A Certain Ratio

The last time I blogged, which I see was a little while ago, I was hopeful of getting back to the sort of times I was doing in 2019. I’ve kind of done it, and learned a lot along the way.

Since the first 2 Curborough rounds I’ve done a Prescott; mixed conditions and not fast enough, Gurston; pretty much on the pace of 2019, and a Loton. Which was very wet at times but more of that later.

There’s not much to say about the Prescott event, which was notable mainly for the fence between the somewhat sparse spectators and competitors. Wet but drying through the day, resulting in a best time 1.2s slower than my best in 2018. I’ll blame the conditions.

Gurston was an altogether different matter. I’ve been to some really hot Gurston events, but this was the hottest. When I put the ignition on the temperature went up to 40C. Without running the engine. The marshalls were running the event so that the queue for the hill was kept to 2 or 3 cars – so it was a matter of firing up the engine and driving to the line and going. Ironically I was running the engine colder than ever before and it liked it; seemed happier revving, and I was quick across the line. Quite quick and still in 3rd. I managed to match my second best time and it got me thinking about gearing.

The other competitors in Pre 94 FF were super helpful with advice on ratios, and when I checked the gearing I was running it was somewhat longer than anyone else was using. If ever they run a sprint at Indianapolis I’ll have some ideal cogs. Armed with the info from Charlie, Les and Peter, I purchased some new ratios from Mark Bailey who was also super helpful. Essentially 3rd became 4th and 2rd and 3rd filled the gap above 1st which I kept. Jo White gave me a masterclass in changing ratio’s in a Hewland MK9.

The first opportunity to try the new gearing was at the Pre 94 FF round at Loton at the weekend. (Beware the book of excuses is in play from now on.) Conditions couldn’t have been more different from Gurston. I don’t think I could have overheated the engine had I wanted to. And at times it was very wet. Difficult to believe it was August. On the first run I couldn’t believe the difference the new gearing had made. Even though the run was dry every time I touched the throttle on the way out of a corner the back started to slide. Different driving style required and almost certainly some adjustment to ARB or dampers.

Photo by Rob Mcdonald

I last ran at Loton in 2017 in the Panda. I was pretty sure I remembered it. But one corner repeatedly threw me a curved ball. Fallow at the top of the hill is a lot further away than I remembered.. That doesn’t make sense so I’ll try again. When you approach Fallow it looks like the corner starts about 50 metres before it actually does. And on the first couple of runs I braked crazily too early. Which was a surprise given that when I was driving the Panda I used to be pretty aggressive here. 4 years is a reasonable amount of time I guess.

And finally the last issue that I had to sort out was a lack of speed on the Cedar Straight. Which isn’t straight. Thanks go to Charlie Reilly who told me that I needed to be in 3rd through the section before the straight, which increased my speed on the straight a lot. I didn’t set the world on fire at Loton (whereas Tom Weaver did with the times he was putting in) but it is still my favourite hill.

So next up I’ve got a session at Anglesey in one of their Formula Fords, simply to get some more seat time and to see what a different car feels like, another Gurston where I hope the new gearing pays dividends, and hopefully the MAC meeting at Curborough in September if my entry has been accepted.

Getting back on track

It was pretty tough getting entries at the end of 2020. I didn’t get out. But 2021 has been different.

I’ve already done two meetings in 2021. Both at Curborough. It’s great to be back.

I’m not exactly at the front of the pack when it comes to outright speed, but the amount of this speed that I’d lost over the enforced break was a bit of a shocker. 16 months without driving a racing car is a long time. I thought that I was actually slower than I’d been in the Panda; that wasn’t the case, but it was a close run thing.

The April meeting was fun, and I pretty much classed it as a shakedown. I needed to be more competitive in May. Luckily I finally found a Curborough simulator track and managed to put in a decent session in a virtual Formula Ford, which seems an unlikely clash of technologies. The results were really interesting.

