This article was first published in AutoSpeed.
Last week we covered the development of a new rear
suspension arm, required to overcome a problem where the rear wheel would patter
on bumpy roads taken at speed. The new arm used much larger tubing diameter and
appeared to be much stiffer than the original. But was it any stiffer? And would
it fix the wheel patter?
Measuring Torsional Twist
The first step in proving whether or not the new
arm was stiffer was to measure on-road torsional deflection.
As covered last week, this measurement is made by
mounting a long vertical arm on the rear suspension. The arm is equipped with a
whiteboard marker that draws on a piece of white laminated chipboard mounted on
the carrier. Any twist in the rear suspension is shown by a horizontal line
being marked on the board; normal up/down movement of the suspension draws a
With the previous suspension, the width of the
scatter of points was about 15mm, implying a lateral tread movement of about
7.5mm. But with the new rear suspension arm in place, the width of the marked
line was... wait for it... worse! In fact, at 30mm it was twice as bad!
So what the hell is going on? The problem is that
the whiteboard on which the pen writes in anchored to the seat, rather than to
the main backbone. Any movement in the seat frame shows up as twist in the
suspension, even if the suspension isn’t twisting much at all. And with the roll
stiffness of the suspension softened since the last test was done, the greater
angle of lean put more twisting force on the seat, resulting in a greater seat
So while I thought this measurement approach would
show changes in the stiffness of the rear suspension arm, it didn’t. However,
physically grabbing the wheel and pulling back and forth clearly showed better
stiffness – so would it be felt in the riding?
The next test was to assess the differences that
could be felt when riding the machine. On dead smooth surfaces there was, as
expected, no discernible change. However, when passing over slow speed cornering
bumps, the rear felt far more settled. This is a really interesting change –
what, precisely, does “more settled” mean?
The change in handling feel was similar to that
experienced on my Falcon when the rear suspension trailing arms and bushes were
modified – see
Frank's Suspension, Part 3
Falcon modifications reduced rear roll steer – so I’d now suggest that the
original trike rear suspension was roll steering.
Rear roll steer is hard to describe – it’s not a
case of feeling the rear steering in a different direction (like it is when
driving a four wheel steer car) but more a case of the whole machine feeling
imprecise in steering and when tracking through corners. A cambered wheel tends
to steer (it’s trying to roll around a virtual cone) and so the twist in the
rear suspension, which caused camber variations, would have been causing some
And what about on the high speed, downhill and
bumpy corner? On my main test hill the pattering was reduced a lot – that’s the
good news. However, on other high speed sections with closely spaced bumps the
pattering could still occur – even in a straight line. And the
latter implies a real problem....
With the rear suspension patter now identified as
occurring in a straight line as well as around high speed corners, it was
beginning to look more and more like it was a damping problem. That’s despite
the patter frequency (say, 10Hz) being very much different from both the natural
frequency of the suspension and also of the tyre.
In a bid to either solve or identify the nature of
the problem, I built a friction damper. This comprised a section of V-belt and a
small alloy V-belt pulley. The belt was partly wrapped around the pulley and
held tensioned by a spring. One end of the belt was anchored to the rear
suspension arm and the other end to the main frame. The pulley was mounted on
the main frame and was prevented from turning.
On suspension bump, the belt easily slipped around
the pulley. But the angles were so arranged that on extension (ie droop or
rebound), the belt wedged itself in the V-groove of the pulley and so resisted
being pulled through it. This gave very soft bump damping and stiff rebound
damping. The ratio of bump to rebound damping, and the stiffness of the damping,
could be adjusted by altering the amount of belt wrap and the tightness of the
This elegant and light system worked fine in
damping suspension movement. But even before full testing of pattering was
undertaken, I decided that the reduction in ride quality that the V-belt
friction damper caused was too great a trade-off. Especially in vibration and
over small bumps, the ride quality was way inferior to having no damper.
I then tried a hydraulic damper. As covered
Building a Human-Powered Vehicle, Part 4, for my previous
suspension trike design I used ex-motorcycle steering dampers. However, this
approach wasn’t entirely successful.
Building a Human Powered Vehicle, Part 7 describes the problems I
had. The main one was modifying the steering damper to give asymmetric
bump/rebound damping. This involved fitting internal valving, something I found
very hard to successfully do. The lack of space to modify the internal piston,
and the way that variations in internal fluid flow speeds gave odd behaviour,
made it a frustrating experience.
However, if the same sort of steering damper could
be found that had external valving, modification would be a lot easier. And, as
it happens, the Yamaha R1 uses a steering damper that has this sort of external
valving. The damper is very much like the ones I used previously except cast
into the alloy housing is a passage that connects the two sides of the piston.
