so that they carry the same standing current as is in R2. The dc output voltages at A and B are set to be the same when there is no signal being received by adjusting the 10k potentiometer represented by R12 and R13. If there is not enough range in the potentiometer to achieve zero balance, try changing either Q1 or Q2 for another BC547B to get a better match. Transistor Q1 is used in a very non-linear part of its characteristic. When a signal is received the collector current increases much higher above the standing current than it falls below it causing a net negative voltage drop at 'A'. The current ripple is smoothed by C2 to give a steady dc output.

Image - FS meter traces
Field strength meter traces.

The above time domain simulation of the meter circuit was done using LTspice (Ref.3). This is an excellent program which is free to all except power supply chip manufacturers - but it does have a steep learning curve.

References

Ref.2 - http://cappels.org/dproj/Azpage/azpage.htm
Ref.3 - http://www.linear.com/designtools/software/#LTspice

-- Return to the top. --