Sprinting an ADR Sport 2

[andylaurence]

I thought I'd start a thread on the car. I've had it two years now and there's been a bit of progress, so here's the story so far...

I did just one event in 2010 at Colerne before parking it up for the winter. Over the winter before the 2011 season, the car went to H's Garage to be checked over for the new season. A new rear wing from ADR and front splitter from 9mm ply skinned with fibreglass weave reduced weight and (hopefully) increased downforce. A new lithium battery was procured to reduce the weight further (it weighed less than the bracket for the old lead battery and is held in with cable ties) and a Racepak G2X data logging device and new dashboard were installed to allow debugging of the car and driver. Also, the extinguisher was replaced with a slimmer unit so I can take passengers at track days.

At the first event in 2011, I picked up my first class win at the Great Western Sprint, the first round of the British Sprint Championship. No changes before the second round and I ended up being pipped to the post by Angus in his Radical on the last run of the day. Onward to Crystal Palace, and the competition wasn't so fierce in class, so I picked up two easy class wins and 4th/3rd overall. I was disappointed to spin twice on runs that were on target to win the event outright, according to the data from the logger. Inexperience! I went to Anglesey for the British Sprint double-header and lost out again to Angus on both days.

To celebrate 100 years of its existance, Bristol Motor Club held a centenary party with the Dick Mayo Sprint on Saturday, live music on Saturday night and a gathering of interesting cars on the Sunday. I bagged second in a very strong class entry and demonstrated the car on Sunday in the parade laps. Happy Birthday to Bristol Motor Club!

The next event was the Wessex Sprint. Unfortunately, the chain stretched in first practice and whipped the tensioner on the downshift into Tower. I didn't realise what had happened until I got back to the paddock, by which time the tensioner was torn to pieces. I had a new one machined from billet, but it was season over.

Over the winter, I looked into making a new diffuser. I posted on the forum for some advice and cribbed a design in Google Sketchup. Somewhat embarassingly, that's as far as it's got to this day, partly because of time, partly because of skills and partly because the car won't go on the trailer with a bigger diffuser.



There were some big changes over the winter:

• The chain tensioner was fixed so it will tension the chain properly and now straps in place to prevent a failure like the one at Castle Combe last year.
• The rear wing mount was milled out and weight has dropped from 7.75kg to 2.5kg!
• The rear diffuser mounts were lightened too.
• Wishbones intruding into the chassis were covered to protect my legs in the case of an accident or the mounting shearing.
• Differential oil seals replaced.
• Rosejoints on the rear pushrods replaced due to wear.
• Alloy plate under engine and a variety of brackets removed.
• Reverse gear selector permanently secured into forwards.
• Battery re-located.

The first round of the season was at Mallory Park. Despite breaking the class record, I was beaten by Clive Wooster. Luckily for me, he took BARC FTD, so I still went home with a pot.

I did a charity track day at Castle Combe soon after and on the second run, felt a bit of chain slap. I'd not noticed the chain had stretched last year and the work I'd had done on the tensioner was based around a stretched chain. I couldn't fit a new chain, so I went home. I managed to space out the tensioner enough to fit a 14T front sprocket (not the 15T I'd run before), but I can't fit a 13T or a 15T anymore. Gah! Simulations with Bosch LapSim had shown that a 13T is the fastest around Combe (and probably everywhere else) and that a 14T is slower than a 15T because I carry one gear higher everywhere.

I ordered some carbon end plates (450mm x 300mm) from LCM Motorsport to replace the piddly items on my wing. I didn't have time to fit them before the next event at Castle Combe. It was to be a testing time! Coming through Camp on my second timed run, the left rear wheel shattered and flew into the catch fencing. Luckily, it bounced back over the track and into the field, rather than into the packed pit lane. Fortunately, the first run was enough to win the class and finally beat Angus on his first visit to Combe. The car was undamaged other than a few mm lost from the diffuser strakes.

Getting new wheels seemed tricky with a 4" PCD and centre locking. In the end, ADR offered me a deal I couldn't refuse and sold me a set of OZ magnesium wheels and new hubs/discs/calipers to convert the car to Dallara PCD. Unfortunately, it took a while, so I missed Llandow and Clay Pigeon.

Next up was Castle Combe again for the Dick Mayo Sprint and I was hoping for dry weather as I'd not been able to find a set of Dallara wheels to fit my wets. Fortunately, it was dry and my only problem was beating Clive Wooster. He did it again, pipping me on the final run to win the class.

[andylaurence]

Mallory Park was the scene for the first event of the season and the first time behind the wheel for about 9 months. What does the data say about how I got on? Starting at the beginning of the data, I’ve plotted RPM for second practice and both timed runs against distance. You can see that I tried 3 different launch strategies in the three runs, but they all worked out the same by the time I’d covered the first 20 metres. What this also shows is that by the time I’m doing 32mph, acceleration is governed by power rather than grip.