I bought one of the dampers (a very lucky and
cheap find on eBay) and modified the valving to give much stiffer rebound than
bump. Achieving this was relatively simple – just take out the standard screw-in
valve block, shorten it and then place a spring and a rubber seal inside to form
a one-way valve. Since in these designs some fluid is allowed to flow past the
main piston, the hydraulic system still allowed rebound extension, but with a
much stiffer force than on bump. By changing the oil viscosity, the total
stiffness could be changed.
Unlike my previous designs, I thought that this
damper was pretty well on the money. Very slow speed bump and rebound stiffness
were of the same rate, but the higher the shaft speed, the stiffer rebound was
in relation to bump.
I fitted the damper, only to find that the
standard oil was too thin: there was insufficient damping. I then replaced it
with higher viscosity oil and went testing. But, interestingly, the presence of
the damper again transmitted vibration and small bumps to the frame (this was
especially visible in shaking of the rear vision mirror), and even more
interestingly, at speed the rear wheel patter was still there!
Solving One Problem
It was time for a very careful re-think.
Altering the rear suspension member for one much
stiffer had improved things but not fixed them. (But it had given significant
benefit in other cornering conditions.)
Changing rear airbag and rear tyre pressures had
also made little/no difference to the pattering problem.
The pattering could occur in both cornering and
riding straight, but only at high speeds (ie downhill fast runs at, say, 50+
km/h) on roads where there were closely-spaced bumps.
And then I had a thought. A few days before I’d
had the back of the trike up on blocks and, for some reason, had been spinning
the back wheel very fast by turning the pedals in a tall gear. I’d then noticed
that the back wheel seemed a bit out of balance.
What if the rear wheel’s balance was bad
enough that at speed it was trying to bounce up and down – and road bumps simply
worsened this behaviour?
I raced down to the workshop and removed the big
rear wheel reflector attached to the spokes. I then spun up the rear wheel and
noticed that while it was still a tiny bit out of balance, it was vastly better
in this respect. And a test ride proved the theory: the high speed rear wheel
pattering was gone.
Yes, after all the effort, it was as simple as
removing the rear wheel reflector....
a single reflector wouldn’t cause much of a problem, would it? I had a physicist
do the maths for a 40 gram reflector and he found that at my top (downhill)
speed, the effective mass of the reflector grew to no less than 3.2kg....
Feeling pretty confident that all was now solved,
I trailered the trike to the Gold Coast where there are flat roads and cycle
paths. There, in one easy afternoon, I rode 75 kilometres. Sitting at a constant
18 – 20 km/h, I was able to feel things that I never feel at home, where it
seems I am always doing either doing 5 km/h up hills or 50 down the other side.
And what I felt was another version of the
rear-wheel patter, this time occurring over short sharp bumps like road
reflectors. Over this type of bump, when pedalling hard, the rear wheel could be
felt to monetarily lose traction. On close-spaced sharp corrugations, the rear
wheel skipped, and over big wave-like bumps the rear simply felt under-damped.
All of which added up to the absolute necessity
for a rear damper...
I went back to my modified R1 steering damper and
re-assessed how I’d previously fitted it. Looking again at the installation, I
realised that I’d pre-loaded the air spring – the full extension length of the
damper had been less than the full extension length of the spring. Was this
preload responsible for the worsened ride quality when the damper had been
fitted? I installed it again, this time with an extension plate that prevented
any preload being applied.
And this time the ride quality remained excellent
but the rear was clearly much better damped.
Removing the reflector and fitting the rear damper
has effectively solved the problem of rear wheel patter.
When going really hard (like cornering hard enough
to being close to be up on two wheels at 30 - 40 km/h on a bumpy surface), the
rear tyre can still skip sideways on the harshest bumps. But this new behaviour
doesn't worry me at all. On the road it's a progressive warning of an impending
loss of rear lateral traction, very much like a RWD car that power-oversteers
out of bumpy corners. It feels completely unlike the previous patter, which
always gave you the feeling of: “Oh shit – where’s the back going?” The new
sideways behaviour is much more: “Ah, starting to go hard enough to be on the
brink of oversteer....”
I should also make the point that all the
behaviour I have been describing is at speeds that, on these surfaces, a
non-suspension recumbent trike would find completely impossible – not only from
a tyre adhesion perspective, but also from a point of view of pounding the rider
so hard they’d find it hard to see and steer...
The front damping (caused by the change in track
that comes with suspension movement) and the rear damping (caused by the
hydraulic damper) are now also very much in synch. For example, passing over a
sharp depression (eg one caused where a filler strip has sunk after a pipe has
been placed across the road), the front and rear bump impact harshness are now
very similar. You also never feel the rear or the front bobbing in an undamped
way – the suspension is working all the time, but it’s not particularly
noticeable in its action.
What you notice is that the ride quality is
incredible – and there’s no damn rear wheel patter!
Next week: the conclusion – the final part in