My first runs are generally warm up/shake down runs, but this time I was more focused on what I needed to do from the off. And I managed to put some of the things I’d practiced into action from the off too. My line at flagpole is much improved. The sim session showed that you could go through the crossing faster than I do in real life, and in the sim I was settling the car with a bit of left foot braking there, which I didn’t do for real this time. The sim session had somehow improved how I approached the circuit and driving it.

In the first run I managed to select 3rd instead of second for the middle but of the run, meaning the car didn’t pull at all. Sorted that for the other 2, though I think it’s screaming a bit in second. Upshot was that I was over 2 seconds faster in May than April. And managed to come 2nd out 3 in my class.

Unless I enter some other rounds this year I won’t be back at Curborough until the HSA end of season event in October, where I’d expect to see a much bigger entry. By then I’ll have done some more time in the sim and a few more rounds. Hopefully I’ll be down to the sorts of times I was doing in the summer of 2019.

Pictures from the May meeting by Peter Scherer

If I can’t draw it, then I don’t understand it

This blog originally started with something about how James Hunt is a big figure in motorsport, then finally totally lost its way describing people I’ve met in the last couple of years.  The last thing this needs to be is a linked-in style, “isn’t everyone wonderful, aren’t I amazing” posting.  For the record the parties involved are listed at the end of this.  Much more appropriate.

So instead, something a bit more factual.  I’ve become peripherally involved in a project to restore the two Dastle F3 cars run by Hesketh in the 1972 season.  These cars are relatively well documented if you take the time to look.  F3history.co.uk has some pictures (http://www.f3history.co.uk/Manufacturers/Dastle/dastle.htm) and a brief chronology, and there are podcasts and even TV interviews which cover the early history of the Hesketh team. Even ebay managed to serve up a print of the classic Druids formation shot.  So there’s no real need to cover the same ground here.

What I’m going to talk about here is something more specific to the world of CAE that I inhabit.  When I rebuilt my Van-Diemen RF84 I started to draw it as I went along.  The idea was to have a reference for new parts and to understand how it went together; but that project went nowhere when I struggled to get the nascent CAD system I’d chosen to actually model a complete car and I discovered that I could buy everything I needed off the shelf from Ken and Simon at Universal.  But the Dastle project is different.  There is no Universal Racing Services for Dastles, and knowledge of how they work isn’t easily come by.  So this time a model seemed a necessity rather than a luxury, and something I can contribute to the project.  To quote Albert Einstein – “If I can’t draw it, then I don’t understand it”.

Being a simulation engineer my plan is to create a simple representation and add complexity as needed.  Already this process is yielding huge benefits, for me at least.  In the total absence of drawings, notes and calculations it’s allowing us to get inside the head of the designer (Geoff Rumble according to f3history.co.uk  – if you are still out there Geoff I’ve got a 1000 questions for you) and reconstruct the most important thing – the original design intent.  You can’t get that by pointing a hairdryer at the car.  Some parts will need scanning, though not as many as people might imagine; mostly this is about putting in the hours with a tape measure, rule and vernier.  And taking enough photo’s, which I almost never do.

This is definitely “a how it started” post.  A real work in progress; I’ll post updates as the model is developed.  As usual a CAE representation has generated significantly more questions than answers and prompted a lot of thought about what the design intent was.  The overarching question for me is what was designed and what just happened as the cars were manufactured.  The model won’t be able to answer that; talking to right people and looking at contemporary references may help there (I’ve got a copy of Racing Car Design and Development by Terry and Baker which may shed some light on the design process in the early 70’s).  But the model does provide a focus for this activity and a means of understanding how things function and how they can be engineered to run properly again. At the very least it’s something to point at and scribble over whilst we try to work out a strategy for the restoration process.

And as the man said “If I can’t draw it, then I don’t understand it”.  TBC.

Who’s who section:

BIOS-Sport: Peter Sneller and Arthur Griffin (Arthur drew the wheels and tyres)

Classic and Sports Finance: Robert Johnson (Owner)

The one where I don’t make a wind tunnel model.

