Thursday, 27 August 2020

Fitting a Marshall supercharger on an MG PA - Part 1

This project deviates from the more usual vintage motorcycle restorations and recreations that are the norm in my home workshop but I have included it in my Vintage OHC AJS blog for the simple reasons that I believe it may be of interest to anyone wishing to undertake a similar project and it is a convenient place to document the work.

I have owned a lovely little 1934 MG PA that I purchased from Barry Walker in 2016. The car had previously been lovingly restored around 25 years previously with workmanship of the highest quality. I have driven the car for the past 4 years and enjoyed every minute of it. However, in spite of having twin SU carburettors, a 4-speed gearbox and not being particularly heavy, the little 847cc engine needs to be revved to get the best performance out of the car. In particular, it can struggle on long hills ….of which we have plenty in West Sussex; I probably wouldn’t care if I lived in Norfolk.

Anyway, I decided to put a supercharger on the car and acquired the important parts from Barry Walker – the Marshall supercharger, a new SU carburettor, boost pressure gauge, oil feed regulator, front pulley and jockey tensioner, belts and some bracketry. All of these are nice shiny new parts ….see picture below which shows all the bits except the SU.

As I have got into the habit of taking plenty of pictures during my motorcycle projects and adding them to this blog I have repeated that process here and in some detail. This may or may not be of use to anyone also wishing to fit a side-mounted Marshall supercharger. Whether I have done this project in an optimum way I have no way of knowing  …this is just documenting my experience.

To avoid this blog turning into a rambling monologue (although that seems to have happened anyway….) I have divided the Fitting of the Supercharger – Part 1 - into 4 main parts, namely: Initial Mounting of the Supercharger; Fitting the Carburettor; Fitting the Front Pulley; Fitting the Front Support Plate and Final Assembly. Part 2 covers Initial Start-up and Performance of the Car.

Mounting of the Supercharger

The first step was to remove the bonnet, the tie rods between the radiator and the top of the bulkhead, the apron at the front of the car and to take the radiator out. It is not strictly necessary to take out all of these parts at this stage but they need to be removed later so it’s just easier to make space to work at the beginning.

At the “supercharger end” of things, there are essentially 4 major components; in order from the engine these are: the pancake manifold sandwiched between the engine intake and the supercharger; the supercharger itself; a small right-angle elbow between the supercharger and the carburettor and, finally, the carburettor.

Attaching the pancake manifold: The standard twin-carb manifold is held on with 5x 5/16” studs. 3 slightly longer new ones were made for the inner 3 and the outer studs slightly shortened to fit the new manifold. A new gasket (flexoid) was made at the same time.

A brief note on materials: for all structural studs/bolts/nuts etc I use high tensile steels or components. In particular, all studs that I make are either EN24T (for 3/8” diameter) or silver steel (not heat treated, ie “as received” condition for 5/16” diameter). Why different steels? I can easily find 3/8” and larger size sections in EN24T but I can’t find 5/16” in small quantities in this steel hence the use of silver steel. Silver steel in this condition is still twice as strong as EN1A and more than equivalent to EN8 ….quite adequate for this purpose.

A gasket was then made for the supercharger-to-pancake manifold (large hole with 4 studs)

When the pancake manifold was offered-up to the supercharger it was found that the 2 lower studs fouled the aluminium cam cover.


These studs were provided with a flat washer, a coil-spring type shakeproof washer and the nut and there was still plenty of exposed stud. To avoid contact with the cam cover 0.1” was removed from each stud and the plain washer and coil-spring shakeproof washer were exchanged for an externally-serrated shakeproof washer (the latter has a larger area).

The boost pressure gauge can be fitted at any stage. It requires a hole to be cut in the dash using a hole cutter (I used a 51mm diameter cutter) and connecting the copper tube to the supercharger manifold. It’s necessary to remove the steering wheel to cut the hole. I would have preferred to have positioned the gauge slightly further to the right but there is a switch beneath the dash that controls the radiator fan that it would have fouled, although this can’t be seen in the picture below.

The oil feed and drain pipes need to be fitted to the supercharger. The oil feed is taken off the cylinder head adjacent to the supercharger by removing a blanking bolt and connected via a pipe to the oil feed regulator on top of the supercharger.

