Locost: Baffled and Gated Sump

This is the first part in a series I like to call “What’s wrong with the Locost?” or WWWTL for short. I promised myself I would do a Trackday this year and as things are starting to slow down for the summer I now have time to prepare the car.

Firstly, the Locost is not perfect; I can easily stand and point my finger at a million things “wrong” with it and there are a few things I can’t really live with that I feel I need to amend before it starts turning laps.

You see, as you fix the fundamental setup issues on your home built race car, and attach a set of half decent sticky tyres, you’ll start to go around corners much faster. This has a big effect on the longevity of the car, increasing the loads through the suspension and engine, and you will definitely find some design flaws if you are lucky enough to have any. If you applied good engineering when designing/building said race car you will hopefully have no issues. You would have considered all loading conditions, and you will suffer no tears/breakdowns/failures.

Something I feel I did not consider enough many moons ago, and potentially completely overlooked, was oil starvation.


The Oiling System

I’ll do a short run through of the oil system in a combustion engine to give you a basic idea of what we are dealing with.

Firstly, oil lives in the sump pan. This is essentially a bucket of oil at the bottom of the engine which stores a supply of oil for the engine; this is directly under the rotating crank. Oil is sucked out of the sump by a crack driven pump and forced through an oil filter, which removes all the small particulates which might potentially cause damage upstream.

From the oil filter it feeds the main oil gallery which gives oil to the main bearings and crank, ensuring there is adequate lubrication and load support for the connecting rods. The main gallery also has a vertical feed going vertically towards the head. This lubricates the cam bearing surfaces and pressurizes the hydraulic lifters.

Oil slowly leaks out of the bearing surfaces, and flows back to the sump thanks to gravity. The restriction between the pump and atmosphere (the effective hole size in which the oil leaks out of) leads to a pressure build up in the oiling system. Once a given oil pressure is reached a blow-off valve allows oil to flow straight back into the sump, restricting how much oil pressure will be achieved. Therefore the less wear on an engine, the greater the restriction and the greater the running oil pressure (until the blow-off valve pressure, which is usually 60-70psi).

As an aside, when an engine is cold the oil is thick and viscous, and therefore the oil pressure is higher.

G13B Oil System

So, if for some reason the engine is starved of oil it will pump air and the oil density will drop, flowing easily through the gap in the bearings and reducing the oil pressure. Air does not lubricate or bear load very well, leading to excess wear and potential engine failure.

In short, oil pressure is an effective measure of engine health.


The Sump

So how does oil starvation occur? Well usually its one of three things, a lack of oil in the sump (check your dip-stick!), aerated oil or oil slosh away from the pickup. Keeping the sump full is easy, and really there is no excuse for having a low oil level, however the other two are not so obvious.

Oil aeration occurs when the crank stirs up the oil in the pan and fully/partially turns it into foam. This can be designed out with use of a Windage Tray; more on that later.

Oil slosh occurs due to the accelerations that are applied to the oil volume. If you achieve a lateral acceleration of 1g (at the apex of a corner for example), there will be a force pushing the oil against the side of the sump equal to gravity and it will set in triangular shape; as illustrated below:

Oil Slosh

In this case the pick-up is partially open to the air and pumps that as opposed to oil. This leads to bearing on bearing interaction, friction, wear and potential engine failure. The secret to good sump design is to reduce the chance of the pick-up being exposed to free air.

You can do this by using a tall deep sump, or by baffling and gating the sump. As the Locost is a small tightly packaged race car its nearly impossible to package a tall sump without running an impractically high ride height, so the sump needed to be baffled and gated, with an inbuilt windage tray.


Old/Poor Sump Design

My old sump was built from the flange of a standard front wheel drive sump, with custom sheet metal work underneath. The pickup was at the front and approximately central. It had longitudinal and lateral baffles with liberal drainage holes between each (making them almost useless) and a bolt in windage tray. It looked a whole lot like this:

Old Sump with Windage Tray

With the windage tray removed the baffles were accessible:

Old Sump Baffles

In hard right hand corners I think it was possible for the oil to slosh to the left hand side of the sump and expose the pick-up; as you can see there is no baffle in the central section where the pickup was located. The only saving grace of this design was its large capacity, giving minimal oil depth change when oil is trapped in the top end of the engine. Fortunately when I put slicks on the car it had terminal understeer and I don’t think I did any serious damage.

Given that the sump was off the engine, it was a great opportunity to inspect the oil/sump for particulates. The oil was clear of shiny aluminium bearing material, but there were some small bits of the cork gasket in the bottom; nothing scary but also suboptimal.


I was happy to move on from this design…


New Sump Design

The new design was going to be wider and shorter than the original, positioning the pickup in the middle of four separate oil chambers, each giving the pickup instantaneous oil in the case of hard cornering. Also, the windage tray would bias towards the pickups central volume, to flood it and reduce the chance of oil starvation.