I hate friends.  No idea why I used a Friends type title.  Anyway.  This is about me not making a wind tunnel model.

wt1

Unless you are supremely gifted, and as we all know I’m not, your first goes at anything are likely to be a bit, well, er, rubbish. In fact I personally don’t think a typical person can learn to do something unless they actually enjoy being crap at it. So you have to get in there, have a go, mess up and then see where it all ends up.  (36 years after making my first toy aeroplane I won an event at the Nationals, so it isn’t always a rapid process.)  Where I’m headed with racing cars means that I need to get a grip on aero (see previous stuff on CFD) and that means wind tunnel testing in some form.  Which means I needed to get in there and mess up a wind tunnel model.

I used to exchange tweets with the recently departed and already much missed Richard Divila.  And in one exchange he said that I’d be surprised how much you could learn from a really basic wind tunnel.  And I guess that means from a pretty basic model.  My son, is, as I write this, finishing his dissertation, which might not surprise people to learn is a CFD study of a racing car.  In fact the one I used to form the basis of Openfoam work in the last blog.  And for that the initial intention was to do some wind tunnel validation of his Fluent work.  It’s entirely possible that even without Covid19 we might not have got round to it.  But I did get what is very much a comic book first hack out of my system.  And that’s a good thing; it really is, because it gets you somewhere good.   So what have I learned

You can’t 3D print everything: Unlike the current UK government I don’t think 3D printing is the only manufacturing show in town.  It plainly isn’t, but this project was my excuse to get a printer.  I chose to get a Dremel because I love their stuff.  It was a great choice and their aftersales service is really good, mind you it had to be after the first one didn’t really work.  And I’ve learned a lot, especially about DFM, (Design for Manufacture) in this specific instance. I’ve been suprised in a good way about some of the stuff I’ve managed to make and surprised in a bad way.  I still haven’t changed my mind about the fact that 3D printing, at least with the machines I’ve used, is like machining something on a worn out milling machine – you can do it but it takes a lot of effort.  But I am pretty sure I could make half decent body panels for models up to about 20%.   What really aren’t good enough are the wheels.  They didn’t ever come out well enough, especially when I can machine a set from solid quite easily.

printer

One area where the printer really comes into its own is the production of small brackets.  These can be produced quickly in reasonable batches.  You’ve got to love that.

Anything that involves thin flat sections should be made from thin flat material, because if you print these you need to make them significantly thicker than scale and the surface finish won’t be good enough. A small CNC milling machine would be really useful here. Floor components, wing end plates and other bits could be profiled from thin carbon fibre sheet.  The results would be pretty good.

Plainly to be useful the model needs to be very adjustable, and in its current form it isn’t.  Ride height and rake need to be controlled very accurately.  I’d say that what I need is some form of cross slide arrangement to support the model from a sting.  Ebay appears to be awash with them. So that’s that sorted.  Another thing that needs to be easily adjusted is wing position and angle of attack. I’ve thought long and hard about this and the only really neat solution is to use a series of mounting plates of different dimensions.  That small CNC mill again.  And finally the hack model uses a completely 3D printed tub.  What it really needs, and I’ve heard this from several sources is a central spine, which is what joins the model to the sting and the wings to the model.

wip

And finally the wings.  Obviously these have to be pretty accurate, and the ones I made kind of were.  They were printed in vertical sections and then assembled on a carbon rod. As a first stab OK and they would have been reasonably representative. At a larger scale I think this approach would have been fine.  So lets describe wing manufacture as a work in progress.  But at least we are in the game now.

It’s quite unlikely that this not really good enough model will ever see the inside of a wind tunnel. (With a bit more work it would be a useful exercise though) But it has taught me a lot about making things with AM and what a useful model might actually be like.  So not building a wind tunnel model has been a hugely useful exercise if I ever do it for real.  Meanwhile the CFD activity continues..

f750_3

 

Frei Aeroworks..