And the oil drain is fed into the central drain tube by silver-soldering a 1/8” BSP fitting in the location shown in the next picture.

I used 3/16” OD copper tubing as ¼” seems unnecessarily large for the low flow rate of a drip every 5 seconds. There are a number of 1/4” to 1/8” BSP fittings required – the oil feed on the cylinder head and the 2x fittings on the supercharger. All fittings on the pipework are silver-soldered. There is a ¼” to ¼” BSP elbow on top of the regulator to minimise the vertical height. The length of thread on the oil drain fitting in the supercharger was shortened slightly to avoid it “screwing into” the gear inside – easily checked because it locks the supercharger. The height of the reglator and pipework above the supercharger needs to be minimised to avoid it fouling the bonnet.

There is a lug at the drive-end of the supercharger that needs to be supported and that is covered later in the section on the Front Support Plate.

Fitting the Carburettor

The next step was to fit the elbow and the carburettor. Both the carburettor and the elbow “as-delivered” are without any holes or threads for attaching the two together.

The first step was to set up the elbow on the milling machine and drill and tap the holes for the studs.

The next step is to strip the carburettor and drill the corresponding holes and then make appropriate 5/16” studs.

The throttle must now be moved over to the opposite side so that the throttle linkage is on the inside rather than the outside. There were various steps involved in this.
The first step was that the cutaway in the carburettor body for the throttle stop must be replicated on the opposite side

Unfortunately the throttle stop now comes up against the elbow and so the elbow must be machined to accept the knurled adjuster.

It was now found that the linkage fouled on the back of the supercharger (needless to say, the throttle spindle on this side has been trimmed to the correct length) and it was necessary to either put a spacer between the carburettor and the elbow or between the elbow and the supercharger to avoid this. I chose the latter and made a spacer, as below, with gaskets on both sides. I would have preferred to have used a piece of Tufnol but used a piece of stainless steel instead as I had that in stock.

The throttle would now rotate without fouling although the throttle linkage itself….

needed to be 2” longer to connect the existing throttle lever coming through the bulkhead and connected to the pedal to the lever on the carburettor. I decided that, rather than buying a new one, I would cut-and-shut the existing 3/16” rod by inserting a length of 5/16” rod, drilled and silver soldered both ends and then chemically blacked.

Having attached the carburettor to the supercharger and the throttle working properly it was now clear that the excess spindle length on the other side needed to be trimmed back. Putting a straight-edge along the bonnet line showed that not only the spindle but also the spring would foul the bonnet.

Although the throttle already had a return spring on the pedal-side, this is one area where a “belt-and-braces” solution is desirable; after all, the original twin-carb installation had a return spring (2 in fact, one for each carb), the “as-delivered” new carburettor had a return spring and this is something I didn’t want to dispense with.

There is not a lot of space in which to engineer a spring on this side and, as there is no space on the other side, it seemed that a reasonably elegant solution would be to use a similar lever arm as for the throttle linkage (part # AUE 180 from Burlen Fuel Systems) and to put a spring between there and the dashpot screw.

Apart from the choke, the carburettor installation is finished.

The solution for the choke is not immediately obvious for 2 main reasons: firstly, the original choke actuating lever on the carburettor is on the outside of the installation (ie next to the bonnet  …it has already been removed in these pictures), which is a long way from the choke lever that comes through the bulkhead and to which it must connect and, secondly, the “direction-of-pull” for the choke lever on the carburettor is orthogonal to the direction-of-pull in the car. The direction-of-pull of the choke from the driver’s perspective is along the axis of the vehicle whereas it is now required to be across the vehicle.

I therefore decided to make an entirely new cable-operated choke operating system.

Although it might be stating the obvious, using a cable allows a “pull” in only one direction and a spring must be used to apply force in the opposite direction. This is unlike the original lever-only choke operating system which can apply a force in both directions, ie for both opening and closing the choke.

The pictures below illustrate the carburettor-end of the cable operated system I devised. The simple bracket, which is attached to one (slightly extended) carburettor fixing stud, is prevented from rotating by abutting against the inner flange of the carburettor. The spring required to close the choke needs to be quite substantial (it is a from a motorcycle clutch) and the bracket needs to be similarly substantial. The bracket is made from 20mm x 6mm stainless steel.