New Sump Flange

Fabrication started by cutting out the main flange to mount to the block. This was bolted to an old junk fitment engine I had lying around (I use this for making engine mounts, brackets etc).

New Sump

New Sump

New Sump Windage Tray

New Sump Pickup

Then the windage tray was cut to match the sump and measurements taken from the chassis.

New Sump Central Chamber

The sides of the sump were then cut and tacked to the windage tray. The central chamber around the pickup was mocked in place.

New Sump Gates

Sump Baffles

Welded Baffles

Then the baffles were put in place to create the four separate chambers. Four gates were added to the central chamber to avoid oil moving away from the central chamber in hard cornering; these were made from steel door hinges! Note that they have limiting tabs to stop the gates going over-centre and killing the engine. The baffles were welded into the bottom plate to stiffen the sump and ensure oil does not escape the central chamber.

Sump Drain

I almost forgot to add a sump drain plug (uh oh!), so I welded in an M12 nut. It turns out M12 course thread is not a standard sump plug size (arg!) so I had to use an M12 bolt with a magnet epoxied too it; could be worse.

Oil Leak Down Test

Once the whole thing was welded together it was tested for leaks using some old oil and left to sit for a few evenings.

Painted Sump

After this it got a snazzy coat of Racing Red!

Closing Comments

The sump is now bolted onto the car and we will see if it causes me any issues. On paper it should be a great improvement over my previous sump and I’m hoping it will give the confidence and peace of mind its designed too.

Before Christmas I will have gathered some track data, covering a large span of lateral/longitudinal accelerations and engine oil pressures. In a perfect world there would be no drop off in pressure over the full span of achieved accelerations; but realistically I’ll  be happy with just very low drop off and a healthy engine.

There is still plenty to do before hitting the track- front wheel arches, rear lights, blah, blah blah… I will get there eventually!


Fabrication: That time I made an Exhaust Manifold

I’m going to try to document a few of my older projects that fell through the cracks and didn’t make it on to here. Hopefully you’ll find these little articles both interesting and informative… and there are pictures!

A couple of years ago I made an exhaust manifold for a friends Seven. Having seen the stainless manifold on my Locost he wanted one in the same “over the chassis” style. The manifold on my Seven was/is OK, it does the job, but its not my best piece of work; I was learning along the way. The Locost itself is a testament to my abilities at each stage of its build; some parts are better than others due to improving my fabrication skills as I went along.

Locost Exhaust Manifold

Locost Exhaust Manifold 2

Locost Exhaust Manifold 3

Locost Exhaust Manifold 4









This second exhaust manifold project benefited from everything I had learn’t and was properly jigged and built close enough to equal length. I built it from separate bends of 316 Stainless Steel with 3/4inch headers and a 2 a inch collector. The primary lengths were specified based on the expanded volume of a single cylinder cylinder, going from atmospheric temperature to an exhaust combustion temperature I found on the internet (I have never measured exhaust gas temperatures before, so I think I can be forgiven for consulting the web).

From what I heard it did well on the dyno, and in truth I was sad to see it go; it took a lot of time and effort to make. Eventually the Locost will get one of the same quality, if not better.

Exhaust Manifold 1

Exhaust Manifold 2

Exhaust Manifold 3

Exhaust Manifold 4

3D Printer: Upgrades 2/3

It’s taken me almost two months to get around to writing part 2 of this series, opps! However this is because I have actually been using the printer, and working on a project for a friends rally car (watch this space).

Now where were we… ah yes, the heated bed. In the previous article I explained why a heated bed is a good upgrade for a 3D Printer, especially one that uses high temperature plastics such as ABS. Installing one is easy straightforward, however my little machine required a few modifications along the way.

To convert your printer to use a heated bed first you’re going to need, you guessed it, a heater to heat the bed.

1. Sourcing a Bed Heater

A simple flat Silicone Heated Bed. Easy to install... and ORANGE.

A simple flat Silicone Heated Bed. Easy to install… and ORANGE.

I chose to use a 12V Silicone Bed Heater. These are easy to get from China and come in an array of different sizes to suit your needs. As I write this, doing an ebay search brings up 47 of them from a range of manufacturers.

Truth be told I took this route because its what everyone else does, however there are some major benefits to this style of heater. Firstly they are simple to wire (4 wires, with feedback), they are relatively thin and they can be driven straight from most standard firmware.


Most common printer PCB’s have the ability to drive a 12 or 24 volt heated bed directly, however I opted to use an external power supply. My heater is rated at 350 watts, so at 12 volts it can draw up to 350/12 = 29.1 amps! I was not willing to push that through a thin PCB, no matter how much the manufacturer says its marginally spec’d to that ampage.

An all purpose 12v DC Power Supply. 240v to 12v made easy.

