Short Side Radius for Sidedraughted Heads

[PersonaGrata]

Hi Guy,
I notice from a few other threads that you use a very small radius on the XE head for good flow. My 4G93 head is quite similar in that the flow in the intake ports is more horizontal than vertical - see diagram below

4G93 DOHC Head

DOHC_Inlet_PortAngles.jpg (71.08 KiB) Viewed 14350 times

From the diagram, how small a radius do you think I could go to? Roughly how small do you go with the XE?


Thanks
Iain

[PersonaGrata]

I've read my own post and decided it was worded badly.

From the diagram, does it look to be of a similarly small radius to the XE after you have modified it?

Or do your XE heads have SSRs that look like they might be smaller?

I can quantify the rough radius in millimeters of the SSR, which I'll try to post tomorrow

Iain

[Guy Croft]

That port has a reasonable downdraft angle but it also has quite a long drop from the port floor (just upstream of the SSR) to the valve. Ports like that can exhibit two markedly different flow regimes: 1) at lower lifts (up to about 5-6mm) the airstream clings to and follows the curvature, and 2) at higher lifts flow separation forces the flow off the SSR completely and it heads straight from the port floor to the center of the valve. The flowtest below (barrel to barrel comparison) shows what I mean. That was a Fiat 20v and getting rid of the shift in flow was by no means easy and if you do it wrong the flow is going to drop like a stone everywhere. I've recently seen another test (eg: http://kansiset.galleria.fi/kuvat/Fiat+ ... at+20v.JPG ) where another guy has worked on that head and apparently not tried to eliminate it.

The valve shape is a clue that this may be such a port and treating it like the XE (one example, there are others) could be a mistake.

Thus it would be unwise for me to advise on this without first examining the valve-in flowtest results for a port in standard trim. Do you have that?

GC

[PersonaGrata]

Yes I do.

This chart shows the standard inlet ports against ones I modified. These values were measured on an early version of my flowbench and therefore cannot be relied upon - at least in terms of absolute values.

Cyl1CFM_1.jpg (37.77 KiB) Viewed 14231 times

Its worth pointing out that the vast majority of reading have been taken at static depressions less than 10inch, then extrapolated. I guess that flow velocity is the most important factor when it comes to a gas making a tight turn - and because my flow rig cannot achieve high velocities at higher lifts, then the different flow regimes might never be seen!

Thanks
Iain McL

[Guy Croft]

Alright, you'll have to do this incrementally and assess the results on your rig. I really mean you'll HAVE to!

Try increasing the SSR curvature as shown, do it on one barrel only and block the other off with its valve for test. Benchmark this by testing as-is and then modified. Do BPF and valve-in tests, that's very important. I do not know what will happen - but's development! Do this with standard valve angles to begin with.


One test of the SSR incidentally is - if you enlarge the port (even just a bit) and the BPF doesn't go up - the SSR is wrong. Don't know whether this will help you any but you may have exp it.


You will know when the SSR is right when the valve-in flow intersects (or near-intersects) with BPF (in the case of a good, unshrouded 16v anyway) at some 'sensible' lift, like the XE type you mentioned. This is only true of real downdraft heads with a good discharge coefficient, ie: the right valve shape and angles, near-perfect SSR with almost zero separation and negligible shrouding. Few 8v heads achieve this, the beautifully downdrafted Lotus TC (below) being a notable exception.

GC

[PersonaGrata]

Hello Guy,
Sorry its taken so long to reply, but its been tough finding the time recently.

I had one cylinder that hadn't yet been modified so I thought I would experiment on that.

I blocked off one valve and set about the other one. Because there is no way on earth that I will ever find the time to do full range measurements I have had to settle for BPF measurements and full lift (10mm) measurements only. I realise that this is less than optimal as we will never be able to work out low and mid range performance, but its Hobsons choice really.