The second thing of note is that there’s a spike in the revs at about 28 metres into the run on all three runs, as shown by the cross on the vertical marker. This means there’s a slip between the engine and the road, so it’s either wheelspin or clutch slip. That could be a bump in the road or a torque spike at that point in the rev range. The spike doesn’t happen at the same point in the other gears, so that suggests it’s not clutch slip. It does, however, happen at the same point in first gear after the spin in second practice, suggesting that the tyres are breaking traction as the engine comes on cam.

Moving up towards second gear, the next thing of note is the point at which I changed gear for the first time. Interestingly, it got progressively further up the rev range as the day progressed and the end result was marginal at best.

The entry into the first corner was tentative in my first run, so in the second practice run, I decided to enter it flat. I was pleased to see that it was flat in third gear, but less pleased to see the chicane appear as quickly as it did. You can see that in the second screenshot, where I’ve removed the first timed run. The blue lines show second practice (top) and normalised RPM (bottom), whilst the red lines show the same traces for the second timed run. The normalised RPM channels are simply RPM divided by the ratio for that gear. These traces are directly comparable from the first vertical dotted line to the solid vertical line.



In second practice, the revs continue to climb (top blue line) in 3rd gear all the way through the corner and speed continues to climb as a result. The next thing to note is that the red traces dip at 317 metres into the run, showing I’m on the brakes for the chicane. At this same point, the blue traces show I was going 5mph quicker in second practice and the dip in the blue lines showing braking start at 341 metres and going 7mph faster than the other trace showed 24 metres earlier, which is probably why I arrived at the chicane backwards. Clearly, I learned too much from that spin and was far too tentative in the afternoon as braking 24 metres later was clearly the problem, rather than the extra speed.

So, how smooth is my driving? Looking at the acceleration data, we can get an idea. The image shows the acceleration off the line at the bottom, moving about a bit and a small amount of yaw when the line moves to the right. A single line connects the acceleration zone to the cornering zone on the left. The line moves about quite a lot on the left side of the graph, suggesting noise, a lot of bumps or a lack of smoothness from the driver. Most likely, a combination of all three. Intriguingly, we know that speed didn’t really vary on this, the final run throughout the corner, so the acceleration G being consistently positive doesn’t stack up. There’s several possibilities; the sensor is mis-calibrated, the acceleration through the corner is making the rear end squat under acceleration and tilting the sensor or the car is oversteering slightly throughout the corner and the acceleration shown is actually a component of the lateral G.



Acceleration data can be viewed in other ways too. Look at this graph showing the longitudinal acceleration stacked on top of the lateral acceleration. The data is normalised so that acceleration is positive for all directions. You can instantly see that the highest points on the graph are when braking into a corner, turning into the first chicane (around sample 300) and also into the hairpin (around sample 950). You can also see that I completed my braking before turning into the second chicane (around sample 600) by the trough in the graph. This shows I braked too early and coasted into the corner, not utilising all of the grip available from the tyres. Tenths to be gained there and also at the similar point in the final chicane (around sample 1040), which I entered with just a lift of the throttle and again on the change of direction in the chicane.

[andylaurence]

Since the broken wheel at Castle Combe, there's been a few updates. I had the option of finding a new set of wheels or replacing the centres (all four to match). The 4" PCD of the ADR Sport 2 isn't such a popular size and the chances of finding a suitable replacement set in time for the next round of the ASWMC Sprint Championship were small. Replacement centres would have to be custom made with a lead time of more than a month. Adrian at ADR suggested a third option to upgrade all four corners of the car to ADR3 spec, which allows fitment of Dallara F3 wheels, which are more readily available. This gives improved braking, lighter hubs and lighter wheels. Most importantly, the parts were all in stock and I could be ready for Clay Pigeon.

I got to work pulling all four corners off the car and carting them down to ADR in Holyport. It wasn't until the rear hubs were separated from the uprights that we realised the ADR3 hubs wouldn't fit my uprights, which are bespoke like most other items rearwards of the driver in order to convert the car from the original live axle to a chain driven double-wishbone arrangement. By this time, the front was already upgraded and the best way forwards was to get a new bespoke pair of hubs sorted at the rear. Luckily, everything at ADR is designed in CAD and a few keystrokes later, a new design was sent out for machining.

Sadly, there wasn't enough time for the new hubs to be made before the next event and they arrived a couple of days after the event. In the meantime, I got to work on other matters. I took the front of the sidepods off to allow air to escape from the front wheelarch, relieving high pressure and reducing lift. As a plus point, it also removed 2kg of fibreglass.