An all purpose 12v DC Power Supply. 240v AC to 12v DC made easy.

In hindsight I probably should have gone for a higher voltage heater and then wouldn’t have had to flow as many amps to achieve my desired bed temperature; especially given the fact my power supply has to drop down from 240 volts! A smaller step would have been more efficient. In fact 110V heated beds are available, their just less common.



2. Wiring the Heater

The heater has four wires, one pair is power/ground for the heater itself and the other pair are attached to a thermister embedded in the heater. The thermister wires were attached directly to the control PCB and this allowed the software to measure the temperature of the bed while printing. The power wires went to the external power supply via an automotive relay (12v 40amp). The relay was switched using the heated bed control off the PCB, which is usually used for driving the bed directly. This giving a lovely 12v output to charge the relay coil and switch on the bed.

Power Supply






I ran the power switch and thermister lines through the case via some two pin connected. Initially I wired theconnectors straight to the aluminium printer chassis but soon realized one of the pins ground through the outer thread! This mean’t I had to print some little top hats to make sure the signals didn’t ground. Printing parts for the printer; it’s 2016.

Case ConnectorsExternal Connections

Working Temperature Feedback





Once this was all wired together surprisingly it worked straight away (well once the above earthing issues were fixed). This gave me closed loop temperature control of the bed, ensure a nice consistent temperature.

3. Fitting the bed

It wasn’t all straightforward. The Chinese manufacturer neglected to give any dimensions when listing the heater, so I had no idea how thick it was going to be. I had a feeling it was likely going to cause issues with the self leveling system, as this requires the bed to bottom out when being installed to get under the sensing probes.

Low and behold once I installed the heater the aluminium bed no longer fit. Fortunately I had my non-heated bed on hand and I used it to print taller probe mounts. Magic.

Raised Probe Towers





So that’s it for this installment. In part 3 i’ll go through building an enclosure and the tricks I’ve learn’t when printing ABS.


A closing thought and something to consider before getting one of these machines. The more I have used the printer, the more I have come to realize that it is as much a piece of workshop machinery as a lathe, mill or welder. It requires maintenance, care and cleaning to remain consistent and usable. In my experience, few people have the patients for this.


Embedded: Tach… Fail?

So I’ve been using the aforementioned tach input for my pneumatic gearshift system with moderate success. However a problem has arisen.

I run a spark cut on upshifts so you can keep the throttle planted and just hit the paddle to get the next gear. This is all well and good but I also need a tach signal during this period to ensure the clutch comes up once the engine speed is correctly synced with the road speed; otherwise you get a massive jolt through the drive train and break differential housings (I’ve broken two!). The target RPM is based on the engine RPM when the shift is requested, the current gear ratio and the target gear ratio; simple.

RPM_{Target} = RPM * \frac{Ratio_{New}}{Ratio_{Old}}

The tach signal I had been using comes out of the CTO pin on the Ford EDIS unit (pin 11, clean tach out) which is a fancy version of what comes out of the IDM pin (pin 2, diagnostic signal). As it turns out this is based on an “EMF Flyback Circuit”, which uses the reverse voltage of the triggering coil when it is grounded (A good reference can be found here).

This is all well and good, but in short: No Spark, No Tach Signal. This is annoying because the ignition unit still knows what the RPM is from its variable reluctance sensor, it just doesn’t output the correct RPM.

To solve this problem I’m going to have to take an RPM reading directly out of the ECU or setup my own hall effect crank speed sensor…

I’ll let you know how I get on.


Analysis: A poor design is a good example.

MeThis is an introduction to a series of posts where I plan to illustrate the importance of analysis in vehicle design; especially before you start construction. Having gone headfirst into a car build without undertaking any proper analysis I have learnt a few lessons along the way! I’m only truly starting to understand how to drive the Locost because I never properly analysed it during the design stage.

The change in mechanical balance under braking, the right springs to use, the effect of ride height change on balance, the amount of cross weight I’d have to run to account for driver mass and more; none of these things I had a grasp of when it came to actually setting up the car.

As I don’t fancy pasting the Locosts credentials online we are going to use another car as the test case. During the final year of my degree, myself and my project group were tasked with designing and building a Formula Student Racecar chassis. If you don’t know what Formula Student is then, in short, it is a racing competition for single seaters designed and built by university students.

For our team it was a design exercise above all other all else, as this is what we were be judge on. I was tasked with designing the suspension geometry and associated components.

We were highly limited by the components we had been given, so the final design was not as good as we had liked. Many late nights and coffee fueled marathons were spent designing this car and I reckon I missed a few things along the way, so without further ado I would like to introduce you to “The Tombstone”:

Tombstone: Chassis



Come back and read future posts where I’ll take you through the suspensions gremlins and iron out any faults that she may potentially have.