Abingdon CAR-nival 2014

The Abingdon CAR-nival is a pretty hectic day for sprinting.  Scrutineering, signing on and the drivers' briefing are half a mile from the paddock and there's a convoy run, two practice runs and two timed runs before lunch.  After lunch, there's another convoy run, two practice runs and two timed runs on a different course.  The day started dry, but it soon turned to rain, which became so bad that they halted the event for 45 minutes as the first corner was barely visible and water was flowing through the paddock.  My boots have since fallen apart!  A tentative morning after my Crystal Palace crash in the wet left me chasing the class lead, half a second behind, so as it dried out in the afternoon (bright sunshine and bone dry) it was all to play for.  A good run put me comfortably in the lead and I was lucky enough to receive my trophy from John Watson!

John Watson

To understand how I went at Abingdon this year, let's compare with last year.  Initially, I made a better start in 2013, but the 2014 gearing was shorter and I was travelling faster within 70 metres of the start.  Last year's car ran the low rear wing, whilst I ran with maximum rear wing in 2014 with the wing in its high up position and the flap at maximum angle.  The rate of acceleration slowed and 200m into the course at less than 90mph, the 2013 car was going quicker.  The 2014 car had pulled just half a tenth in this period of better acceleration.  By the time I hopped on the brakes in 2013, I'd gained quarter of a second in just 350 metres.

Abingdon Data 0-500m

This is where the time was gained.  I was far braver (60 metres braver) on the brakes into the chicane.  This gave me an overspeed of some 15mph during the braking zone and netting 1.13 seconds before the apex.  A slower exit from a poor line through the right-hand of the chicane led to a speed difference onto the straight, even though much of the speed difference was pulled back by the straight.  I presume a wider exit cost time on the final left in 2013 and that allied to the greater downforce and shorter ratios in 2014 closed the gap.  The gap in speed carried on all down the straight, costing 0.33 seconds in 2014.

Abingdon Data 500-1000m

I was again far faster in 2014 on the way into the corner.  On this run, I dragged the brakes slightly, scrubbing off 27mph by the apex, but being 9mph faster than 2013 at best.  Apex speed was almost identical, but I did have tremendous understeer in 2014 suggesting that the high downforce setting was bringing the centre of pressure too far rearwards.  I'll try a lower wing angle at a later date, which will reduce leverage on the front, hopefully giving more overall grip and lower drag.  Another run showed a 4mph advantage in apex speed, suggesting I was also off the boil on that run.  The later braking gave a tenth.

Abingdon Data 1000-1500m

Helpfully, my data logger made this year's best run a few tenths quicker than last year's.  Given that the official timing made last year's a 44.85 and this year's a 44.59, I think we can say that's within tolerance.  Sadly, I don't have data from my best run last year, which was a second quicker than the reference lap shown here, but there's enough information here to say that the high downforce configuration is considerably slower as a much better drive resulted in a time just a few tenths quicker.  Next year, I shall revert to a lower downforce configuration.  It is worth noting that despite a further year of (ab)use on the same tyres, that apex speeds seem remarkably consistent.

Motorsport at the Palace 2014

You may have twigged that I have skipped over the Llandow Sprint.  That's mainly because I have no data as the battery was deader than I thought and I'd not replaced it at this point.  The day ended with a 2nd in class pot and a snapped chain after just 8 starts - the company that did my spanner check over the winter and fitted the new chain didn't understand the importance of a good one and fitted something cheap!  I bought two DID X-ring chains and cut them to make 3 chains - the benefit of a short chain.  I also had the car corner-weighted and a full geo setup at Bridgwater College, who did a great job of telling me the car was heavier than I thought.  The bad news was 475kg wet with some fuel in and 43.7% front with me in it.

I do enjoy Crystal Palace.  The atmosphere is so relaxed and everyone seems to have fun.  I again entered for both days and found myself up against Rob Hume in his Van Diemen and David Seaton in his BMW-powered Pilbeam.  Sunday was beautiful weather, as we've come to expect, but I couldn't get my head in gear.  I took the class win by a narrow margin from David, despite having the floor peeled off the car on the way to the start line by the matting used to ensure the grass didn't get churned into mud.  Several marshals (including a fabricator of 30 years) pushed me out of the way and repaired the car over their lunch break.  To all of you (and the organising team for giving me a practice run after lunch), I am very grateful.  Scott Boulton took a bunch of brilliant photos.