Whilst in the garage, I spied my spare rear wing, which is a dual element design. Having recently purchased some new larger end plates, this seemed the obvious chance to upgrade. I used Foilsim 3 to estimate the performance of the wings, which suggested 152lbs of downforce at 10 degrees around Camp corner for the main plane. Now I need to establish just where the flap needs to be mounted and at what angle to achieve similar drag and, hopefully, better downforce. CFD Gurus, please apply here!



With all that analysis to be done, I took the car down to ADR so that the new uprights could be fitted. At the same time, I had all the rod ends and bearings in the suspension replaced as after 8 years, I've no idea of their history. The corner weights and the alignment had to be reset at the same time. The car now has ADR3 lightweight hubs, new discs, new calipers, new rod ends and bearings throughout, new magnesium OZ wheels and, because the new wheels are wider, a new set of tyres. If I don't see a step in performance, I'll be disappointed!

Dual Plane Wing Update

This weekend, I spent some time fitting a flap after the main plane of the rear wing and some new end-plates to facilitate this. I used the old end-plates as templates and drilled a couple of holes to fit the main plane. Once the end-plates were fitted, I used a flat-bladed screwdriver to remove the foam core around the edges. Some black epoxy filler sealed the gap and after 24 hours to cure, a quick once over with a sanding attachment in the Dremel made them good enough. If I was fussy, I'd make it a bit neater and use a polishing bit too, but I compete against the clock, not the flannel!



In the absence of any data to decide where to mount the flap in relation to the main plane, I took an educated guess with a 6mm slot. Marking in the right place, I clamped both end plates together with the holes for mounting the main plane to ensure the flap was mounted symmetrically and drilled a hole. After mounting the end-plates back on the main plane, I bolted in the flap. Using the angle brackets that previously mounted the flap, I relocated them to fit the slightly wider main plane and angled the flap to the same angle as the trailing edge of the main plane. Dremel to the rescue and I drilled the second locating hole. Wing fitted!



Admittedly, there's a few issues to be sorted out. The flap is narrower than the main plane, so there's a 12mm gap between the end-plate and the flap at each end. Also, the flap is flexing and I think it's safe to say it's flexible enough that it will close the slot gap between the two at speed. A couple of supports along its length should solve that minor problem.

Of course, I've no idea of its performance, so the next task is to fit the spare camera to the rear wing support and tuft-test the underside to ensure the flow remains attached in normal conditions. The data from the next event will show whether it's producing more downforce and more drag. I'm hoping for lots of the former and very little of the latter!

[andylaurence]

I've ordered myself a IOIO (pronounced yo-yo), which plugs into an Android phone to provide 48 inputs and outputs. £37 delivered is an absolute steal! It's just a bare board, so will need a waterproof box to sit in. I plan to use Tupperware! I've got a few sensors that I can hook up to it; wheel speed, fuel level, brake pressure, dry sump oil level, header tank water level, water temperature, oil pressure and oil temperature. I have also purchased a selection of potentiometers from Maplin to measure clutch pedal position, throttle position and steering angle.

I've also found WifiLapper to use with it. It's a great piece of software that uses the phone's internal accelerometer, a Bluetooth GPS unit and the IOIO for collecting data. It shows the current lap delta on the phone's screen (red for slow, green for fast), so you know how well you're doing at a glance. Best of all, if you have your laptop on the pit wall, it'll upload the data as you whizz down the start/finish straight. Of course, that's not much use in a sprint where most events are less than a lap.

The software isn't an analysis tool, so with my RacePak G2X collecting RPM and the phone collecting other data, I'll need to combine the two. I could do that in Excel, but I'd much rather use a better tool. I've found Dashware, which seems to do the job for about £30 ($50) and also synchronises with the video from the GoPro to create video with a nice data overlay.

[andylaurence]

Two events in, I arrived at Castle Combe keen to continue my form and take another class win. Sports Libre had been split into <1800cc, >1800cc and rally cars leaving four of us battling in the small engined cars with Colin Early on his own in the unlimited class and a solitary rally car. Colin took part in the battle for the category win, comparing times with us after each run.

The car had been significantly upgraded since the last event with four new wheels, new tyres, new hubs, new wing element and everything else you've probably read about on this blog. I was keen to improve on my 68 second run of 2011. I took it easy in first practice as you would an installation lap. I was surprised to get a 70.06 on that first lap and it showed promise for later on. Clive and Colin both put in a 68 and Tom Arnold was only half a second behind me. In second practice, I put a bit more effort in but was red flagged when Tom had a spin at Bobbies. In my re-run, I was surprised to put in a 66.26, which was 0.9 clear of Colin and 1.66 clear of Clive. I was confident I could go quicker in the afternoon and Clive was clearly worried.

