Right. After climbing on the back of the car and rocking side-to-side,
the front windscreen portion was found to have so much lateral movement
that a serious rethink was needed. The first step was therefore to brace
the top part of the windscreen. This actually has a benefit as well, in
that although it means ducking a bit to get through the door, the side
glass can be flat.
The second phase of stiffening up the roof is to brace the lower part
of the windscreen frame against the front door pillar (marked in green).
As it stands however the bracing is insufficient: the corner where all
three pieces (green) join together is free to move up and down.
Additional bracing (purple) which will need to continue through to the
other side is needed: that corner then will not be going anywhere,
and neither will the windscreen.
Showing the windscreen (red) which has (had) considerable lateral freedom.
What happens now is that actually the bracing has transferred that lateral
freedom to the uprights just behind the wheels, so it is doing some good.
There will need to be additional bracing (marked purple, left) and also
lower down as well: there is also vertical shear occurring that needs to
be stopped.
In an ordinary car, where the windscreen pillars were at the far outer
extreme edge of the vehicle, this type of windscreen bracing design would not
even be possible: there simply isn't anything to brace against. However
in this design, given that there is a 10in air-box on either side of the
car, there are *double* pillars to brace against (lower green bits).
This is why it's possible to get away with windscreen pillars made
from such narrow tubing, thus apart from anything increasing forward
visibility.
The door, below (green parts), amazingly, still fits and there is still
plenty of leg-room. However, as the stock of 40x20mm box section is running
low, and the weight is going up, the door is going to have to be remade
*again*, but this time mostly out of 15x15mm box section, with only the pieces
for the glass made out of 40x20mm. The primary reason for using 40x20mm for
the door's window frame is not because of the window itself but is to give
some bracing against the roof. Also, for when the doors are opened: 15x15mm
box section is not self-supporting at a length of 40cm (height of the
window) whereas 40x20mm is.
Ok, so the support diagonal is in (green, right) which stiffened up the
windscreen nicely, even though the car's entire opposite (right-hand)
side bracing support hasn't even been done, yet, at all. The only problem
was: the lateral movement was simply transferred to the uprights. So,
a couple of diagonal braces (green centre and right) were needed. Also
the front wishbone supports, cut out from the Suzuki, really had to be
braced properly as well (green bottom, partially obscured).
The new welder (inverter) from SIP is so much more superior to their Arc
welder. Instead of operating at 50hz (mains AC frequency) it operates
at what is probably close to 15kHz. I turned up the amps somewhat and
redid the welds along the wishbone supports in one straight 6in run,
and it came out clean, immediately. This is quite impressive, because
one piece was completely rusted; the other still had paint on it, from
where it had been cut out from the Suzuki, and the previous welds were
heavily blobby. All this made absolutely no difference to the Inverter!
However, the only reason that the amps could be turned up is because
the metal being welded was at least 2.5mm, possibly 3mm thickness.
The 40x20mm box section is only 2mm, and the rods go straight through
it at 100 amp. oops.
Showing the front diagonal bracing a bit more clearly, although this is hard
to do. Standing on the back, now, there's virtually no movement of the
entire left side of the vehicle, all the way from front to back. This is
a superb result, because there aren't even any doors in.
Once the right hand side is done as well I will double-check and then take
the car out for some photos, outside.