4G93p Porting Progress

[PersonaGrata]

I've thought about where to put this post - it being mostly about progress on my porting efforts on the Proton 4G93p, but since it ends in a question I guess that this is the best place for it!

This is the second 4G93p head I've ported. The first was enlarged right through the port, and flowed pretty damn well, but didn't seem to produce the goods low down the rpm range with the standard cooking cam. So the effort this time was on achieving good results but not at the expense port volume and velocity. All the work so far has been on a scrap head with the intention of transferring the 'best practice' to my good head with 12k miles on it.

From the first head, it was obvious that profiling the valve guide produced good results, so this was carried over to the scrap head. Here's a standard guide:

FT003B-HP_Maxxed_Diam_sized.jpg (93.71 KiB) Viewed 11266 times

And here is a profiled one:

Profiled_guide.jpg (67.91 KiB) Viewed 11266 times

Experimentation with different valve seat profiles and valve shapes had been completely fruitless. A number of different profiles were tried, all SERDI compatible cutter with 4 angle profiles and maxxed out to the full 33mm of the inlet valves. In addition to that I'd tried manually produced seat profiles using a Neway. Here are a selection the valves tried:

Valves.jpg (138.45 KiB) Viewed 11266 times

To cut a very long story short, none of it worked. (You did predict that the valve shape couldn't be bettered Guy)

Something that didn't proved fruitless though was reshaping the SSR. So, guided by the flow bench I reshaped and smoothed until I got to this position:

SinglePortFlow_sized.jpg (43.78 KiB) Viewed 11266 times

Just as a footnote to this picture: Its pretty easy to machine a profile onto the SSR that 'looks' like the finished article. But without the flowbench its absolutely impossible to know!. Its a bit like navigating a car to a destination on a cloudy day by just using the sun - but every so often you get out of the car to use a GPS to find out where you actually are. As it result it takes maybe 5 or 6 iterations before you get to the BPF you want.

So the increase in the in the range of 19.2% max, with no increase in the port mouth size, parallel section size or valve seat diameter. So for the vast majority of the induction tract I can fairly confidently say that velocities can only increase. I also haven't had to perform any deshrouding work in the combustion chamber so static CR remains unaffected. Finally, because the seat does not need to change and nor does the valve, there has been no adverse affect on low to mid range lifts.

So overall, I'm extremely pleased really.

The max flow is just under 70CFM for a single port at 10mm lift and 10 inch depression, so I'm guessing that the cylinder flow will be between 135 - 138 CFM for the same conditions. I'm not going any further than this, in that all other mods from this point onwards will reduce velocity in other parts of the port.

So, finally, the question is this (and apologies for its vagueness) : Do you think that this flowrate will enable a good BHP/litre? - given the right induction and exhaust, cam, CR etc? I've got this target in my head of 180BHP at 7500rpm or less, and if this isn't achievable with 135+CFM then I will have to consider further work.

Iain McL

[Guy Croft]

What is the engine cubic capacity?

Sorry to be so clueless!


G

[Nobby]

Standard 4g93 engines are 1834cc.

[PersonaGrata]

It is completely standard in terms of cc.

[Guy Croft]

With cams to suit valve-in characteristics plus aggressive lift integral and springs that can take it, plus the highest CR you can get (up to 13/1) and the right header and calibration - that flow will take you way past your suggested bhp threshold.
FWIW one of my 8V Fiat TC could meet that bhp even in quite a mild state of tune with typically only 108cfm BPF @ 10".

Well done - great work and nice to read about it.

G

[PersonaGrata]

Thats music to my ears! Thanks

To all that might be interested in the spec of this engine - this'll is it:

1834 DOHC 4G93
Home gas flowed head - approx 135CFM at 10mm / 10inch
ZX9R 'C' carbs on a bespoke Bogg Bros manifold
Ford wasted spark / EDIS / 36-1 tooth system driven by a Megajolt
The highest static CR that I can get away with
Hotbits 4-1 manifold into a straight-through 2 1/4 inch stainless system (no cat)
Cam duration yet to be decided but will be a regrind Kent profile approx 295degrees or thereabouts
Adjustable cam pulleys
Hydraulic lash adjusters.
Uprated rod bolts
Fully synthetic oil

Standard Clutch
Standard 'box

The only thing left to buy is the cam - but there is still a lot of work to do before its running.

