Carburetors and inlet port diameters

[harrynaerger]

Morning All. Following on from my last post about bearings, the engine I am buying from Mike Smailes as a back up for our racing Jupiter is presently having the cylinder heads engineered. Any thoughts on the best carbs to use and what we should aim for in terms of inlet port diameter? Our current engine has Webers. I've not had a chance to look at our inlet diameters. I'm told by Mike Smailes that Richard Gane uses Zeniths (he thinks) and again we are not sure on inlet port diameters. Unfortunately Richards mechanic Dave is poorly at the moment so I can't pick his brains. Many thanks for your help.

[Keith Clements]

The main issue will be getting through scrutineering.
Usually carbs and port diameter have to be standard.
I use DelLorto as did Mike. When racing I fitted a pair of hyperbolic throats after a run on the dyno showed large loss through air filter.
There are detailed pics on JT of Richard's carbs. He does use Zeniths but has altered inlet size and length.

[Chris Spencer]

Pretty sure that Julian Crossley runs a single webber carb centrally mounted to a custom inlet manifold on his Javelin - with a relocated electric water pump - in fact just found the image that I took in the pits at Goodwood Revival 2019 which confirms that - Julian obviously got it accepted through racing regs - not sure how but you could surely argue a period racing mod - unsure of what was fitted to the Le mans / period competition Jupiter's

20190915_135644.jpg

Jupiter's in race format are very rare I doubt that race scrutineers hold a in depth knowledge of the carburation / fuel / manifold systems & period race modifications - a conversation with Julian would qualify if he has hit any objections with his mods

[Keith Clements]

At Goodwood some others in Julian's class had to un-modify their cars prior to racing, notably on the ignition. Similarly at Le Mans Richard had to do some difficult changes before being accepted.
Whilst Jowetts can be easier to get through, the Jowett racing team in the 90's had to show evidence to get mods on the FIA papers. Mine was standard in 1989 for rallying and my FIA papers only show other final drive ratios homolagated, larger bore and weight reduction to 15cwt, but I think overdrive and some carbs were later homolagated. Tyres and rim sizes were the main issue which now seem to be easier.
Julian has done extensive suspension mods.
Goodwood seem to pay little attention to originality or period mods so are not a good guide, although they do like cars to look good and perform well. This is a shame as cars are modified beyond the point of no return and are nothing like what was raced in period.
Other series will have their own regs and the scrutineers and other competitors will get to know the class and any deviation from the regs.
In the end you need a back up plan, so if you do mod heads and carb then you carry original spares to change back on scrutineer failure.

[harrynaerger]

Very many thanks for the replies. I think the only time we would run into trouble with scrutineering would be Classic Le Mans (we had only planned to take this as a backup motor...) as none of the other racing we've done requires an HTP and as Chris rightly points out we are such a rareity that I'm sure most scrutineers won't know any better!It also helps that we are always at the back of the field. I guess it might be more tricky if we were ever to get near the podium!

[Keith Clements]

What John says is not something I would advertise. Round ports should be better from a fluid dynamics point of view. I think the issue is finding enough metal to get the required area round hole.
There was some research done by Frank Wooley at fitting a twin Weber and separate ports, rather than the siamese.
The unusual firing order makes the fluid dynamic modelling complex with siamese ports and balance pipe all involved.
I do not know if there is any quantative evidence for increasing the inlet track length (or even for the partioning of exhaust manifolds and pipe tuning). The sort of revs (7500rpm) being used in racing is way above the design compromise so breathing is going to be an issue.
The use of tuned camshafts by Dave on Richard's car also puts more demands on the carb.
There was one Javelin fitted with water injection which on a trip up to a Scottish rally I experienced the advantages of when a head gasket went.

[David Kemp]

I am going to use a single SU HIF44 on my car. Any suggestions on manifold design?
I also intend to use an electric water pump, again any suggestions?
Would the owner of the Javelin pictured please give me photos of the conversion & or Advice.

