So after slicing everything to bits I needed to start reassembling the subframe in a way which would hold the engine and all of the gubbins required – primarily that damn manifold and the new manner in which the engine mounted.

At this point I had no idea how the engine would mount, just that it would be significantly different to the way it mounted in the RX8, with different mounts and different bushes. The stock RX8 mounts extended further than the chassis rails of the MX5, which pretty much immediately ruled them out.

Through a process of offering the engine up, taking the engine out, cutting and repeating the process, I started to get more of an idea of what needed to be removed and how it was all going to fit together.

A major factor playing on my mind at this point was the angle of the differential was at. This was crucial to the position of the engine, as where it sat within the bulkhead affected the angle of the propshaft. Being a one piece propshaft meant that tolerances within the angles of the UJs was way less than a two piece. What I needed to do was to start measuring stuff.

So, with a variety of creative measuring techniques I began measuring everything on the car, and making a CAD drawing what I was trying to do. I started with a downloaded blueprint of the MX5 which I checked for accuracy. The door’s length, wheelbase, height and arch diameter measurements were millimetre perfect, so I assumed this was a good starting point.

Above. In scale millimeter(ish) accurate CAD drawing. 

I then measured the floor with a spirit level and made sure the axle stands were all equal, before creating exact scale replicas of the engine, subframe, gearbox and differential. This method gave me greater peace of mind than simply eyeing the thing up, and also saved the potential nightmare scenario of the prop running out of balance when it was finished. In short: I found out that the engine needed to sit at 3 degrees back. Sounds like a piece of piss but actually took me bloody hours.

The Weldathon Begins

As Christmas set in I used my last few hours of sobriety to see my mate John back in my native homeland of the north east. From his habit of only purchasing rusty old shitters, his welding game was significantly more developed than my own, with some ten years experience under his belt.

Above. John.

As a result, his welder was better than the one we had in our workshop, a British made Portamig 186 which was equally comfortable welding thin plate at 15a or blasting through 5mm steel.

A chassis brace from a Mk2 made an ideal brace for the back of the subframe once it had been chopped down.

Next up some old gym equipment was recycled, as it was made of decent 2mm plate rather than the usual monkey metal home gym equipment seems to be made of.


The usual mouldings through which the engine mounts had to be chopped back in order to let the new quarter-pipe sit, so it was necessary to somehow put it back. We chopped a long 10mm channel through the bottom of the subframe to allow for a kind of elongated plug weld to be laid down, tying the whole thing together. On the topside a couple of slugs tied the top of the quarter-pipe to the crossbeam, hopefully rivalling the rigidity it originally had.

Before it was even New Year I was back down to Sheffield and into the workshop, tying up loose ends and making sure that the revised subframe would be strong enough to hold the new engine. The front crossmember, which had previously provided a considerable amount of rigidity (in its current state the flex could be felt by pushing the two sides in opposite directions), needed some reinforcement. Overkill perhaps, but a meaty 1x2” section of 3mm wall steel box fitted nicely into the gaping hole left by what had been cutaway.

By jamming the front face of the subframe into the pillar drill I was able to make around 12-15 12mm holes in a zig-zag fashion right across the front, skipping out the steering rack mounts. After cleaning out all of the swarf and metal grounds with an airline and some brake cleaner I clamped the two pieces together and laid some big ugly plug welds in there. Though not the strongest type of weld on their own, several of them together can provide a tough, rigid product.

Above. About 15 plug welds smoothed over. C-spanner used to simulate a 6mm gap beween the engine and subframe, which is about all there is. 

Next it was simply tying up what was left of the top and bottom sections of what was left of that front cross member. It was clear that this method was adding weight to the overall piece, but between chopping out so much metal in the first place and using a lighter donor engine, I figured I was okay.

After this it was simply filling in some of the leftover mishapes that had been cut, bending some more 2mm steel into shape and grinding it til it fit.

On thinner plate work such as this, it was clear how outclassed our welder was compared to the Portamig. The Sealey Supermig 130 is a decent welder for the price, but with no gas flow regulator, achieving the same consistency is difficult. Even with the same gas settings, wire speed and power settings, it’s not unusual to get differences in the quality of the bead and ultimately the integrity of the weld.

As I was beginning to fabricate engine mounts and gearbox mounts, tension was piling up as to whether this particular welder was suitable for blasting a deep, penetrating slug into 4mm steel. Compared to the welds I was getting with the Portamig, the Sealey just wasn’t consistent enough.


Above. Decent slug done with the Portamig.

So now I have loads of pieces of metal cut into relatively intricate shapes, waiting for a decent semi-professional welder so I can join them together properly.

Above. Making the uprights for the engine mounts.

Above. PU bushes and pipe for the engine mounts. Awaiting a welder.

Above. Assassins Creed pendants. 

Luckily, I made the sensible decision to bankrupt myself (again) and get a new welder.

Previous Next >