For the first port/valve the table of results is below:
PS. V6.B means BPF figure for Valve 6, V6.10 means figure at 10mm lift for Valve 6

PortValve_6.jpg (86.05 KiB) Viewed 14055 times

The headlines are:
1.Reprofiling the SSR gave significant improvements in both BPF and Full Lift. BPF more than Full Lift in this case.
2.Deshrouding made next to no difference
3.Profiling of the Valve Guide made an appreciable difference
4.Blending the seat into the throat reduced CFM. Throat was opened up slightly here, but I was careful to leave about 2mm of the 75/60(?) degree section remaining.
5.Widening the controlling section gave improvements in both Full Lift and BPF initially, but further widening only improved BPF
6.Deeping the 'pocket', after the guide gave a big improvement in Full Lift, with little or no improvement in BPF.

Thinking I had learned from this set of experiments, I went on to change the order in which tasks were performed for the other valve/port. Results below:

PortValve_5.jpg (91.94 KiB) Viewed 14055 times

Headlines this time:
1.Reprofiling the SSR again gave significant improvements. This time Full Lift was favoured more, but ultimately the figures were smaller overall. I had two attempts at this because the figures were so modest first time around
2.Deeping the pocket did nothing positive this time, perhaps because the port shape wasn't yet compatible?
3.Widening the controlling section had no effect at this point.
4.Profiling of the guide again proved beneficial
5.Blending the seat to the throat had no effect. This time I tried using a ball type stone right on the union between seat and throat to produce a very slight 'venturi' shape
6.Deshrouding was pretty much a waste of time - again.
7.A final further widening of the controlling section saw the BPF jump up with little improvement in Full Lift.

So, taking all the above I've come to these flaky conclusions.
1. Reprofiling the SSR is the best way to improve flow. It is very difficult however to replicate the shape from port to port and so different flow figures are almost guaranteed.
2. Reprofiling the Valve Guide is the second best way to improve flow. (pics later).
3. These heads would almost certainly benefit from wasted valve stems given the experience of the Valve Guides - and the near certainty that most of the gas has to navigate either the stem or the guide (if this were not the case then the guide profiling would not have had so much effect). By the same token, modifying the valves to be more Penny On A Stick-ish would give big improvements given the amount of cross flow, rather than axial flow.
4. I have got no where near making significant improvements to the seat
5. I don't yet understand the interplay between port and throat areas.

More to come.
Iain

[PersonaGrata]

Oops. Last night I spent about 20 minutes adding to this thread and then must have forgotten to press 'Submit' at the end of it all (I like to preview first). As a result I've wasted that time, and now I'll have to do it again! (I was going to have to revisit it anyway because the images were too large)

From the previous post I had decided in my infinite wisdom that there was precious little flow symmetry about the axis of the valve at mid to high lifts, and that is why profiling the guide had so much effect - and why a waisted valve stem would be a good idea. It also seemed worthwhile back-cutting the valve to offer less obstruction to the flow shearing across its back face.

One point to note at this point is that the Proton valves are NOT axially symmetric themselves - the 30 degree section has been cut non-concentric (if I remember I will try to post up an image that shows this clearly), so the flow characteristics of the port change depending on the valve orientation at the time! When I back ground the valve the pattern produced wasn't axially symmetric either, but I cannot figure out if that was because of sloppy machining or the original asymmetry. Anyway, this was just for experimentation and a machine shop will be given the job of finishing the job for all the valves. Here is a prototype valve next to a standard one:

BackGround3VersusStandard1 (1).jpg (99.58 KiB) Viewed 13971 times

Note much reduced 30 degree section on machined valve

3Versus1SideOn (1).jpg (105.1 KiB) Viewed 13971 times

Ultimately using this machined valve gave over 5% improvement at mid lift and almost 2% at full lift so it seems like the right thing to do.

Wondering just what the non-axial flow might look like brought me to set up a little experiment: I got an old Hayfever drug atomiser, filled it with Engineers Blue, blocked off one valve then opened the other at 5mm. The flow bench was started up (pulling about 12inch WG) and the atomiser sprayed twice in the vicinity of the inlet. I tried to keep the atomiser a decent distance away so as not to bias the result.