In the first timed runs, I followed form and went slower than practice with a 67.67, whilst Clive went one hundredth quicker than I had done in practice and Colin trailed me by two tenths. Tom, meanwhile, had broken the 70 second barrier for the first time. It was looking like rain and the last run would decide the winner. Any of us could win it and I wanted it to be me. I set off and took Avon Rise faster than I've ever done before, made a good exit from Quarry, locked the rear going into The Esses (brake balance needs to go forwards), kept it pinned through Old Paddock and made a brave stab at Tower (still more bravery required) before fumbling the downshift into Bobbies and accelerating towards the finish in the wrong gear. It felt good. A 66.57 was the result and to add insult to injury, Clive followed me through the timing beam with a 65.54. Second in class. Colin was unable to improve, but Tom put in a 68.73 to lop about 4 seconds off his PB. He must now be considered a contender in class.

A quick skim of the data suggests that the car's not really much quicker than before, but I'm getting to grips with the circuit, braking later and carrying more speed everywhere around the apex. Most intriguingly, I've had a photo of the rear wing sent through by Scott Boulton, which clearly shows the wool tufts on the rear wing. Whilst the ones on the flap are showing attached flow, the main plane is showing separation. It's stalled, so I need to lessen the angle so it's working like a wing rather than a parachute as it must be at the moment!



Unfortunately, the car needs yet more work as the steering lock was inadequate at best. The wider wheels contact the steering arms at half the lock it had before, so I've had to get the mounts machined to get more clearance. Keeping a close eye on Ackerman, it's going at add 50% more steering angle, which should be enough for paddock manouvres and getting around Clay Pigeon Raceway on Sunday.

Keep an eye on the blog for a data debug including a comparison of data between the ADR and Tom's ever-improving Spire GTR. Here's hoping he doesn't go any quicker!

Data Debug

It's almost two weeks since the Dick Mayo Sprint and so it's time to have a look at the data and see what happened. This, however, is a special data debug as Tom Arnold has provided data from his fastest run (68.73) in the Kawasaki ZX-12 powered Spire GTR to compare with mine (66.57). The two times were set just a few minutes apart, so utterly comparable. Tom's Spire is somewhat heavier (at least 100kg) than the ADR, although Tom's a few kilos lighter than me! The engines should be of similar power and the frontal area won't be much different. Tom's tyres are a harder compound, so mechanical grip is likely to be less.



So starting at the left, the most obvious point here is that Tom appears to accelerate quickly off the line before slowing down around 40mph. The reason for this is that Tom's data comes from a rear wheel speed sensor, so this is actually wheelspin. My data is from GPS, so wheelspin is not shown. Our 64 foot times are actually within 0.01 seconds according to the official results from TSL. So this means that Tom's data is actually shifted to the right, which explains why his trace looks slightly off for the rest of the run.

Accelerating up Avon Rise, the cars are very similar. I'm first to lift off going over the rise with a small confidence lift. Tom appears to be on the brakes soon after, whilst I'm on the throttle again. At this point, I'm about 7mph quicker before we both get on the brakes in a similar fashion, Tom perhaps slightly later. We carry similar speed at the apex, but Tom picks up the throttle earlier, yet doesn't get as good a drive off the corner. This could be because of how early he was on the throttle or it could be because my car accelerates better at that speed. It's actually because Quarry is borderline between 2nd and 3rd gear in the Spire. Whilst 2nd gear would be faster accelerating, it'd require an upshift very quickly, which might not work out quicker - something to try, Tom!

Once into the power band, Tom starts to close in over about 90mph. The slower exit from Quarry means he's 3mph slower down most of the straight and he brakes earlier into The Esses. Despite the earlier braking, Tom carries more speed through the apex and gets on the throttle earlier. This works well, especially as I miss a gear between The Esses and Old Paddock, but whilst I take Old Paddock completely flat, Tom's slowing down. There's almost 10mph between us at Old Paddock and it takes the whole length of Hammerdown for Tom to recover that speed. This clearly shows the straight line speed advantage of the Spire.

Into Tower, I'm slightly later on the brakes than Tom and brake less, resulting in a higher apex speed. Tom's trace suggests he realised he braked too early and accelerated into the corner as the dip appeared immediately after braking and before a long period of acceleration. The trace also shows Tom was on the throttle early and this is shown in the peak speed before Bobbies, which was about 2mph higher. I was again slightly later on the brakes by about 5 metres and braked harder, but slowed to an apex speed 2mph slower, the reason for which isn't apparent until you see the video. I missed the apex by a country mile as I struggled to slow the car for the corner! It all evened out by the time we both crossed the line at 84mph.

