Projects...

The land rover came with a standard induction system consisting of twin 1.75" SU carburettors fitted to a 3.5 litre 9.35:1 compression ratio Rover V8 using electronic ignition. The vehicle employed a stainless steel exhaust system coupled with log style exhaust manifolds. However, even after rolling road jetting of the carburettors and with good and tight linkages for both throttle and choke, they were the usual SU horror to use. The mixture was generally fairly good on cruise and part throttle but cold starting was erratic to say the least, and carburettor icing occurred regularly unless the weather was warm. It was quite typical to experience boggy stumbling running when having the audacity to approach traffic lights if the engine was anything other than fully warm and even then, the engine idle speed was impossible to set with any degree of precision.

Many will disagree and some off roaders have very good reason to prefer carburetion over fuel injection, but speaking personally, while I admire the elegance of the constant vacuum carburettor, I have had first hand experience of both Stromberg and SU carburettors on three vehicles in my life and have equally loathed them all. This paper describes the conversion of the existing carburetion system to a fuel injection system in full and is designed primarily as a maintenance aid.

The engine control unit (ECU) of the 14CUX system is run by a Motorola MC6803U4 8-bit microprocessor, which is an uncommon variant of the otherwise ubiquitous 6803. The processor is mislabelled in the ECU, possibly to deter any reverse engineering efforts. The 14CUX determines fueling values for each bank of the V8 separately, which requires the use of two dedicated timer outputs for independent control of the fuel injectors. This requirement helped to drive the selection of the MC6803U4 part, which has three timer outputs available (in contrast to the single timer output on the standard 6803.)

Minor design updates were made to the ECU through its life, with the earlier units given the part-number prefix "PRC" and the later units "AMR". The firmware and data used by the microprocessor is stored in either a 27C128 or 27C256 EPROM (depending on the revision of the ECU), which is soldered in place on most PRC units and socketed in some late PRC units as well as AMR units. Only half of the 32KB PROM space is used, so the code/data image appears twice, duplicated in the upper half.

The manual prepared for the maintenance of 14CUX system is available here...

Before I had completed the task of fitting the Bosch fuel injection system to the engine, the problem I faced was dealing with a no start condition. How would I know where the problem was? How could I diagnose the issue with little or no telemetry? The design of a fuel injection monitor system (FIMS) was a logical extension of the project and provided a great opportunity to explore PIC chips for the task.

This first version was cumbersome and also rather simple... only providing a real time display of the outputs from the two O² sensors fitted to the left and right exhaust pipes. But, out of all the signals you could monitor in order to figure out the health of an engine fuel injection system, these are by far the most useful to watch.

This 2nd generation telemetry monitor was designed to be fitted immediately to the left of the dash inside the vehicle to dynamically display the state of the lambda signals from both O² signals on the left and right exhaust pipes, while also displaying a number of other critical data values read directly from the 14CUX serial port (throttle position, mass air flow sensor, engine speed, temperator sensors (coolant and fuel) and the vehicle battery voltage.

The unit also post processes a square wave signal that should come to the ECU direct from a road speed sensor (a chopper transducer) but which due to the gearing on this vehicle runs too fast and so confuses the ECU as to the point where very slow speed cruise converts to a halt condition. The unit takes in the raw road speed transducer signal, divides it down and then outputs it back to the fuel injection ECU.

The unit is small and unobtrusive and uses an 18F4620 PIC chip programmed in C coupled with an I²C colour LCD touch screen for the UI.

If you're thinking of removing the heads of a Rover V8 for re-engineering - feel free to review these work notes to see how I approached the problem. Modern heads have to work with poor quality fuel and so valve recession is an issue that will affect performance. Reworking the heads periodically to cope with recessed seats and damaged valves is well worth while.

The idea of a series III heater is a rather good example of an oxymoron. Yes, it did work... sort of. Ironically it was rather good at heating the vehicle cab in the summer and I have to say, even in the depth of winter it was just able to demist the front windscreens. However, in regards to winter driving creature comfort, it was profoundly inadaquate. Partially that was down to the open and rather leaky design of the bulkhead on a Land Rover Series III but it was made worse by the design of the original heater assembly itself.

Newer designs employing the same size heater matrix and a similar fan motor manage to obtain far better results mainly due to the way the air is routed though the unit.

These notes descibe the steps necessary to fit a Defender heater box into an old Series III land rover and if you're anything like me - you'll thank the heavens each and every winter you use it.

I'd experienced problems with the cooling system of the Land Rover in the winter of 2012 - where a relatively small amount of antifreeze would leak from somewhere and end up pooling around the base of the distributor (located at the front of the engine). A lack of any obvious cause was very puzzling - for the hoses in the immediate area were all new and tight and the nearby thermostat housing was also well sealed.

After up-rating the heater system for the vehicle (see other project notes above) I decided to look more closely at this issue and found out that the old standard expansion tank was actually venting coolant through its expansion pipe over the front of the engine. The question was why?

These notes discuss the problem - and describe the eventual solution employed, and which is still running very successfully today and able to cope with extremes of temperature from cold winters through to very hot summers.

Way back in 2000, an aggregate company asked a colleague of mine if it was possible to write software that could simplify the task of laying paving slabs of different sizes to make a random patio. Later we chatted about the problem (he likes maths as much as I do) and after a while we came up with ideas to test some of the principles involved. It sounds quite easy to do, but combinations of stones can lead to areas that can't be filled, or will be likely to result in tram lines (where the paving slabs form long unsightly lines running in either X or Y across the whole plot.

It turns out this this problem belongs to a class of math problems known as non polynomial indeterminates (NPN Hard) characterised as having no right answer, just grades of answers ranging from really bad to really good. The trick was to code something that could get very close to a good answer but then provide additional editing tools to allow further improvement.

The software was coded in VB6, distributed for free and each year was revamped at the companies request so that new ranges of pavers (with different dimensions) could be added. In the link below a company is distributing it within their body copy. From time to time I get complimentary emails from landscape gardeners about this software, which is always a bit of a treat for me.

Scientists and Engineers have been waiting for a successor to the magnificent Hewlett Packard HP 42S calculator for years. In the interim, a bunch of smart chaps got together and repurposed a different HP calculator to create the interesting WP34S, a reverse polish notation calculator based on the HP 30b (financial) calculator.

Version 3 of the original WP 34S manual written by Paul and Walter is available below - but if you struggle reading that manual (like I did) then by all means have a look at my beginners guide. Originally this started out as a personal study aid as I explored the WP 34S and grew from hand written notes into a more formal user guide as I delved deeper into the machine. Its an open source device, so it does feel right and proper that community ideals should be respected and so I decided to make the guide available freely across the web.

Not all sections are completed, and some haven't yet been drafted (for example there is no matrix section) so keep an eye on my site - and feel free to share links. The document does contain a draft number on the footer of every page, and a summary change history table tucked in at the start to make clear what version changes have been made.

The guide was useful for me, I'm hopeful it may be for you too.

Block 11

Block 12