So where did I lose the time?  Buzzing the limiter into the first corner cost me 4mph from running too short ratios and my apex speed at the first corner was down by 3mph.  Acceleration through turn 2 was much better and I took the corner flat for the first time, getting that lost time back.  Buzzing the limiter again cost me some time down to Big Tree Hairpin and a slightly slower apex speed cost me time through such a long, slow corner.  Again, peak speed down the straight was down, partly because of a missed shift, and I was very slow on the first apex of the chicane.  Another attack of the limiter cost me time down to North Tower and I was a whole 5mph down before braking into the corner.  Lots of time was lost being tentative through North Tower and that, on the whole, explains why I was so slow this year.

Crystal Palace Speed Trace

On the Monday, the heavens opened and flooded the paddock.  The cars were lined up on the return road and the track shortened to compensate.  The others in Sports Libre mainly gave up, but I soldiered on, despite my wets that had been purchased from a shelf where they'd sat since being taken from a Formula Renault in 2004.  Rob Capper was the only fast single seater and he clipped the bank at Big Tree Hairpin in practice, so I sniffed a chance of my first FTD.  In the first timed run, the heavens opened just after the roadgoing classes had gone out on a damp track.  A Subaru Impreza and Nissan GTR went quicker than me by some margin, but my run was scrappy at best.  I could go quicker.  On the second run, I was sat in the car as it was drying out and I was debating slicks.  Then a Datsun took out a tree and the 20 minutes lost left time for the rain to come down.  I pushed too hard, locked the front up and slid into the barriers at Big Tree Hairpin, smashing the front clamshell, the splitter and all the body mounts.  At least I won the class!  Scott Boulton managed to get a photo as I locked up.

Front Clamshell Damage

Winter Update 2013: Part 2

I can't go six months without an update to the site, so as tomorrow marks 6 months, it's now or never!  The RaceCapture/Pro has been upgraded to the latest firmware and I'm on the beta test list for v2.0, which is imminent.  I've yet to fit any extra sensors, but I'm hoping to get some students from Bridgwater College to fit them.  Adding to the front aero from the previous update is some wheelarch vents.  They reduce pressure in the wheel arch, reducing lift.

Wheelarch Vents

To balance out all this aero at the front, I had a mould made for the rear diffuser, although the diffuser wasn't ready for the first round at Clay Pigeon.  It wouldn't have helped there, so it made no odds!  Clay Pigeon was a bit of a struggle as the charging system wasn't working as I discovered after flattening my main pack and the spare.  I finally got the car running (and I think I broke the Sports Libre record too), but the two packs were dead.  I ordered a new lithium pack to replace them, which has an integral button for showing charge status.  It's a bit short on grunt, so this one doesn't start the engine at all.  Anyway, back to the diffuser, I finally had it fitted for Llandow, which let me remove some of the stiffening braces for the rear wing to reduce weight.  Here's the new mould on the left next to the old floor on the right.  Chunky!

Diffuser Mould

Winter Update 2013: Part 1

After the previous blogs on aero modifications, it makes sense that the car should see some aero mods for 2014.  That's not all it's had though as there were still two weak points to the car; steering lock and the chain tensioner.  I started at the front and the first change was to move the timing strut from the middle of the car back to the side of the splitter.  Not for aero reasons, but because it kept getting in the way when removing the front clam!  A bit of fibreglass and some matt black paint sorted that nicely.


Timing Strut: Black and Vertical

Moving backwards, a vertical gurney has sprouted on the leading edge of each wheel arch.  This causes low pressure behind it, which helps draw the high pressure air out from ahead of the front wheels.  The net result is reduced lift and drag.  The downforce at the front had started tearing the splitter floor off the front crash structure, pulling the rivets out.  They've been replaced with high tensile bolts and skid plates have been added under the splitter to try and reduce the wear on the leading edge.  Added to the trailing edge of the splitter is a small diffuser to try and reduce the pressure under the splitter further.