I should mention that there will be no data debug from Clay Pigeon last week as I failed to make it to the event when the gearbox on my tow car failed. The next entry on the blog might be entitled 100 reasons not to use Brunel Ford given their ironically titled Service department's ability so far. Many thanks to Torbay Motor Club who kindly gave me a refund on my entry fee despite me pulling out by telephone after signing on had begun. I didn't ask for it and the sentiment is appreciated! Here's hoping I make it there one day...

[andylaurence]

It all started over a beer (don’t they all?) whilst celebrating a friend's birthday. I said to Dave Greenslade that the MIRA event I’d entered had been cancelled and I was considering entering the Curborough event on the same day instead. “Fancy it?” I said. “Yeah, alright then” replied Dave. It was on! Entries in the post, we decided to enter a team in the teams event as well as individual class entries. Neither of us had been there, so it was a bit of a punt.

I’d not driven the ADR since the Dick Mayo thanks to the drama of my tow car breaking down on the way to Clay Pigeon and putting me out of action for Llandow the following weekend too. I’d also not tested the new steering setup that had been altered to get more lock for the tight corners after wider wheels had reduced lock to practically nil. Curborough being tight, I was worried, but not as much as the fact that I had no wets. Not knowing whether I could retain the wider wheels, I’d held off buying a second set for the wet tyres in case I had to change to different wheels and the forecast was horrendous!

I drove through rain on the way to the event and dark clouds loomed ominously. The forecast showed rain coming later on, but nobody really knew when or how much. Dave arrived in plenty of time (yes, really!) and we walked the course to see what we’d taken on. Our summary was that the first left was completely flat (we didn’t bother walking down there) and it really was quite tight in places. Maybe 4th gear at best down the straight.

Dave was out first in the Evo Half and came back in saying it was like an AutoSolo. Both of us have done quite a few AutoSolos, so that sounded good to me. He also said the first corner was not flat. With less power and more grip, I didn’t think it’d be a problem for me. He’d done a respectable 69 second run though!

I headed down for my first run. By the time I’d got to the line, the temperature gauge was showing 103, which was giving me palpitations! I took off from the line and snatched second before the first corner. It’s not flat. I took a big lift and dabbed the brakes into the second corner. I drove slow and steady as I had no idea how much grip there was and whether I had enough lock to get round the corners. I gave it a bit more on the second lap and recorded a mid-61. My target for the day was sub-60 seconds, so I was happy to be so close after such a tentative run.

After second practice, both of us had improved. Dave was in a commanding lead of the class, 1.5 seconds clear of David Pearce in his vastly more powerful Impreza. I was 0.36 behind my nemesis, Clive “I always pull one out of the bag on the last run” Wooster. With only 40 cars on the entry list and just 60-70 seconds per runner, we had time for the first timed run before lunch. Black clouds were looming, so it was now or never.

Dave was doing well and pronounced he’d do a 65. I was doubtful and said I’d buy him a pint if he managed it. He did. That put him fastest of all production cars, irrespective of engine capacity. Not bad for a £1000 car that’s not even seen a sponge for years, let alone anything to make it go quickly! I shall be very glad to buy him that pint at club night on Tuesday; a well deserved prize. I also improved, getting down to a low 58 second run, just 0.04 off Clive’s time. As the rain came down during the final few cars, that appeared to be the end of competition for the day.

The rain quickly let up and the track dried over the lunch break. When battle resumed, Dave was able to get within a few tenths of his earlier time. His closest challenger went 2 seconds faster, eclipsing Dave’s best and taking the top position. I went 0.7 quicker, but Clive lopped off a second. Two bridesmaids. Or was it…

We were so far ahead of time, there was a third run. Dave took to the track, but was unable to improve his time. Second place was all he could muster from his pretty special time. The only consolation was that he recorded a 2.09 to 64 feet and that the man who beat him had enough power to be 8mph quicker over the finish line, which probably made the difference.

It was all or nothing and, as thunder and lightening crackled and sparkled around us, I headed to the start line early to try and get a run in before it was too late. Clive was glued to my rear wing. For the first time ever, I did a burnout up to the line to warm the rear tyres. I lopped 0.1 off my 64 foot time, so it was worth it! Clive was 0.23 faster still. I was scrappy, but not overly so. Coming into the second lap, I missed my braking point into the tight right-hander and started writing a postcard to the apex. The rest of the lap was good until the finish straight and I missed a gear. I finally found the right cog, but the time was gone. I crossed the line at 96mph compared to Clive’s 108mph.