Back to blogging; I’m certainly not doing much driving.  My blogs aren’t popular at work these days, so its a kind of mixed CAE/racing car offering here this time.

CFD_Image_11

I’m convinced that in the Tony Southgate autobiography he says that there was a time when you could increase the downforce a car had by a third in an afternoon in the windtunnel.  (As an aside I managed to get hold of a copy of this before the price got crazy, it’s on Amazon for £550 at the moment. It’s good, but not that good.) Problem is that I can’t find that reference.  Maybe its in one of the other designer biographies; if its in the Murray book I’ll need to get training before I try to pick it up.  Maybe I imagined it.

Anyway, it struck me that much of the club racing scene must be operating like that still.  People put wings on cars and run floors and diffusers. But how optimised are these generally?  Not very I’d say, admittedly from a position of near complete ignorance.  Modern high end aero packages are designed and optimised using wind tunnels and Computational Fluid Dynamics, way beyond the scope of the average competitor.  Its how far beyond that I set out to discover.  It would be a lot of fun right?  Is there still the scope for big gains made relatively easily?  First, for me anyway, the easy bit. CFD.

Openfoam has been something of a curiosity for me for some time.  It’s a really decent CFD code, used by at least three F1 operations to my knowledge, and its free.  Not free like Nastran; that’s not really usuable in its free form.  Openfoam is free and really good.  But because there aren’t any decent free pre-processors (the bit where you actually build the model rather than solve the equations) I hadn’t tried it.  I’m not entirely sure how I stumbled across the FreeCAD openfoam pre-processor. But stumble across it I did.  And frankly its bloody marvellous.  The test was to see if I could create a vaguely useful, full vehicle CFD model that could tell me useful stuff.  Currently I only have an 8 core machine (with a low end GPU) so that set an upper limit on what could be achieved.  Anything I did had to mesh in less than 1/2hr and solve over night.

CFD_Image_3

Luckily my son has created a car model for his final year project.  Its a model of Jo White’s latest, so ideal on many levels.  Before attempting a full vehicle I ran through a few self created tutorials that I used to learn the system, especially the mesher, which as things went on seemed better and better.  Its a free version of a commercial program called CF Mesh written in Zagreb by a company called Creative Fields. If the idea of the free code is to publicise their work its certainly done its job.  Getting the CAD model right for meshing was the usual headache, but no different to other CFD codes I’ve used.  One thing that I thought might become a headache was selecting the internal faces of the flow volume for boundary conditions and mesh refinement.  I kind of stumbled into how you did this, and that meant the whole car workflow was feasible..  Another interesting thing about FreeCAD is that its all scriptable in python.  Which means I might be able to automate the whole process at some point, which would be really neat.

Paraview is another free program that is little short of incredible.  Its free, amazingly functional and again can be scripted. All the results here were plotted with it.

So the upshot of all this is very promising.  This is CFD, and I guess you need to know a bit about what you are doing, so its not for everyone.  But the results I’ve got so far with meshing times less than 1/2hr and solves of  3 to 4 hrs have been very encouraging.  I still need to add a moving floor (I know how you do it, just haven’t yet) and look at ways of doing drag/downforce calculations for more than just the whole car.  But I think I’ve proven that this sort of thing can be down at no more cost than that of the hardware. (Which now makes me want a 16 core machine for my office at home.)  This, I think, has big implications for club motorsport competitors, or  if not them, the smallest car constructors.  If you are the sort of person who is happy to hack about without support or training, and is as interested in the process as the result, this is really something. What we need now is a low cost wind tunnel to validate everything!

As an addendum to the above I’ve also been using a free multi-body dynamics code called FreeDyn.  I’ll probably blog about it at some point in the future.. It looks like it could be the MBD openfoam.

free_dyn_4

And I’ve also been playing with rendering.. which gets me thinking about doing an OF model of Matt’s MW01.  There’s a whole lot to do.. its just about finding the time.

render_1

 

 

 

 

The shock of the new.