The next pictures tell the story

No blue on port floor!

CleanPortFloor (1).jpg (82.73 KiB) Viewed 13971 times

Narrow band of flow concentrated around middle of port roof

SplitterLee.jpg (98.59 KiB) Viewed 13971 times

Note shadow of flow in lee of guide and stem - and concentration to inner side of port

StemAndGuideShadow3 (1).jpg (101.02 KiB) Viewed 13971 times

Shadow

StemAndGuideShadow (1).jpg (97.39 KiB) Viewed 13971 times

NO blue at all on SSR side of valve!

P1070101.jpg (84.01 KiB) Viewed 13971 times

Valve in correct axial orientation - but upside down to show flow

UpsideDownValve (1).jpg (138.51 KiB) Viewed 13971 times

So, there is a major concentration of flow along the middle top half of the port, and it can now be seen why profiling the guide had so much effect. The original post was about the SSR radius, but its looking like the contribution of a machined SSR in this engine is just to provide a larger aperture for the almost straight flow into the chamber (for mid to high lifts it looks like I will have to give up on the idea of getting the flow to 'stick' around the SSR). Further hand machining of the guide profile should give extra results, as should further back grinding and stem waisting. (Can you machine a waist onto valves that never had waists??? and if so how much is safe? - stem is 6mm)

What do you think Guy?

Iain

[Guy Croft]

On ports like that the flow can bias towards the roof as the lift increases and only very well downdrafted ports will exhibit flow distributed quite evenly around the port and valve. Some ports exhibit 2 different regimes as I mentioned earlier.

I like your tests very much and given what they can show I'd be for trying at 2/3/5mm too see if the bias is throughout the range.

G

[PersonaGrata]

I think I will try just that. Maybe one at 2mm and one at 3.5 depending on the result of the 2mm test. Whatever the results there will be one value of lift where the flow will transition from 'valve axial' to shear. At that point maybe the modified port will flow less than the standard one?

On the valve front Guy I'm really keen to machine a waist on the valves but am worried about the effect on the structural strength post-machining. You must have done it before over the years; is it safe? and if so how much could a 6mm valve be waisted?

Thanks again
Iain

[Guy Croft]

I don't agree with the practice which is by way of saying I would not undertake it*.

You - however - have complete freedom to explore it! Remember that as the valve goes onto lift - the full form diameter will be exposed to the airstream. You could go down to 5mm I guess but do have a care for the bending load and unsupported length of the valve stem at full lift. I definitely would not do it on a turbo unit, even the in-cylinder pressures can bend the valves.

GC



* in the nicest possible way let me just remark that if I recommend something every man_and_his_dog will be asking me to do it for them. I have to be constantly mindful of 'where to invest the money, which is often better spent on other things. The question (on the phone) usually begins with 'would it be worth (doing such and such)'.. and there is a new breed of fool who, when I say 'no' instantly retorts with 'oh, why's that then?' as if he knows something I don't...

[PersonaGrata]

I'm not considering 'waisting' these valves now - especially considering that the valve seats have been cut and lapped for the valves, and machining might distort the shaft.

In a previous post I mentioned profiled valve guides. Here is a photo of a standard guide

Standard guide in a standard port

Cyl3_standard_inl_aligned_seat.JPG (202.31 KiB) Viewed 13760 times

and here is a photo of the profile guide

View up the port to profiled guide in machined port

P1070138_lower_res.jpg (65.98 KiB) Viewed 13760 times

Here is the same guide from above

View from the chamber onto profiled guide

P1070135_low_res.jpg (61.39 KiB) Viewed 13760 times

It takes about 35-40 minutes per guide with a variety of small files and has given me a fierce shoulder ache. It does give results though, and I'm pretty sure that the support provided is still likely to be sufficient.