A new clevis pin has been made (EN24T) with an additional central hole through which the cable is passed before attaching the nipple. An alternative solution would have been to use a trunnion between the cable and the bottom of the jet but this would have had the disadvantage of increasing the length of an otherwise compact arrangement. The disadvantage of this design is that the cable must be inserted through the clevis pin before soldering the nipple.

The outer cable has an inner PTFE lining and does not require lubrication.
A brief note on attaching nipples to cables: For many years I have used the time-honoured method of splaying the cable in the nipple and then soft-soldering (plumbers solder) the splayed cable to prevent it pulling through. This works without a problem on motorcycle throttle, air, ignition and clutch cables. However, there can be problems with the front brake cable on vintage motorcycles. As an example, one of my bikes is a 1923 500cc Model 18 Norton that weighs next-to-nothing and goes like stink – it’s good for 70+ mph but has a 5” front brake and a rear brake that is essentially a lump of rubber pushed into a dummy belt rim. This gives the bike stopping characteristics normally associated with a fully-laden oil tanker and I therefore try to avoid situations that require emergency braking. However, on those occasions when substantial braking is required, it is just a natural behaviour to keep yanking on the brake lever in the misguided belief that extra pull equates to more braking. It doesn’t. And the most likely outcome is that something will break and that is invariably the nipple that is pulled off the end of the cable in spite of being properly fixed in the first place.

For this reason, I have taken to silver-soldering the nipple onto these cables and the same has been done here because of the strength of the spring that is used/required to close the choke.

Great care has to be used in silver-soldering nipples to avoid heat damaging the cable and destroying the temper. I clamp the cable (sufficiently to hold it but loose enough not to squash it) in aluminium vice clamps (to conduct away heat better) with just the nipple and the required amount of cable within the hole visible. The cable is splayed although this is not as critical as when soft-soldering. Oxy-acetylene with a #1 nozzle and the smallest flame possible is then directed very carefully at the nipple (trying to avoid the exposed wire within the nipple hole), having previously fluxed the hole, and the silver solder is then applied.

If carried out correctly, the resulting bond is much stronger than a soft soldered nipple and I have yet to have one of these break or pull through.

The final part of the choke operating mechanism is at the other end of the cable. Here, I made a “top-hat” arrangement by TIG welding 5 pieces of stainless steel together, a cable adjuster and making a trunnion to take the cable to fit on the driver-side of the bulkhead and attached to the existing choke mechanism. It is important that the trunnion is made with a slot to take the cable otherwise it could not be assembled by passing the cable and the adjuster (minus the locking nut) through the bulkhead without drilling an unnecessarily large hole.. 

Here, I have drilled a second hole in the original choke operating lever closer to the centre of rotation to get more leverage as the leverage that would have been obtained from the choke lever attached to the carburettor is now absent with the straight-pull cable.

The Front Pulley

The first step was to expose the end of the crankshaft by removing the Front Engine Support/Radiator Support and the End Cover.

First, the engine needs to be supported under the sump with a block of wood and a jack to avoid it just dropping by pivoting about the rear engine mount.

The Front Engine Support is fixed to the chassis by 2 bolts and to the End Cover by 4 bolts. The picture below shows it after removal.

The next step is to remove the End Cover, shown below, by first removing the crankshaft end nut and the existing pulley.

The end of the crankshaft is then exposed.

There are 2 machining modifications that need to be undertaken on the end cover: the 4 fixing holes need to be countersunk to enable 4 new countersunk screws to be used (this is necessary because the pulley would otherwise foul the bolt heads) and the inner diameter of the end cover needs to be machined to take an oil seal. It turned out that the second of these modifications had already been done on my engine although the oil seal needed to be replaced with one with a slightly larger ID. The new seal was SKF CR35x45x7HMS5RG. If the machining for the oil seal had not already been done the end cover would need mounting in the 4-jaw chuck and carefully set up to machine the recess.

The next step was to fit the pulley and here a problem emerged. The new pulley is shown in the 2 pictures below.

The problem I encountered was the pulley would not go onto the end of the crankshaft because there is insufficient space between the shaft and the chassis cross member immediately in front. The pulley was simply too long to go over the nose of the crankshaft without fouling the cross member.

