(With thanks to Kees Stravers. See his excellent site for more information)

The Senster used two Hewlett-Packard doppler units (with custom made gold plated antenna horns) to detect movement near its 'head'. This system worked on 10 GHz.

A spare unit was built, but it was never used. This part was found a short while ago (2001 ?) and is shown here:

The images below are scans of technical documents about the radar system that were sent to me by Peter Lundahl. The text, as far as I can make out, is as follows:

The Senster Doppler Radar

1. Introduction

These notes describe the doppler radar system supplied for fitting to the Senster. Included are the power requirements.

2. The System

The system can be described using the scheme shown in Fig. 1.

This shows two radars, A and B illuminating a single target X. Movement of the target produces doppler signals which are in general different in both amplitude and frequency in the two channels. The signal amplitude in this system, however is kept constant by using high gain amplifiers and allowing them to limit. Thus a squared waveform is produced most of the time. The frequency of this square wave is proportional to the velocity of the target X towards the individual radars A and B, as calculated on a vectorial basis. It was considered cost practical for the radars to be mainly restricted to detecting people moving radially towards or away from the Senster. Other movements are detectable but less easily classified. Indeed, even with radial movement, there may be some confusion when several separate movements occur simultaneously. Used as indicated, the ratio of the frequencies from the two channels provides useful bearing information, Thus if theta1 = theta2 then a person at point X moving along the direction CD will produce equal frequencies in both channels. If the two frequencies are not equal their ratio indicates the angle ratio theta1/theta2, To assist the frequency measurement process, a lower-gain output point, which is free from noise, is provided on each channel. This can be used as a sensing point to indicate the main output is limiting to a square wave rather than providing output noise.

2.1 Angling the Horns and Range

The optimum inclination of the horns has been determined by experiment. The range of the radars is about 8 meters.

3. Block diagram of a single channel

A more detailed description of one of the channels is shown in Fig. 2. [circuit diagram] As can be seen the microwave power from the Gunn source is routed by a 3dB coupler. The doppler detector is integral and placed in the reject arm where it receives about 50 microwatts of microwave poier, because of the natural unbalance of the coupler. Optimum DC bias for the detector is set by a preset potentiometer placed on the input end of the amplifier board. Signals out of the amplifier are clipped to 7 V p-p square waves, at both the low and high output points. Each channel has its own microwave generator and the two will naturally be sufficiently different in frequency to prevent their cross mixing resulting in a frequency within the pass band of the amplifiers. This would otherwise produce a spurious [unreadable scan]. A second type of interference can be considered in which [unreadable scan] radiates into the other, both directly and via a target. This allows [unreadable scan] microwave source to produce a doppler signal in both channels in response to target movements. However, when both oscillators are switched on, each channel has a preference for its own signal which is a result of frequency beating rather than the otherwise video detection. This problem, incidentally, rules out the use of a common microwave oscillator for the two channels.

3.1 Gunn Source

The Gunn Sources are Mullard CL8370 5mW ones set to a nominal 9.6 GHz. They each require a stabilised 7 volts supply at about 100 mA from a supply, which is separate from that of the amplifiers and that has not more than 1 mV p-p ripple. Ripple produces microwave AM which then appears as a detected signal at the amplifier inputs. If the Gunn supply is common to the amplifiers then the oscillation in the amplifier-Gunn loop can occur due to the amplifiers causing a similar AM.

3.2 3 dB Coupler

These are an MRL design only available in sample quantities. They are printed in stripline and use a Sylvania D5504B diode. As stated, Xtal bias is provided via the adjuster on the amplifier board.

3.3 Amplifiers

The circuit of one of these is shown in Fig. 3. [circuit diagram] Each amplifier has a TAA310 integrated circuit providing 85 dB of gain, followed by a single BC107 transistor providing an additional 20 dB of gain (Gain stated ignores differences between input and output impedance). These two gain levels are brought out to pins, labelled "high out" and "low out".

The overall frequency response is -6 dB at 4 Hz and 4 kHz. The lower frequency allows indication without loss down to a velocity of 200 metres an hour (i.e. 1/25 normal walking pace) ; the upper allows a reasonable square wave form when the amplifiers are limiting.

Xtal bias adjustment is provided on the front of the amplifier. This should be adjusted for maximum signal using targets small enough, or far away enough, not to cause the amplifiers to limit. If required these controls can be used for gain adjustment to match the sensitivity of the two channels.

3.3.1 Amplifier Power Supplies

The power required is 7 Volts 9 mA each and the source resistance should be less than 1 ohm. This supply has to be separate from the Gunn supply to avoid interaction and oscillation. These supplies are described in section H.

3.3.2 Amplifier Components

The electrolytic capacitors used in the amplifiers are tantalum ones and suitable Philips types are:

100 uF	15 V	C421AM/CP100
47 uF	6 V	C421AM/BP47
22 uF	10V	C421AM/DP22

None of the components need to be close tolerance except perhaps the BC107 where a large departure from average can cause the output DC voltage working point to be less than ideal.

Test Data (7.0 Volts HT)

High out 3.8 Volts plus/minus 0.5 V DC
Low out 3.5 Volts plus/minus less than 0.5 V DC
Square wave output from both is usually better than 1.2:1 i.e. 1.1:1
Square wave output voltage is approximately 6.5 Volts p-p.
Gain:	85 dB from TAA310
	20 dB from BC107
Total:	105 dB

Caution:

The use of a soldering iron, which is not earthed to the amplifier printed circuit board before use, particularly when used on the signal input and low output connection, will usually damage the TAA310. Such damage is usually obvious by the resultant 30 dB or more lower gain.

Plug in sockets are fitted to the board to facilitate such replacement, which would not normally cause or require additional adjustments to be made. No failures have occured except in these circumstances.