I came back to the paddock where Dave was waiting and he said I’d gone quicker, beating Clive’s previous best by a few hundredths. I cheered with excitement. Could it be that I’d finally beaten Clive with a scrappy lap that I thought was rubbish? No. Clive had knocked 0.3 off his previous best and pipped me again. I’m sure that without that missed apex and fluffed gear change, I’d have scalped him on my first ever visit to Curborough.

[andylaurence]

On Friday we headed to Llandow for a bit of familiarisation with the circuit and some more seat time. It was also an opportunity for Dave to get his first drive of the car. We planned a 90 minute session in the morning and 90 minutes in the afternoon. Each session was split into 6 stints, with each stint consisting of two 1.75 lap runs with a minute break between them. I took the first stint to check everything was OK and laid down an 84 followed by a 75, which would have been enough for a top 5 overall and a class win had I made it to the last event there. Not bad for my first two runs on the circuit, but the weather was better, so the results are not directly comparable. We alternated stints all day, which allowed the car time to cool down, although the tyres were badly overheating with the temperature of the day and inadequate time to cool down.

Dave's first run was tentative as he got to grips with the direct steering, slicks, wings and a sequential dog 'box, all of which were new to him. He picked up speed quickly and got progressively quicker through the morning, soon getting the hang of the extra grip, especially in the slower corners.

We were testing out WifiLapper as a new data logger during the test. A wheel speed sensor and brake pedal sensor had been hooked up with the rest of the data coming from the phone's GPS and accelerometers. Sadly, the phone's internal GPS is poor and the data was not good enough to be usable. The wheel speed sensor didn't work as expected, but the brake pedal sensor worked flawlessly. Unfortunately, the poor quality GPS and the broken wheel speed sensor means that whilst I can be certain when the brake pedal was pressed, I can't be certain where it happened, so it's not much use! There's onboard video to review and hopefully data from the RacePak logger, although I've not found time to check yet.

In the afternoon, a worrying wobble surfaced, which led us to the top mount on the upright. The brand new rosejoint had some play on each side and was allowing the front wheels to steer a few degrees without moving the steering wheel. At the same time, the gear lever developed some slack, preventing clean changes. The lever was because of a broken bolt, but with these two issues and just 30 minutes of the session left, we called it a day. The car had completed the equivalent of 4.5 sprint events in just 2.5 hours with nothing more than a quick spanner check over lunch. That was good news, but now there's data to debug, work to be done on the data logger and some fixing to be done. This might be the right time to fit that paddleshift...

[andylaurence]

I nipped out to the garage last night to investigate the cause of the dodgy steering at Llandow. We'd ascertained that the top rod end holding the upright to the wishbone wasn't firmly attached to the upright, but was it play in the rod end, the interface with the upright or the bolt? It turns out there's the best part of 1mm play in the rod end. By the time it's multiplied out to the edge of the tyre, it's a lot of free play in the steering and a little camber too. It must be a faulty batch as both sides have failed and they were brand new before the Dick Mayo Sprint, so they've done 4 laps of Castle Combe, 10 laps of Curborough and a few dozen laps of Llandow before failing. I have spares as they were only replaced due to unknown age, rather than wear. I'll order up some new ones too.

I also had a look at the data from the IOIO. There's no wheel speed data coming in. I'm not sure whether it's looking for a high or low signal. Also, as I'm just grounding the pin, it's very noisy when not grounded. I'll connect it via a resistor to ground and another to +3.3v and use the switch to either ground it or tie it to +3.3v and see which works. The brake pedal sensor was successful and you can easily see when the brake's applied. I'll do the same wiring as the wheel speed sensor to remove the noise though.

Also broken on the day was the gear shift. I've replaced the 6mm bolt and nut and lubricated all the joints, which seems to have freed it all up. The spring has definitely not broken as you can feel it by hand once the linkage is removed. That's good news as it's probably an engine out job to get to it. There's still some play in the linkage, but it's minimal and I think that's all in the rod ends. At some point, I'll weigh the linkage as it's quite substantial and that might offset the weight of the paddleshift kit, making it a more attractive option.

The final piece of work is to fix the second plane of the rear wing by attaching a slot gap separator again after the last one fell off at Llandow. I didn't prep the surface for the epoxy, so it came off with all the vibration. I also need to sort out something neater and tidier than the bits of skirting board I knocked up to get it in place quickly over the Summer!

[andylaurence]

I've knocked up a quick schematic of how I'll be hooking up the sensors to the IOIO. The brake pedal switch is too noisy, so I'll tie it to ground normally, with the switch pulsing high. Wheel speed currently grounds the pin, so I've done the same and tied it to ground with the sensor pulsing high each revolution. I suspect WifiLapper will pick that up rather than the grounding of the pin I'm doing at the moment. I'll also hook up the fuel sender, oil pressure switch and oil temperature senders that are currently unused. After that, I'll work out a strategy for the sensors currently plugged into the ECU. I need to be sure that tapping the sensors does not affect the ECU and that I can get a 0-3.3v signal.