I dropped the front of the engine further and further in an attempt to fit the pulley but there was no way the pulley would go on. To lower the front of the engine as far as possible I ended up: disconnecting the track rod; removing the seats and the floor to allow access to the rear engine mounts to slacken off the pinch bolts/remove the locating pin to allow the engine to pivot about the rear mount,

and then removing the tachometer drive from the rear of the cam cover to avoid it fouling on the bulkhead, slackening off the bolts holding the propshaft tunnel sheet metalwork and removing the odometer drive ….but still there was no chance of the pulley going onto the end of the crankshaft. The only way I could see that this could be done with the engine in situ would be to remove the chassis cross member and I was not prepared to do this.

Plan B. I made a new 2-piece pulley with a hub that would fit onto the crankshaft and a pulley that would then bolt onto the hub.

The new hub is made from EN24T, the internal keyway is spark eroded and the existing pulley was machined to fit the new hub. They are held together with 4x 5/16” BSF high tensile cap screws.

And now it fits!

It is important to check that the back of the pulley does not foul the end cover and that the engine turns freely when everything is bolted up tight.

The belts can now be put in place and the front engine support reattached.

The Front Support Plate

The next sub-project is the support plate that is bolted to the front of the cylinder head with 2x 5/16” BSF studs/nuts. This plate supports the end of the supercharger and also carries the jockey wheel tensioner for the belts.

The parts to do this were provided, namely the main plate, an additional small plate with an offset bend that requires welding to the main plate, the jockey wheel and its support plate. These can be seen in the picture at the start of this blog.

However, I decided that instead of using the main plate and the plate with the offset bend that had been provided I would make a complete single-piece plate and (luckily) Barry Walker had a drawing for this that had been made by, I believe, a German enthusiast, Oliver Renzow, that had fitted a similar supercharger. The one-piece plate is straightforward to make from a 200mm x 200mm x 5mm piece of steel and is shown below before the tensioner was fitted.

A few points worth noting: the studs in the cylinder head are threaded 3/8” BSF in the head and 5/16” BSF externally and new, slightly longer, studs need to be made.

A stud supporting the supercharger to the plate needs to be made. The hole in the supercharger lug was of nominal 5/16” diameter but with about ¼” length of 3/8” BSF thread. Rather than extend the threaded portion (which would be a viable option), a shouldered stud with 3x threads was made that can be firmly attached to the supercharger.

And, together with a spacer and a high-tensile half-nut and serrated spring washer, the supercharger is fixed to the plate and with adequate clearance from the belt.

The holes for holding the jockey support bracket are 5/16” and 3/8” and it is convenient to tap these into the plate (as it makes for easier adjustment) with a locknut on the back.

A couple of points to note here: in the picture there are multiple washer acting as temporary spacers on the bracket holding studs substituting for the radiator support bracket (coming next) and the length of the bolt head closest to the inner belt was subsequently trimmed slightly to avoid possible interference with the belt. There are 2x 1/2" BSF nuts on both ends of the jockey shaft.

The last bracket that requires attention is the top radiator support bracket. Because the radiator support bracket has been pushed forward by 5mm (the thickness of the supercharger support plate) the radiator has also been pushed forward. But, because the radiator tends to pivot about the bottom fixing (probably constrained by the bottom hose) the displacement is magnified at the top. The consequence of this is that the distance from the scuttle to the top of the radiator is approximately ½” greater and the bonnet no longer fits properly.
The solution to this is to “cut and shut” the radiator support bracket and the only really fool proof way to do this is to first cut the bracket in two, tack weld the 2 parts back together in situ and then remove and complete the weld.

Incidentally, the “nick” that can be seen on the left of the upper part of the bracket in the picture is to avoid the belt and bracket interfering.

The picture below gives an indication of the amount of overlap required to get the radiator back to its original position.

The final part that needs to be remade is the radiator tie rod so that it goes around the oil feed regulator and the dashpot on the carburettor. This is a length of ¼” diameter steel, threaded at both ends for the trunnions and shaped as shown below.

Installation completed!


I decided to pour a stiff G&T and take the rest of the day off.

I would like to acknowledge the valuable advice given to me in a few phone chats with Oliver Richardson.

1 comment:

  1. Great work :) look forward to reading the next part ..