[PersonaGrata]

As can be seen from previous posts, the valve guides in my 4G93 heads have been carefully profiled to reduce flow loss in the important port roof area. Given the improvement seen in flow due to this work, I decided to experiment with the complete removal of the valve guide within a port in my scrap head - I think most people call this 'shaving' the guide.

To be completely honest, this was just for interest sake because I wouldn't feel comfortable running an engine with shaved guides on the track.

Here are the results:
Test 1: BPF of single port with the profiled guide in place : 73.3 CFM 10"/10mm
Test 2: As above but with a valve stem in place too - no valve head : 72.19 CFM
Test 3: BPF of the same port but with the valve guide completely shaved : 73.3 CFM
Test 4: As above but with a valve stem in place - no valve head : 70.11 CFM

So the completely bare port flow is totally unaffected by the presence of the profiled guide - from the point of view of flow it doesn't matter whether it is there at all.

When you introduce a valve stem in there its a different story though - you end up loosing just over 1CFM. And this makes sense really, given that the profiled guide was all but invisible to the flow. When you remove that guide the overall effect is to increase the amount of stem visible to the flow - and at the very part of the port where it is most critical!

So shaving the guides in this engine at least will reduce the flow and make the engine less reliable. Not a winning combination really, and one I will be avoiding in my engine.

Guy, do you wan't me to continue this on another thread? Instead of the Q&A? I'm conscious that this section might not be the ideal home for what is in effect a bit of a 'blog'

Iain

[Guy Croft]

I'll take a detailed look later but I like your thread and there is no reason to move it form GC Q&A but thanks for asking,

G

[craigb]

Iain

Theres some encouraging results there ! Have you determined the valve to piston clearance as yet on these engines , as I'm getting various bits of infromation as to what skim a std head will take before clearance becomes a problem .

The duration of your proposed cam sounds interesting ,?

Guy, is a straight corrolation that with a greater duration cam , benefits are greater in a plenum fed engine when the CR is rasied to accomodate the effect of the increased duration and overlap

[PersonaGrata]

I performed a test last night, comparing my modified ports against the standard ports but using both valves instead the single valve testing performed thus far. To say that the results were a disapointment is the understatement of the year.

If I get the chance then I’ll try to post up the chart at some point but suffice to say for now that there was virtually no improvement till we got up to the high lifts, and even then the flow never got beyond 4-5% on top of standard. So improvements of 15-19% seen at the single port level had magically transformed into two tenths of a nats whatsit. [Edit: Here is the chart]

4G93FullCyl.jpg (32.09 KiB) Viewed 10694 times

The natural temptation was to give up but then it occurred to me that I’m not comparing apples with apples.

The problem is this as I see it: when testing at the port level my flow rig can generate quite respectable static pressure differentials (I test at just over 12inch and interpolate down to 10inch), whereas the rig can only support just over 4inch when testing at the cylinder level which is then extrapolated up to 10inch). This would be OK for a straight pipe but not for a curved port where flow is partly a function of velocity.

My work so far has concentrated on the SSR and the valve guide and it seems to me that these features are only important for high velocity gas, especially for sidedraughted ports like the 4G93. The critical aspect of an SSR profile is the range of velocities that gas can traverse it without loosing ‘adhesion’, and the critical aspect of the valve guide profile is the drag caused when the flow runs across it towards the far side of the valve. So the improvements in these areas can only really be seen when the velocity is sufficient to cause problems in the standard port.

I don’t think 4.36inches static depression creates a high enough velocity to cause significant problems for the standard port hence the small improvements apparent under these conditions.

When you think about it there is some logic behind this theory; as flow rates decrease towared zero the velocity vector of the gases exiting the port will straighten up to be axial to the valve stem. As the velocities increase that velocity vector will approach (but never meet) the angle of the port axis. The modifications I have made specialise the port shape to suite the latter set of conditions.