This shouldn’t come across as criticism, but although the Radical SR1 is in many ways an incredible racing car, in one critical respect it differs from most others I’ve come across.

radical

When I bought the SR1 it only had one set of harnesses, so in what passes for me as deal making, I got the seller to knock the cost of a new belt off the asking price.  On top of that, when I drove it, the seat was a bit far back; however initial attempts to adjust the seat were doomed to failure. Fitting the new belt wasn’t any form of priority all the time the car wouldn’t go on the trailer, so I left it all alone for a while whilst the new ramps were fabricated. (Very nicely it has to be said.)

ramps

With the 2019 season now history I started work on these two issues.  Removing the seat seemed like a good place to start.  It would allow me to work out what was wrong with the adjustment mechanism and wouldn’t do any harm when it came to swapping the belts around.  Most of my projects start with me staring at things for a while, but staring at the SR1 didn’t progress things any.  How you removed the seat totally escaped me. So I emailed Radical. And was sent some instructions.  Which still made little sense, and were ultimately proven to be quite wrong.  Cue more staring at a car in a shipping container. The stand-off was broken by a friend who said “ the only way you’ll get that apart is to drill those rivets”. Once I’d done that, undone 2 bolts under the riveted section, removed the roll hoop stays and loosened the sidepods, I could actually get the seat out.  Well, I could after I’d lifted it up, reached underneath and removed the crotch belt connections.  (I only loosened the sidepods because taking them off meant removing the water radiator and oil coolers, because, barely credibly, these were fitted into the fibreglass mouldings themselves.)  The SR1 is in no way designed for rapid maintenance.  Like a racing car is.  I’d expected everything to be instantly removable.  You can achieve what took me over 4 hours in less than 5 minutes on my Formula Ford.  And that’s one which is famously difficult to work on. I’d guess that with a bit of practice I could get the time down to a couple of hours.  Certainly not the sort of thing that could be done at an event. I guess it’s all about the profit motive because making everything removable properly would add cost.  I’ll put it all back it as it is for the moment, its just too big a job to re-engineer it all properly now; annoying things like having to go to work see to that.

So my new, well it’s a lot newer than my Formula Ford, SR1 has sprung a few unpleasant surprises, but funnily enough I don’t seem to like it any less.. Before long I’ll have it all back together and can’t wait to get driving it again.

 

 

 

 

Jumping through hoops

The HSA season closer at Curborough is a great event, even though it was slightly blighted for me by a sticking throttle cable last year.   After the success of my last trip to Curborough (first trophy – did I mention that?) I wasn’t actually expecting great things, especially given the size and quality of the Formula Ford entry.

At the risk of looking like I’ve been thumbing through the racing driver’s book of excuses, I wasn’t feeling great.  So I was really happy to be half way up the class after the first two runs.  The first timed run was faster than the two practice runs, but most people seemed to have gone faster than me.  I’d fallen through the field somewhat.

I managed to psych myself up a bit for the last run, and it did seem a lot quicker.  Well it did until Flagpole at the end of the first lap. I overcooked the entry speed and then selected 4th instead of 2nd.  I headed out of the corner in 4th, and by the time I had realised what had happened the run was ruined.  I completed the second lap in 4th and my overall time was just over 70 seconds.  That, combined with the fact that my first split time was 6/10’s up on the first run, convinced me that I’d have been on for something a bit better than run 1. But it’s not like it was critical for the championship.  Clutch woes earlier in the year had seen to that.

One thing I should have done for this meeting is to have stiffened the front anti-roll bar.  The picture, by Andy Leivers, shows why.  The front suspension seems to have moved a lot in the corner, the nose is sticking up in the air, and I managed to slide the rear end at Flagpole. I guess I’ll be trying that stuff early next season.  And by then I’ll have done the corner weights and undertaken a dyno run. I’m going to do 2020 properly. 72388160_1348975145258395_1557035230890557440_n

But more annoying than messing up the last run was having the scrutineers taking an interest in my ROPS (roll hoop).   For 2 seasons nobody has picked up any problems and I’d even talked to the MSA about its legality during the rebuild.  Before I panic and get a new hoop fitted I’ll see if I can sit lower in the car, although taking the seat out will result in me sitting a bit close to the battery.  If it is the answer I’ll have to make a new, lower, seat.