[PJGD]

The standard inlet port diameter is 31.75 mm diameter. There is a rule of thumb observed across multiple disparate engines that at maximum power the flow velocity through the port will be close to 80 meters/second. In the Jupiter, that works out to very close to 4,500 rev/min. and sure enough if you look at the standard Jupiter power curve, that is where maximum power occurs - see graph:

Jupiter Performance Curves.png

As for the question of optimum manifolding, the fact that due to the firing order, each bank sees two induction events close together followed by nothing for two strokes is not ideal, and I am pretty sure that it would be better if the induction strokes were 360 crank degrees apart (as they are on the Bradford for example). This would be possible if we had individual inlet ports on each cylinder head and not the Weslake configured siamese ports that we have. With individual ports we could connect cylinders 1 & 2, and 3 & 4 with cross-over manifolds and a single central carburettor. Better still would be to arrange for the manifolds to be tuned using a Helmholtz resonator to maximize volumetric efficiency at whatever speed you want peak torque to be. Generally, the longer the inlet tract, the lower the peak torque speed will be, and our very short tract likely accounts for the relatively high peak torque speed of 3,000 rev/min in the standard engine.

To get more power from the engine, one either needs to increase engine speed without degrading volumetric efficiency, i.e. breathing, or increase volumetric efficiency at the current speed (or of course increase both). To increase either or both requires a serious upgrade to the valve train for the Jupiter; things such as lighter pushrods, stiffer rocker shaft, perhaps hollow [lighter] valves, computer designed cam profiles, etc. With this done, one can then take your new estimated maximum power engine speed and use the 80 m/sec port velocity value to work out what the new larger port diameter should now be.

In period, the Le Mans Jupiters used regular twin Zeniths as can be seen here: https://www.motorsportimages.com/photos ... jowett&p=2

[ajackson]

Fascinating Philip, thanks for posting this.
I'd be interested if you can add the intermediate steps of the calculation..
to aid my understanding.

All the best
Andy

[PJGD]

Hi Andy,
I think that I am guilty of being too glib about this topic which is in fact a lot more complex than it first appears.
A rough approximation for inlet port flow velocity can be obtained from (bore diameter^2 [mm] / inlet port diameter^2 [mm]) x mean piston speed [m/s] and this is plotted for the Jupiter below. Mean piston speed is RPM x 2 x stroke. The two curves are one for the inlet port (valve seat) diameter, and the other is for the port diameter at the cylinder head/carburettor joint. Ideally, the latter should be a little larger than the former, but in fact they are back to front (valve seat is Ø31.75 mm, and port is Ø30 mm).

Jupiter Inlet Port Flow.png

What this chart suggests is that if you improve the valve train so that the engine is now happy to run to higher speeds, you will then want to increase the port diameters to bring the velocity back to the 80 or 85 m/s range by re-running the calculation.

[ajackson]

Thank you Philip, somewhat clearer!
Andy

[Keith Clements]

As Philip says the calculation is considerably more complex but simplification helps understand the parameters. The limiting factor in my opinion is the valve opening. I spent a lot of time getting the profile of the valve seat to give the best flow and the valve lift to be the best. The cam profile is key to this. There is also a considerable effect on the flow by tuning of the exhaust and inlet tracts to give a scavenge and thrust pulse to the flow at the critical time when the ports are open. As the factory tuning notes show combustion chamber shaping also assists charge quantity.

[PJGD]

My graph 3 posts above is not correct because I picked up the incorrect values - notice that the lines do not converge on zero; what a give-away! This graph is better:

Jupiter Port Flow Graph.png

[PJGD]

Not directly related to the issue at hand, but of interest nevertheless, here is a graph of various competitive engine performance curves dating from the early 2000's, i.e. around 20 years ago, collected by (and confidential to) General Motors. The curves are Brake Mean Effective Pressure [BMEP] against engine speed, and the reason for plotting this in terms of BMEP instead of BHP or Torque is that it provides a direct comparison of the level of development between engines of varying displacement and performance. The higher the BMEP, then the higher the amount of development and very likely the sophistication that has gone into the engine. That explains why the 4-valve/cylinder engines, many with variable cam timing and higher volumetric efficiency, show higher BMEP values.

Note that these emission compliant engines have all been run with current petrol of 91 Research Octane Number [RON], and I assume that they are all naturally aspirated, not boosted. Onto this graph, I have plotted the BMEP curve for the standard Jupiter which of course was developed by a company with very limited resources on petrol of an estimated 80 RON which dictated a maximum compression ratio of around 8:1 as against something like 10:1 for the recent engines. What this shows is that with better fuel and some similar technologies such as electronic mapped ignition and petrol injection, the Jupiter would likely be very competitive with other relatively modern 2-valve engines.

Jupiter Comparative Performance.png