Very cool...a bit of lock up in that photo :)
For facts and figures you need to contact the real brains behind this project , David Wedge . All his details are on his website which is DTMPower.com . As well the upgrades mentioned he has a special 4.0 litre 308 2 valve in build as well . Here are some photos of his IDF manifolds in mid development and a sectioned head which was used to help the port design as well as confirming his theory that bigger inlets and smaller exhaust valves would still flow enough air . Image Unavailable, Please Login
Hi Everyone. Thanks for all the comments on the yellow top 2 valve engine. We have been working on this engine on and off with dell550 for quite a few years, and recently acquired commissions have meant that we are having another go at some more power from the 2v, well more torque really (bhp = tq x rpm/5250 don't forget!). Who said that "the 2v heads suffer from poor flow"? Well actually most OEM heads suffer from poor flow, but ours don't. Our current iteration of the 2v Ferrari is doing more again than before, and way more than the standard 4v heads. That said we could get even more flow from the 4v heads, just need a client for some. Our high velocity port strategy makes good torque through a wide rpm range, this is what makes race cars quick, not headline grabbing peak power figures. An example is the F355 engine we did a few years back. The standard engine dynoed as a baseline did 320bhp at 8500rpm, the peak was about 500 rpm wide. We couldn't believe it at the time so we tried another engine, and another dyno... all the same. Gasflowed heads were the only mechanical mod, then correct mapping with Motec M4 ECU. Result was 420bhp peak with a 1500 rpm band above 400 bhp. All the ancillary bits you can see on the yellow engine are still available frmm us, and some have been re-designed and updated recently. The Mercury CDI coils you can see on the cam cover proved to be a bit unreliable and we moved on to some Piaggio ones with far more spark energy. BTW... the original equipment carbs were 40 DCNFs we tried 44 DCNFs which were a little better, but the underlying problem with DCNFs is that the progression is poor and this leads to flat spots as you have to run rich in some areas so that the lean areas are not dangerously lean. The IDFs have much better progression, still not perfect, but massively improved with aux vent mods and the high speed jets you can see in the trumpets. So we now have 3 areas of jet adjustment not just two. AND... we have an even better system in development for carb mandated race series.
Kind thanks for joining us and your hard work with these 2v heads! I agree IDFs do indeed have much better progression, it's just a more flexible carbureter all around but I never knew of a 2v manifold existing! The 2v intake port itself is a horrendous thing, rather than reshape the whole thing which would essentially be reinventing the wheel, I just installed 2mm larger ID and OD valve seats and ported the bowl, throat, and called it a day. I saw about 13% flow increase on average, while lift over .300 to .450 saw much higher gains than that. I also used a nail head style Manley valve with heavy heavy undercuts which attributed to quite a bit as well.
granted the following data is purely CFD output I find it interesting in results. from utilizing the resultant data and inputting that into dyno-sim programs it shows promise, but as always the proof is on the track/road. The stock port slows velocity down at the roof where there is a bubble. too large a valve and to much lift as a similar effect, it's a bell curve as the lift increases the flow does till the flow drops as there is more lift. I also feel that static head flow isn't an accurate predictor of power supported, the dynamic model with wave length and pressure pulses need to be taken into account. that of course gets very laborious and deep into mathematics and thermodynamics. Excellent work btw, and yes it's possible to get really good numbers out of the 2V head, it's just that the 4V is easier . But we work with what we have Image Unavailable, Please Login
+1 great post. The bubble can be alleviated by altering the intake port entry position raising initial entry say about an inch or so. lots and lots of welding and then new cutting. It's able to be done, but not without massive time.
The problem with CFD models is that they are massively wrong. The airflow just doesn't do what the model says it will. Do we really think that the velocity gradient at the mouth of the port is zero, as shown by the constant light blue colour in smg2's picture? I'm very much agree that "static head flow isn't an accurate predictor of power...". We regularly see transonic speeds in the port entry (in the region of 340 m/s, green in the CFD illustration), I doubt we ever see that velocity just after the valve where the pressure drops as the charge pops into the combustion chamber, certainly never down the cylinder wall! Also this particular CFD snapshot has no carb and no intake trumpet. These both have a massive effect on air flow, quite pointless to analyse without them. What most airflow developers do is to improve on what has gone before, but we have found that what has gone before is not that good, so we have started from scratch and found that air actually flows in a way that has not yet been represented by maths. Sure 4v flows better than 2v (but 5v doesn't flow as well as 4v!!! sorry Yamaha, even Ferrari have gone back to 4v since the 430 engine). Our brief in these heads was to comply with race regs. Soon as we get a commission to build a 4v Ferrari... watch this space!
No offense in following intended... Just to clarify, the gradient color map drops to zero but the flow in the model does not. there are over 50 velocity points in the color shift, in order to read what is occurring at any point in the model requires a coordinate point or surface boundary to pull from. The posted map is a more of a visual snapshot. CFD software is very accurate, it just requires knowing how to read the output. Yes more accurate results will be had by taking into account the entire system, which I do -see attached below (positive pressure / dynamic model). All the CFD results are really best viewed in animated form, they are however such massive files that it's not practical to share on a forum. the section plots are a zero depth frame, not a uniform distribution, more on that later. there is a lot going on in the cylinder, massive temperature changes as well, combine temp fluctuations with varying gas densities and a dynamic movement of the piston & valves, pressure waves, incoming, outgoing and the overlap on the cam timing. pressure differentials that effect velocity and flow and the effect it has on gas density, phase transition and shear. I've also included a couple more shots showing the flow trajectories, as seen there is no zero velocity points, and where there are higher then one would expect flows along the cylinder wall it's not uniform distribution, notice the swirl of the incoming flow. I've found the CFD results to be within 2% of testing. I'd love to get my hands on the Ricardo program, I just lack the millions needed Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login
Yeah made it through ok - had to try real hard to get a break from the F40 and the 355Challenge guys . Beating the 355's was hard work and the old GT4 has to be wrung right out to do it , but the F40 was another thing altogether . Once they were sorted there was no catching them . My brother and I spent the best part of a season finishing third and fourth behind a pair of F40's !
Tried to get the club to allow turbos but they nearly had a fit . At the time we were not even allowed to fit 4 valve heads or 3.2 blocks . They even banned fuel injection if your car was not already fitted with it from the factory !! We had to stick with 3.0 litre 2 valve and carbs so that's why we did the IDF conversion . Now anything goes hence the 3.3 conversion and throttle bodies .
Peekaboo !! Here is a pic of my brother and I ( he's in the GTB ) and you can just see my GT4 behind him and the F50 looking for a way through .....this is off the start line at Croft circuit and we hassled the F50 so bad he went off at the first corner . Image Unavailable, Please Login
my point was not that the flow was zero but that the difference in flow velocity across the port was zero, or at least not much, as illustrated by the constant colour of the snapshot. Nearly all the pictures show little velocity difference in any cross section at a normal to the flow direction. This is not what we find. We find a huge difference in the velocity from the port wall to the centre line. We exploit this in our port design to get far more gross flow volume than conventional methods can access. If you continue to believe that CFD is correct you will only find small incremental improvements in engine performance. Using what we have discovered about port flow patterns in addition to large increases in intake system flow and velocity, we can now for example; increase flow through mandated NA intake restrictors, and produce turbo heads that have zero boost lag without having to set fire to the exhaust system! BTW I really love your second law of Murphy