So 2019 didn’t pan out how I wanted, but was pretty much as expected as soon as I had to tackle the the clutch. To prevent a re-run of 2019 in 2020 I need to make sure of the car’s legality as soon as possible.  Watch this space.

 

Vehicle Dynamics

There is no other way to begin this post.  I’ve finally won a trophy. In a motorsport event.  As you can tell I’m quite happy about that, and it’s made the BARC 2019 season closing event at Curborough quite special for me.  But that isn’t what this posting is about.  What it is about is the fact that at last I think I can tell what the car is doing.

I mentioned in the post about the Gurston meeting that I was beginning to have some ideas about adjustments.  This stems from the feeling that the rear end tends to break away before the front.  I’m not sure if the video shows this effect or my inability to drive properly.

Whether it is or it isn’t I’ve felt this at Gurston too – coming out of the Karousel the car seemed very tail happy.  Same on the way out of Ashes. Maybe even on the verge of spinning.  I’m going to stiffen the front ant-roll bar a bit, and see what that does. If it does anything I’ll have made a big breakthrough.

At Curborough something else that I noticed was very large drop-off in grip on the early runs.  P1 was probably the coldest run I’ve made in some time, especially at Curborough, and grip levels, I’d imagine as a result of this, were very low.  These levels improved during the day, but were never near what I’d experienced at Gurston or even some tests at Curborough in high summer.  I need to work on getting some grip back, possibly by lowering the tyre pressures.

So I’ve reached the point where I think I can start to make decisions about set-up.  Its nearly as gratifying as getting a trophy.  OK, nowhere near as good, but still gratifying.

 

What a difference a day makes.

A year ago I ran at Gurston and was frankly a little puzzled that I wasn’t faster, and what was more puzzling was that I couldn’t work out why, or what I could do to increase my pace.  I get that a lot.  But on Sunday it was quite literally all change. I took 1.54 seconds off my PB and know how I could go faster, though with the back end going light through Hollow it might be some time before I go quicker there.

The clutch issues are well documented in previous blogs. After the run at Bicester to check that the clutch worked I wanted a proper test to see what impact it had on how the car actually went.  A spare space on a test session the Thursday before the Bank Holiday Gurston seemed ideal.  For once the AIM datalogger was actually working and after the test some data could be coaxed out of it (this was because temporarily the device firmware and PC data download software actually worked together.  It didn’t last.)  Sprint and Hillclimb times from the device are almost useless, but speed data does provide some useful info.  And the top speed down the straight was up 4MPH on the last visit.  Promising.  Working out what was happening with the engine RPM was somewhat less straightforwards. I need to work on that.

curb_data_22nd_aug_2019

My first practice run on the Saturday at Gurston was just over 41 seconds, pretty typical for me, but encouraging as a first run.  The next run was a PB by 0.63 of a second, but as we all know they don’t count in practice.  It was also tantalisingly close to my personal target of 39.something at 40.05.  The final run of the day was 40.32, which I would have thought fast last year.

gurston august 2019_3

Sunday’s practice run was a decentish 40.7, leaving me with something to prove in timed run 1.  After a bit of socialising with the great and the good of my hillclimbing and sprinting world it was time to run.  I got a decent clean start (clutch still working then) and took a reasonable amount of speed through Hollow and into Karousel.  I still wasn’t fast enough through Deer’s Leep and was OK through Ashes, but that was because I’d been paying attention there on previous runs.  Getting a decent run from Ashes to the top of the hill isn’t as straightforwards as it sounds, but I made a decent stab at it.  I was out of breath at the top, so had obviously been trying quite hard.  At some events you get to see your time after you cross the line, but at Gurston you have to get a ticket from the timing hut.  As getting in and out of the RF84 is a bit of a headache I stayed in the car and headed back down the hill not knowing my time.  A state of affairs which continued, although Jo White told me I’d done my 39 when I saw him near the start line.  Eventually I ended up in the start line office and was given my time directly from the system. 39.14.  Not just a sub 40.00 but over 1.5 seconds under my PB.  To say I was pleased was a bit of an understatement.  A little while later the printer produced written proof.

