I bought the series of interviews on CD and used to listen to them in the car on the way to work (stuck in traffic)!
Everyone thought I was mad.
But I will look at a few more. The Reher Morrison Tech pages are very interesting.
By way of progress, I now have a model of the cylinder pressures (vacuum) that drives the port flow. Now I need to work out what the relative cylinder timing is and feed that into the simulation program. The peak vacuum is very high just before the valve opens and has no real relation to what the vacuum might be at full lift. So itāll be interesting to see what impact that makes with low lift geometry of the valve seat etc.
Iāve attached here a graph of the inlet āpressureā simulation. Superimposed is the inlet valve lift curve so you can reference pressure against valve events. I have not sanity checked this yet. But atmospheric pressure is 14.7 psi ā¦ around the middle 15 line on the graph. Anything below this, and its sucking in air/fuel mix. Anything above this, and its resisting.
In terms of what might be more normal on a flow bench, the peaks and troughs equate to about +/-100 inH2O. I think most flow benches run at 28inH2O? So the peaks are some 4 times what a steady state flow bench might simulate. The peak suction event occurs around max piston acceleration down the bore, I think (about 425Ā° - I should graph that as well). What is curious to me is that there are 3 peaks per cycle (720Ā°). I āthinkā the big peak (600Ā°) to the right of the inlet valve closing is the āair piling upā against a closing valve and rising piston. The strange kink at around 350Ā° will be exhaust valve overlap.
I guess none of this is particularly meaningful without more data. I havenāt recorded it with the graph, but I think this was at 6500 rpm. And the inlets are about 280-300mm in length. (Need to double check that). My objective was really just to try to see the kinds of variation that could be expected. The real point is that peaks of 4 x steady state raise questions about the relevance of a steady state flow reading.
What would be interesting would be whether after a number of engine cycles, this graph settles, or whether peaks and troughs move across the graph. Put another way, can you predict or demonstrate reversion will occur with a certain camshaft at a certain rpm.
Similarly, you should be able to use it decide the peak time to inject fuel at lower rpm when the duty cycle is low.
Hi Mark
The steady state flow benches provide a useful benchmark for the efficiency of the inlet design. Make a change and if the flow at the same pressure increases then you have improved the design. This is all to do with geometry and surface finish. As you discovered,ā¦a small radius vs. large radius or double angled seatsā¦all contribute to the effectiveness of the design to maximise flow at any static pressure. The other effect is the dynamic effects inherent in the moving fluidā¦but an efficient design will perform better in the presence of whatever static pressure is available due to the dynamic effects. So it is a good way to benchmark an existing system. On top of this you can make improvements with your CFD analysis to optimise the dynamic effects. CFD is a relatively new tool made available by gamers pushing the performance of desktop computers!!! So not available to Engineers of yesteryearā¦unless you had access to some serious water cooled computing rooms!!! ā¦but what a great tool it is!!!
Regards
Matt
Hi Mark
The tube starts at the bell mouth end at 38mm and increases in size toward the filter can.
I looks like the other way in your model??..or is it the beersies in the sun after work???
Regards
If you are interested in making big power from the V12, you may be interested in the book " Basics of the Internal Combustion Engine" by Richard Stone (not sure about the first name). It had a description of the development of the V12 engine and also described how offshore powerboat racing used thick metal plates to raise the heads from the block to allow for stroking the engine. They also had to increase the timing chain length.
That was the good news. The bad news was that drivers of Cummins diesels in the 30s had to pit stop every hour for a coffee refill and more biscuits and gravyā¦
I had an accelerometer that would measure take off G force and 0-60 times. I used it everytime I made a change. The air inlet opening made the largest jump as far as performance goes. Almost a full .5 seconds of of the 0-60 time. Overall, I took the car from a stock 9+ second time to a 6 second time for 0-60. Later I put a 6 speed manual but did not have the measuring device any more.
I have been a bit worried about the impact of the big valve cut outs in the piston crown, and wondering whether their geometry gives an ability to shape the swirl and tumble, or whether there is bad shadowing.
Anyway, I thought Iād try to run an analysis with 5mm of valve lift. It turns out that the piston is almost BDC by the time the valve is open 5mm, so there is no issue with the pockets shadowing the valve when it opens. (The overlap of valve and piston actually occurs during valve closing).
The attached image are the flow lines with the colours showing vorticity (amount of local turbulence).
This analysis took my computer about 1hr to run (and included the full inlet trumpet and ports). Of course the danger is to believe what you see! But anyway, I might try a few tweaks to see how it might impact this. Iāll also try to refine the model to simulate the predicted vacuum at the various lift points and see how it āprogressesā through the stroke. In theory, I think the computer can run solutions for multiple piston positions and ārun a movieā but I suspect thatād take more computer power than I have available.
To put things in a little perspective, assuming the port velocity is in the order of 300-400ft/s (100 m/s) then there is nothing very āslow motionā about what is happening. The piston is probably taking about 4ms (4/1000th) of a second to descend. And the inertia of the column of air in the port should keep it flowing while the piston rises and starts compressing the mixture.
It also starts to become apparent why spark timing is so critical - it wouldnāt take much spark scatter and youād have the spark delayed or advanced by a critical amount. Bit of wobble or play in the distributor bearings? Wouldnāt take much. Good reason to go all electronic. An engineer I know builds formula ford engines, and one of their big innovations was to reduce the spark scatter caused by the oil pump drive cog ā¦ but I digress. One of the advances on the Chevy LT-1 was to reduce spark scatter from +/- 2.8Ā° to +/- 0.5Ā° . Be interesting to know what it is on the Opus ignition.
only time will tell if all this makes a substantial performance change(increase)!
i think the high ridge of piston edges should be as tight as possible(without touching head), so as to produce some squeeze just when combustion pressure is highest ,pushing out toward center of bowl?
for sure an all electronic ignition system , actually building with this much reengineering would be redundent not to use electronic!
i dont think anyone knows what the spark scatter is in a good running1970s Jag V12.
i have modified some GM distributor shaft, removing the bushes, and installing ball bearing parts.
of course never knew if it helped, run good both ways!
All interesting stuff and lots of work here, Iāll be interested to see if you find anything new. The late Group 44 heads have 42mm inlets and I think are superior heads top the TWR offerings.
Not too sure why you have the āstepā in the trumpets, looking like the XJR13 efforts to me not like 44 or TWRā¦
If you want dimensions of original stuff 44 or TWR just let me know.