[andylaurence]

I've been wondering whether the previous diffuser design knocked up last winter was too conservative. Given that there's no regulations over the shape of the underfloor, why should I start the diffuser at the rear of the chassis? Why not start it behind the front wheels? The sidepods are not structural and there's space either side of the chassis. This brings the centre of pressure forwards and provides a much larger plan area for the tunnels, thus a little more downforce, I hope. This theory would allow a tunnel each side of 3.1m, which means that even an 8 degree angle would result in a tunnel exit of around 43cm. That's high enough to mean that the driveshafts and wishbones will have to penetrate the tunnels and also high enough that the rear bodywork (~45cm) will practically meet the trailing edge. In fact, here's a quick graphic to show just how large the tunnels would be.



The tunnels can simply replace the existing sidepods. Two pieces of flat material are all that are needed to fabricate the replacement sidepods, so it's easy to build and, if I pick the right materials, I could save a good amount of weight as the sidepods are made of beefy fibreglass and clearly overweight. In order to fit the new sidepods, I'll need to relocate the radiator, oil cooler and the exhaust. The radiator can be canted over and placed over the tunnel. This raises the centre of gravity, but hopefully not too much. The same can be done for the oil cooler on the other side of the engine bay. The exhaust is not so easy and I'll have to come up with a solution to that. The final thing to think about is that in order to get it on the trailer, I'll need to pad out the bed so the car sits a couple of inches higher and clears the front crossmember of the trailer.

So there's a weight benefit, but can I quantify the aero benefit without trying it out? I don't think I can without decent CFD and that's beyond my means. So the question is whether I fancy spending the time and effort on something that looks like it should prove beneficial. Comments appreciated!

[andylaurence]

It's been four months since my last post. I had a winter project, but with a lack of time to work on it in November and December, I'm out of time, so that's shelved for the future and my attention has turned to getting the car ready for the new season. The focus is on fixing broken parts, reducing weight and reducing drag. It's clear from last year's data that there's too much drag on the car as it runs out of puff down the straights at Combe.

The rear wing is a big dual plane device after last year's upgrade and hanging it high up in the airflow means it not only gets nice clean air for best downforce, but it also adds to frontal area, increasing drag dramatically. The simple fix is to move it downwards and rearwards to fit behind the rear wheels, slightly above the exhaust. There's two main benefits to this; a reduction in frontal area and corresponding drag and better interaction with the underfloor thanks to the proximity to the diffuser exit. A tertiary benefit is that the proximity to the ground allied to the large end plates means it will benefit from some ground effect.



The sidepods are effectively working as big parachutes. Whilst I cut the front open last season, I found no way to exhaust air from the rear, so the high pressure air from the wheel arches is collecting in the sidepods causing drag and lift. I've worked out that a single aluminium panel can be cut simply with two folds to replace the existing sidepods. The sidepods will provide a good exit for air from the wheel arches, exhausting it to the side, with the whole sidepod weighing around 7kg. This should give a weight saving, reduce drag and reduce lift.

The new sidepods also provide an opportunity to add a small front diffuser to the trailing edge of the front splitter, which will again be cut from a sheet of aluminium with 6 simple folds. The final piece of aero work is to block up the air intakes on the front clamshell as they are causing a buildup of high pressure air under the clam-shell.

Most importly, the broken things need fixing. Fresh oil and new rod-ends to replace the broken ones on the front uprights are the key things to fix. The rear wing endplates also need re-attaching as the existing mounting isn't up-to-scratch as well as a replacement spacer between the main plane and flap.

[wesley123]

really cool to see how your effort continues!

How did you exactly come to the conclusions of the things you improved? Have you done some windtunnel testing or CFD?

The front diffuser is something I am in particular really interrested about.

On the drag side, I think you will gain more in instead of cut drag, dry to improve horse power. Also a lower angle rear wing might be useful in this case(iic you had some stalling issues?). I am not certain about the lower wing height, it sure will cut drag, but by the position you will push the CoP rearwards. On the other hand also the bodywork over the rear wheel will block the rear wing in my view.

Do you got some more pics of your car now? I would love to see how it has improved!

[andylaurence]

Luckily for me, an ADR was put in the wind tunnel and the results written up in Aerobytes in Racecar Engineering last year, so I have some raw data to work with. I've talked with Adrian Daniels of ADR about what was done to achieve these improvements and based my work on that. I'm also at the very bottom of the hill as far as aero goes, so I'm at the point where you can see changes that will improve things with your eyes and CFD/tunnel analysis is overkill!