39s

I said that I was so happy that I’d be waving to the Marshalls on the way up the hill on my last run, not just the way down.  As so often with my final runs I might as well have waved at everyone whilst progressing sedately up the hill.  But frankly I didn’t care.

gurston august 2019_2

The big point about my performance wasn’t just going sub 40.00, and not that far off a 38 (which would have been respectable as Sam Lester’s fastest run was 37.95), but the fact that I can see where I can go faster.  (Looking at the time sheet you can see my finish line speed is 4 mph down on my fastest run for example.) And for the first time I’ve even got some ideas on car set-up.  Mind you I’ll need them.  It looks like Trevor Willis has signed up to drive in Formula Ford at the HSA season closer at Curborough.

gurston august 2019_1

Clutching at straws and beyond

The clutch issues that dogged the final runs at the Gurston practice day didn’t get any better.  And I didn’t want to blog about them until I’d sorted them.  I didn’t envisage it taking this long.

The first thing I did was to flush the clutch hydraulic system to remove all traces of air.  It seemed a bit better, but not radically so.  I entered the May Curborough event, with the idea of seeing if things were better.  Spinning off on the first corner of the first run wasn’t exactly on the planning sheet, and finding the battery didn’t want to restart the engine wasn’t (as an aside I seem to need to replace the battery every season, which is odd.) I managed to put in some semi decent times in the end, but it was evident that the clutch still wasn’t good.  At that point I made the decision to investigate things properly, and that meant removing the back of the car.

towed

I needed to make removing the gearbox and suspension a one man job, so I made a trolley, which with appropriate packing pieces, allowed me to simply undo a few bolts and slide the back of the car off.  It was nearly that easy.  When I got inside what I found was a bit surprising.  The bobbin which actuates the release fingers has a bearing pressed into its ID.  And that bearing had fallen out and was sitting on the input shaft behind the release fingers.  I wish I’d taken a picture of it. This and the resulting millimetres of clearance between the bobbin and shaft may have gone some way to make the clutch of the RF84 a somewhat unpredictable device.  It would also have explained something that happened the first time we ran the car in Bicester.

When I depressed the clutch pedal in those first runs there was an odd modulated grinding noise.  It caused a lot of confusion amongst some pretty experienced people, including one who had made lots of numerical models of clutches and release finger operation in the late 90’s.  I’d say that the bearing had parted company from the ID of the bobbin pretty much instantly.

The bearing and bobbin weren’t usuable, so I had to source a replacement.  And this provided another puzzle.  I measured the dimensions, as Mark Bailey told me that there were two sizes.  Od thing was that the bobbin in the car wasn’t either of the standard sizes.  So it was back on with the rear of the car to measure the freeplay.  Which indicated that I needed the longer of the two options.  So I ordered this and refitted it. The longer bobbin definitely made the freeplay more reasonable.

With the rear of the car reassembled I tested the car in the drive (not ideal but there were few other options.) The clutch felt instantly better.  In fact totally different.  As the car has only ever been driven by myself and my son its possible that the clutch had never been right, and perhaps we should have known that.

Finally I managed, this week, to drive the car round a track; I hired the circuit at Bicester Heritage for an hour.  Again the car felt very different, and when I returned to the pits after the first run, the track manager said “there’s not a lot wrong with that”. He had a point. In that run I’d have to say that the car felt really good, and if anything seemed to be running better.

I’ve got another test session booked at Curbourough next week and am running at Gurston over the bank holiday.  We’ll see if the improvement translates to times by then.