The rest of the changes were carried out to a much higher standard by Jo at Vulcan Dezign.  First off, the steering arms were changed to gain clearance to the front wheels.  Then the lower wishbones were replaced by newly fabricated units with a little more wheel clearance.  This means the rack now goes from lock to lock without fouling.  A new column with a new boss for the steering wheel gets rid of the minor slop from the wear that was in the old boss.

Lower Wishbones

The real work went into fixing the chain tensioning mechanism.  Failures of the previous mechanism had cost me money, grief and results, so Jo thought long and hard about how to resolve it.  The answer is a rotating rear transaxle.  The Quaife unit is able to be mounted in several orientations, which means it can be rotated slightly to tension the chain, removing the need for the idler gear and the tray it sits on.  Hopefully, I'll gain reliability, some power (fewer losses without an idler gear) and the ability to change ratios with a greater latitude in tensioning.  I've done 3 seasons geared for >130mph and if I ever broke 120mph, I didn't have the data logger running to prove it.

Tensioning Adjustment

Bolt on right tensions chain

The rear sprocket was worn beyond tolerances, so a new one's been ordered from Elite Racing Transmissions.  The new one will arrive in the new year and has 25 teeth instead of the previous one's 20 teeth.  This means that my 17 tooth front sprocket is the appropriate gearing for Castle Combe, rather than the 14 I have previously run (13 was optimum).  The 14 tooth is correct for Crystal Palace as I ran this year, but instead of using 1st and 2nd gear, I'll now be close to the limiter in 3rd.  This focuses more on gear changes, which means a flat shifting system is high on the list of priorities.

Whilst all that's been done, I still have to build a rear diffuser extension to balance out all the downforce that's been generated at the front.  The centre section was folded from aluminium whilst at Vulcan Dezign, but I need to mount it and build the outer sections to match.  I also need to find some arch vents to further reduce the pressure in the wheel arches.  The car also needs to go back together with a few extra sensors for the RaceCapture Pro data logger; ride height sensors (to validate the aero mods), RPM (so I don't have to correlate with the RacePak G2X logger) and gearbox speed (for analysing clutch slip and wheelspin).

Aerodynamics on Demand Part 2

After the first round of CFD analysis, I summed up that the car would benefit from reducing pressure in the wheel arch, bringing in the CoP forwards.  This would enable the use of a small front diffuser on the trailing edge of the front splitter to draw air into the newly formed low pressure region.  In addition, some vortex generation at the side of the front clam would reduce pressure in the gap between the tyre and bodywork, helping to reduce pressure ahead of the front wheels.  To bring the balance rearwards, a larger diffuser would reduce the pressure on the whole underside, especially at the rear.


So with that information to hand, I modelled a few changes (Google Chrome recommended).  In addition to the changes above, I also raised the front of the splitter so that it was angled closer to the direction of airflow towards it in an effort to reduce the separation caused by modelling with square edges!  Looking at the pressure map, you can see that I made a small error in the right-front wheel arch, which resulted in a high pressure area all around.  On the left side, you can see that there is low pressure on the front diffuser and that the pressure on the leading edge of the tyre is lower than that of the other side.  In hindsight, I would have modelled just one side of the car with the changes to more easily see the differences.


There are some meshing problems on the leading edge of the splitter, hence the high/low pressure variation.  You can also see high pressure build-up on the air exit on the sidepods, mostly down to the crude implementation of the changes.  At the rear, there's a much larger low pressure area thanks to the extra floor area of the larger diffuser and the low pressure it generates, specifically at the point of angle change.


Skipping the streamlines and slices, let's move onto the drag and downforce.  Drag is up 20N, whilst lift is also down 46N.  It's clearly at the rear too as the CoP is 65 metres ahead of the car showing that there's lift at the front and downforce at the rear.  The downforce/drag distribution graphs show this well.  The splitter produces less downforce now, although this may be down to the meshing issues exposed in the pressure map.  The figures at the front of the car are also skewed by the modelling mistake I made in blocking off the rear of the wheel arch.  The main point of downforce generation has been at the diffuser.


Most interestingly, I've also got the full results as a download and can process this in the tool of my choice, which happens to be the free Paraview.  So here's a little view of the front diffuser in action.  You can see the vortex flowing into the sidepod and the air washes down the side of the car and under the rear wing.  The high speed air (red) shows why it's low pressure under the splitter.

Front Diffuser

Additional information