— PROJECTS —

Each pod has a 12V lead acid battery and these were to be charged from a single solar panel — A traditional PV shunt regulator with all 3 batteries connected in parallel would be wasteful so a regulator that switched the panel current to each battery as needed was designed and proved to be more efficient as the panel current was always being directed to charge one of the batteries

The control system and amplifiers for the pods had to be be very efficient so Class D (Modified Tripath) amplifiers were used and to further increase efficiency the "H" bridge supply was envelope modulated using information from a custom modified MP3 player ahead of the signal output

As part of the celebrations for "Liverpool 08 European Capital of Culture" We provided design consultancy to the artist Peter Appleton for laser links between the two Liverpool Cathedrals and the magneto acoustic drive for the stringed instruments used within the Cathedrals

The original intent of the project was to transmit audio between the two cathedrals using analogue modulated green lasers that I had designed when working for Vector Technology but the fast modulation required was not possible with laser powers higher than ~50mW and the visibility of these was considered low

It was also considered that 2 parallel visible lasers would look awkward so we installed a separate infra red digital communications link which interfaced to the rest of the project digital audio equipment easier and later installed a parallel 2W visible laser shinning from the Anglican tower to the Catholic cathedral

Summer 2008 the 2 churches were connected with beams of low power infra red light transmitting digital audio signals between their sound systems — The intent of the project was "highlighted" in September 2008 with a 2W green laser along the same route — see BBC News — programme

The picture on the right was taken from a balcony inside the Anglican cathedral while we were investigating cable runs between the laser digital audio link equipment in the tower and the sound booth at ground level

This large area of coloured glass made an ideal termination for the high power green laser which along with other safety precautions at the Anglican cathedral end ensured the system could not cause harm to anybody

September 1998 — I was contacted by an ex colleague working on a European Space Agency project in Zurich because the group he was working for could not get anyone to meet their spec. for an aerospace VHF power amplifier — I had to admit the size and efficiency requirement was very tight and at this stage I was not even considering the limitations on component choice I would later find due to the aerospace qualification required

After many emails and a few visits to Switzerland and reading through my old BBC engineering notes and text books I realised the efficiency was not such a problem and the small size could be met but not using PCB stripline matching — I modelled several sample designs using lumped component matching and quoted for the first stage of the project to produce a design that functioned as a SPICE model

My input was a very small part of a large project and it was not until April 2004 that I got to actually build the first prototypes (left) and supply 2 working devices for testing on the Atomic clock at Neuchatel

The very small prototypes looked a bit messy but the size and performance more than made up for the appearance — The only problem making the final article would be finding non plastic inductors and no variable components and making a PCB with copper filled vias that would pass a vacuum and temperature test

The amplifier was to be used in space and driven from a VHF oscillator whose output level and frequency would be carefully controlled to provide a specific level of hydrogen plasma and its full title was Space Hydrogen Maser Hydrogen Dissociation Oscillator Power Amplifier or SHM HDO-PA

Many of the inductors (now fixed value) were available as expensive space qualified parts and to overcome the need to tweak the lumped matching I was allowed to make open frame coils using silver plated copper wire and this further improved the efficiency over the prototypes

Apart from the usual tests like linearity - gain -- frequency response and power made with RF analysers — The final check is can it power a loop antenna to induce a plasma in a low pressure gas vessel

The answer as you can see is yes it can and it does this with much higher efficiency and in a smaller space than the very tight specification required

The frequency required to get 1,400,000,000Hz which divided by 14 = 100MHz (the clock output) and two other frequencies to automatically tune the clocks narrow band sapphire loaded cavity (ACT) were generated by direct digital synthesis (DDS) using the 100MHz output as the DDS clock — The 2 ACT signals either side of the hydrogen frequency get measured and when equal amplitude the cavity is assumed centred

The discrete digital nature of the DDS signals and the fact that the final detection of the 2 ACT signals was done in software meant that there would always be errors in cavity tuning and ACT could not be independent of the hydrogen frequency which is "pulled" by the cavity tuning — The problem was solved using mathematics and the fact that the complete space clock also had a cesium part (PHARAO) for use as a reference

I proposed a simple analogue solution using homodyne detectors for the side frequencies that greatly reduced the complexity of the ACT circuit and in addition separated the ACT signal path from the hydrogen frequency signal path which could then be simplified using limiting amplifiers (as opposed to flat response wide band linear) further reducing the design complexity and improving noise performance which for an atomic clock equates to improved accuracy

The sapphire cavity of the SHM was buried deep in several layers of heat insulation with several heaters in between to maintain its temperature accurately — The temperature stable cavity was then "tuned" by an electrical signal from the ACT circuit which controlled a varactor diode connected to a wire loop inside the cavity

The time constant for the ACT signal needs to be very long and this was achieved using traditional op-amp integrators which combined with the other filtering required to keep the varactor control signal "quiet" meant that the ACT was affected by temperature changes in the varactor signal electronics

I suggested that the temperature control circuit should be used to vary the cavity tuning thus simplifying the ACT and giving it a natural long well damped time constant — The cavity could also be used to determine its own operating temperature which would be more accurate — This principle could also be applied to the ovens of crystal oscillators so no external noise gets in

Due to the many changes of companies undertaking this project I doubt my design will now become a flight model but I would like to thank Phil Smith at SemeLab especially for the open top transistors for proving the thermal analysis and Sam Samarasinghe at Express Circuits for help with the copper filled via PCB heatsink

I suggested converting all the documents including calculations and PCB layouts into pdf format and that then led to hyperlinking them together to form a single document for each section of the system — GPT had to agree that pdf was an "electronic format"

In 1999 rather than move for another job I contracted to Vector Technology to help them develop and improve their range of laser communication products and introduced closed laser loop modulation and a range of complimentary optical detectors capable of detecting acurately modulated signals in the presence of interfering ambient light

Click here for LiD3010 data