Talk:Reece Fish Carburettor

I have some practical experience of this carburettor, having been an engine builder/tuner all my adult life. One of the principal drawbacks of conventional carburettors is its reliance on 'progression' circuits to overcome unstable mixture control during transitional periods of operation; traditionally, with fixed-choke units the idle circuit is fed by air being bled via a fixed jet into a gallery that is in turn fed with fuel by the idle jet, and the resultant mixture is then admitted to the venturi via the slow-run or 'idle volume' control screw. Ideally, the relationship betwen the air bleed jet and the idle jet should feed the idle circuit at a level where the 'idle volume' control screw is between 60 - 70% of its total effective travel; any weaker and the engine will hesitate when throttle is applied - in some cases stall altogether - any richer and the mixture will get progressively richer still as the engine speed increases. Most units get over this 'flat-spotting' on progression by setting the primary curcuit as weak as possible, and fitting an accelerator pump to make good any shortfall on fuelling levels; however this is an "indefinite science" that is subject to engine tolerance and wear levels, and also introduces more moving parts that can, and frequently do suffer high rates of wear. The Reece-Fish unit overcomes most, if not all of these drawbacks by its basic design; it is composed of a single float chamber that supplies fuel, via a one-way valve, to an adjoining sub-chamber that is roughly triangular in shape, the upper corner of which is bored through the entire main casting to accept the throttle spindle, which is hollow. The driven end of this spindle had flats machined to engage the mechanical linkage for the throttle, but at the opposite end it was threaded internally to accept a brass jet, which also served as a retention bolt to secure a pendant, swinging 'sweep arm' that operated within the confines of the trianglar chamber. The jet size was chosen according to the relationship between the engine size and the size of the choke (venturi)and regulated the overall flow of fuel to the engine, but the really clever thinking involved the sweep arm and the way that the fuel was drawn into the engine via the hollow spindle that carried the throttle plate. At the base of the sweep arm is a single drilled port that coincides with a progressively-widening groove that is machined into the rear face of the main casting, and at the same radius; at the lowest end it measures less than 0.2mm, and at its widest it is over 2mm in diameter, mirroring the diameter of the port in the sweep arm. The port connects to an end-drilling that transits the length of the arm, and emerges behind the 'main jet' that secures the arm to the throttle spindle; from there fuel travels down the hollow spindle to exit ports that are drilled at roughly 90 degrees to the orientation of the throttle plate itself, facing upstream of the incoming air through the carburettor. It should be noted that the throttle plate could be adjusted for angle independently of the spindle by means of two clamp bolts, thus allowing a degree of leeway when adapting the unit for a different application, i.e. a smaller or larger engine. The discharge of fuel directly over the throttle plate had a significant effect on the quality of atomisation of fuel entering the engine and consequent gains in smoothness and efficiency; however as with other fixed-jet designs, there remained the problem of snap throttle openings causing hesitation and weak-mixture misfires and blowbacks. This is the really clever part of the design; because the sweep arm operates within a closed chamber, any movement of the sweep arm has the effect of compressing the fuel trapped in front of it, and effectively discharging neat fuel, at pressure, through the spindle and into the airflow. This then gives the engine the enrichment it needs on sudden throttle openings, and also (as I found out on my 3.3 Litre Ventora!) overcame the need for fitting a manual enrichment device; four pumps on the pedal allowed the engine to fire up, and after ten seconds at 2000 RPM it was happy to pull away without any problems.