The front diffuser is quite small, just 50mm high and 180mm wide on each side and attaches to the trailing edge of the 1800mm wide front splitter. It's simply a piece of aluminium folded in the right places so that it effectively clamps onto the trailing edge of the splitter in double-shear. Its shape defines its strength. I'll post pictures when it's been cut - I have a template here in cardboard, but it's not very neat!

I'd love to improve power, but that's not cheap. My parked winter project was going to improve power significantly. Simulations suggested a couple of seconds a lap, but I don't have time to complete it before the season starts and once it gets going, the season is pretty hectic so there's no time for mid-season upgrades.

I've already lowered the angle of the rear wing and stalling is no longer an issue. I agree that the CoP will likely move rearwards, but the addition of the front diffuser will bring it forwards and the reduced efficiency of the rear wing from the less suitable location will offset the leverage it has from the rearwards placement. I've confidence that the rear wing will continue to work in the less favourable position as Dauntless Racing have been able to make theirs work, both on track and in CFD.

The car's unchanged since September. The work is upcoming in the next few weeks. The only change so far is a trial fitment of the rear wing in its new location.

[andylaurence]

I've been playing with Xoomspeed's Xoomcentre, which has been recently modified to import CSV files. Download it here. It's extensible with plug-ins that you can write yourself and it's really quite easy to use. The CSV importing is brand new, so it needs massaging of the CSV files to import correctly. It's also reliant upon GPS longitude/latitude data to accurately report the run times, so as my RacePak logger doesn't output that information, I don't get the run times and I have to manually set a few things like graph ranges as the time slip shows massive variations after the finish line!

Despite these shortcomings, I've been able to make good use of the software to show me things that I've never been able to see with the RacePak software. I've written a plug-in to handle the RacePak CSV format and it's now a breeze to import any of my logs. As a test, I decided to compare my final run at the Dick Mayo Sprint at Castle Combe in 2011 and 2012. I imported my 2012 data, set it as the reference run and then imported the 2011 data. After removing the friction circle graph and gear data, I added speed from the reference run to the 2011 graph. It took 90 seconds from start to finish and this is what I was left with.



In 2011, I managed a 68.29, whilst the 2012 run was a 66.57. The graph hasn't lined up perfectly and there's an error of almost 0.5 seconds in favour of the 2012 run, yet the official results show the 2011 run was faster in the first 64 feet. From then on, the accuracy is much better, so I'd put this inaccuracy down to the inaccuracy of GPS at low speeds. Immediately, it's obvious that in 2012 the car was a little faster from 40-90mph. This will be down to the gearing, which was shorter in 2012 thanks to the 14 tooth front sprocket replacing the 15 tooth of 2011.

Moving on to Quarry, I was tentative over Avon Rise in 2011, some 8mph slower in places, having lifted off over the crest and this cost a tenth. The downshifts happened in a similar manner in both years, whilst in 2012, I carried 6mph extra apex speed gaining another 0.16 through the middle part of the corner. The lighter weight and bigger rear wing were likely to be good reasons why I was faster, allied to greater familarity to what is a surprisingly tricky corner. The higher apex speed resulted in a faster exit, which led to a 2mph run down the first half of the Farm Straight, netting a further tenth until the longer gearing of 2011 caught up thanks to one less upchange.

At the end of the straight, I was much more agressive in 2012, braking 35 metres later and gaining 3 tenths. I lost two tenths of that with a lower speed through the first half of The Esses. From memory, I'd braked too late and missed the apex. It was even by the exit though and pulling more revs through Old Paddock from the different gearing meant another half a tenth. At the end of Hammerdown, the ratios pulled another half a tenth and I again braked 30 metres later, being some 15mph quicker on the way into Tower and making up a whole six tenths from the braking point to the apex. Despite the massive gains, I'm sure there's more to come here too.

A higher apex speed and better gearing going down to Bobbies netted another half a tenth, followed by a further half a tenth on the brakes into Bobbies. The final time came from getting on the power earlier out of Bobbies and I gained another half a second in the process. In summary, I can see that there was a marginal benefit in the shorter gearing, but that most of the improvement in lap time came from behind the wheel with increased knowledge of the track and a bit more bravery.

Most importantly, this shows how quickly and easily I was able to review the data with Xoomcentre and whilst the export of the graph could have been quicker, it still took less than a minute to do. I shall be reviewing all my data with Xoomcentre from now on - my laptop might even make it to the paddock for lunchtime reviews!

[flynfrog]

Excellent write up mate. There are some older threads on here from way back in the day it airfoil equations used by ground effect F1 cars for there tunnels. They could be a good starting point.