[Edit: Here is a schematic showing my interpretation of flow in the standard ports at low rates, then flow at high rates in the same port. ]

StandardLowFlow.jpg (15.33 KiB) Viewed 10694 times

StandardHighFlow.jpg (15.69 KiB) Viewed 10694 times

[And here is a schematic explaining how the modification help with the high flow situation]

ModdedHighFlow.jpg (16.14 KiB) Viewed 10694 times

[Edit: in all pictures the red arrow represents the overall flow vector as the gas enters the cylinder]


If this theory is correct then the improvement seen during single port testing should decrease as the flow/velocity decreases. Well, I can actually test that from perhaps 2 inch to 14 inch static at full lift. That’s the next job then!

Craig, embarrasingly, the cam spec is nothing than a single number plucked out the air right now. At some point this’ll need proper analysis – in tandem with valve to piston clearances. But I haven’t got around to that yet.

[Guy Croft]

Just wanted to say how sorry I am that I have not been able to remark or even follow this thread in recent weeks.

Hopefully soon and in the meantime keep up that smashing dev work and great posting,

G

[PersonaGrata]

No probs Guy, I know what its like to run your own business so understand how little time you have.

I haven't followed this thread up as yet because the experimentation is not yet complete - on account of my manometers all being frozen solid. This has not all been bad however, and has allowed me to focus on other tasks that would otherwise be neglected. Hopefully, I will be able to resume testing towards the end of this week when I get back to Lancashire.

Iain

[Guy Croft]

frozen?

You can add antifreeze, I use Unipart, it's sort of fluorescent & shows up better than water too.

G

[PersonaGrata]

The weather improved!

Its been a while I know, but there was a lot of prep work to carry out on the car for a recent trackday at Oulton. The truth behind the strange results has turned out to be quite involved and complicated - and a indirect result of the Heath Robinson type approach my rig employs.

The way I calibrate the rig is by placing a Sharp Edge Orifice plate over the adaptor tube, running the rig at a variety of static depression whilst recording the values registered by the venturi (and these values are some proportional function of actual flow). At each depression the performance of the SEO can be competently predicted by the calculators freely availabe at Efunda.

This allows me to chart out venturi measurements against Efunda/Actual flow.

But, and this is where the problem roots lie, the chart is a Scatter chart of sampled points rather than a continuous function. I need a continuous function and so use Excel's Solver tool to produce a quadratic interpolation of the scatter plot.


Which is all well and good and here is an example - its not a bad fit really...........

Quadratic Interpolation - in range

QuadraticInterp_26mm_SEO.jpg (29.22 KiB) Viewed 10188 times

This approach works well UNTIL you use a value beyond the range of the data used to generate the coefficients in the first place. The thing is that the squared term has a negative coefficient and this is the term that dominates with larger X values. The result is that the curve takes an increasingly tight trajectory then dives inexorably for the X axis:

Extrapolation!

Extrapolation_26mm_SEO.jpg (29.25 KiB) Viewed 10188 times

In fact the problem is really that I've used an interpolation as an extrapolation

So that why. And now I know, I can breathe a bit easier and rest on the laurels of up to 18-20% improvements that were measured per port, rather than the single figure improvements measured at the 'cylinder' level

[PersonaGrata]

I got some cams reground - even though billet would be better I just can't afford the extra expense. Kent did these for a mere 215+VAT which was excellent I thought

Del from Kent asked me 'what do you want? Power or Torque?' I said 'Torque - at high RPM...'

So he concentrated on duration rather than max lift, and because I was sticking with the hydraulic lash adjusters (at least for the time being) he made sure that valve train peak loads were kept under control.

In the end he managed to squeeze 304 degrees duration out of the cam (measured at 4 thou clearance) and here is the new profile against the standard inlet profile

Kent304.jpg (33.45 KiB) Viewed 10133 times

Because a sizeable amount of material had been taken off the BCD I had to shim up the lash adjusters from underneath (being careful that didn't lead to oilway mis-alignment), but it all worked out OK in the end.

I think that the final profile actually fits really well with the modded cylinder head because the 'low lift' flow has increased massively due to the increased duration, and the high lift flow is enhanced by the cylinder head work

The exhaust cam has exactly the same profile as the inlet, which might not be the ideal situation as it seems that most 16v setups have shorter exhaust duration than inlet. I think if I was do get a second set ground at any time it would less duration with more max lift, less exhaust duration than inlet and solid lash adjusters.