All other things being equal, a car body that is stiff in bending and in torsion will handle better (the suspension movement will be in the springs and not in the body), will be more durable and also has the potential to do better in a crash.
That’s all well and good – but one of the difficulties for amateur builders of cars is to turn the above type of generic statement into something that actually works on the road. In other words, if you’re not an engineer who also has access to complex computer modelling software, how do you devise automotive body structures that are strong and light?
Most recent literature gives you very little clue. Because of the high-end software tools that are used, and the associated high-end engineering skills needed to understand those tools, there is very little around that you can look at and then go: “Aha! I can see how that works to make the body stiffer without adding more weight!”
The Ancient Austin
Recently I came across an engineering study approaching 50 years old. It was published in Automobile Engineer and covers the development of the Austin 1800, a sedan produced in the 1960s, engineered in the UK, and built in the UK and Australia.
The body design of the Austin was significant for a number of reasons:
The result was a lightweight, extraordinarily stiff body that used only conventional pressed and welded steel panels.
So, how light? The complete, assembled but unpainted body had a mass of just 347kg – extraordinarily light for the size and carrying capacity.
In terms of torsional stiffness, the quoted figure was 13,300 lb-ft/degree – about three times that of contemporary cars and amazingly stiff even in today’s terms.
Finally, when bending testing was carried out by supporting the body on the axle lines and placing a mass of 454kg mid-way between them, the maximum deflection was just under 0.5mm!
I have owned an Austin 1800 sedan and while the car is certainly not beyond criticism, it still impresses with the ways the doors shut and the stiffness of the body on the road. I can well believe the figures quoted above.
So how did it achieve this?
The following material is based on the article appearing in Automobile Engineer, February 1965.
The Austin 1800 used a largely flat floor characterised by three design features:
The floor was made in two sections, spot-welded together. The front pressing (at left in this diagram) was made from 0.9mm sheet. The rear pressing (right) was largely made from 1.2mm sheet, while the rear seat pan, a reinforcing pressing located beneath the seat pan, and a transverse channel section at the front of the boot were all made from 0.9mm sheet. The floor of the boot was formed from heavily swaged 0.9mm sheet, supported by a 50mm deep channel-pressing longitudinally placed beneath it.
This diagram is a little deceptive as the side-sills look relatively shallow...
…however this photo of a floorpan cut from a car better shows the amazing depth of the sills.
The side sills, shown here in cross-section, had a maximum depth of 150mm and were pressed from 0.9mm sheet. The vertical stiffener (arrowed) was made from 1.2mm sheet and contributed to beam (bending) stiffness. The sills extended from the firewall to the rear edge of the rear seat pan (as we’ll see in a moment, effectively connecting the body points bearing the main suspension loads).
Trailing arm rear suspension was used with compact rubber springing feeding longitudinal loads into the seat pan area.
The roof and pillars were of conventional construction – in fact, given their slimness, it’s likely in a modern context that the roof and its supports were quite weak.
So in terms of bending strength, the body gained immense rigidity from its massive 150mm-deep sills with the internally welded vertical stiffener, and the widely used floor ribbing and central hump.
The body gained its torsional stiffness through the use of a very strong and carefully engineered firewall and ‘front-of-boot’ panels.
The design of the front suspension meant that the rubber spring units were placed transversely, being housed within a cylinder (arrowed). The main firewall panel was pressed from 0.9mm sheet, while the transverse cylinder comprised inner and outer sleeves – the inner being made from steel 1.6mm thick and 132mm in diameter, and the outer 168mm in diameter and also 1.6mm thick. The tubes were joined by means of two pressings, all four parts being welded together. At each end of the tube two forged and machined flanges were welded in place; these carried the inner pivots of the suspension.
This diagram shows how the transverse tube was assembled to the firewall. The projections downward from the tube (pressed from 1.2mm sheet steel) carried other parts of the suspension: almost all front suspension loads were fed directly into this firewall that acted as a major load-bearing bulkhead. The arrowed bracket carried the rack for the rack and pinion steering
This cutaway pic shows the way the spring/damper was mounted within the tube, and how most of the suspension arms were hung from the same basic assembly (an additional strut ran forward to prevent fore-aft movement of the wheel).
A frame side member (arrowed) was bolted to the flange positioned at the end of the transverse spring tube and welded to the inner guard and firewall. It also connected to the triangulated downwards projections from the transverse tube.
This shows the view from within the engine bay; the openings in the inner guard gave passage to the air outflow from the radiator.
The dashboard steel pressings (arrowed) were integral, stressed members of the structure.
At the front of the boot substantial cross-bracing was provided by two triangular pressings...
…while at the very rear of the car a tall vertical panel added further torsional stiffness (and also gave a higher sill to lift luggage over!).
Looking at the design from nearly 50 years later it still strikes me as amazingly good.
By using the transverse tube to absorb the front spring loads, the stressing of the firewall panel on two wheel front bumps was very small (consider how the two spring forces react against one another); on single wheel front bumps the great rigidity of the firewall panel being acted on transversely (and in plane) was obviously easily sufficient. At the back, the rear spring loads again acted in the same plane as the resisting panel, this time acting fore-aft against the floor.
The strong and stiff firewall that accompanied the front suspension design provided body torsional resistance; the heavily swaged triangular panels across the back seat must have been very light and yet effective. (But why didn’t the designers go for full triangulation rather than triangulating just to the centreline of the car?) The rear vertical panel would have also helped in torsion.
The use of very deep sills (complete with the internal stiffener), the “transmission tunnel” and the ribbing of the floor would have all helped in bending stiffness.
The fact that even the dashboard support panel was a stressed part of the body structure shows how the designers wanted to make use of every piece of metal at their disposal. Another clue to the approach the designers took is how there is barely a square inch of the interior panels that do not have swaged stiffeners, holes for lightness (and then rolled flanges around holes), pressed indents and the like.
Crash safety when measured against modern criteria? Almost undoubtedly, very poor. But as a lightweight, stiff and strong body where the techniques that were used to achieve those aims are on public show, I think it’s incredible.