Swirl in a 4 valve per cyl head

Question and answer direct with GC. Competition engines and 'live' projects only. Good photos to illustrate your post are expected.
PersonaGrata
Posts: 52
Joined: October 8th, 2009, 7:42 pm

Swirl in a 4 valve per cyl head

Post by PersonaGrata » September 2nd, 2010, 4:27 pm

Guy,
I was a little bit dissapointed in the performance of my first Mitsubishi/Proton cylinder - specifically at lower rpm where it seemed to lack torque. For this reason I am trying again, but with a different emphasis.

This time I want to keep the port size more or less standard (only blending where neccessary) and only work the valve, the valve seat and the SSR. Improving low to mid range lift flow without affecting velocity in the ports and controlling sections is the idea. This is so as to keep the low rpm performance good.

I've also had this mad idea about increasing swirl by improving only one side of each pair of inlet ports - again with the intention of improving/maintaining the low rpm performance. I understand that 4v heads have very weak swirl in comparison to 2v for obvious reasons and this hurts low rpm performance.

On the basis that everything I ever think up has been tried millions of times before by millions of other people, I wonder could you give me your opinion on this approach?

Regards
Iain McL

Guy Croft
Site Admin
Posts: 5032
Joined: June 18th, 2006, 10:31 am
Location: Lincoln, UK
Contact:

Re: Swirl in a 4 valve per cyl head

Post by Guy Croft » September 2nd, 2010, 4:38 pm

I want some photos please,

G

PersonaGrata
Posts: 52
Joined: October 8th, 2009, 7:42 pm

Re: Swirl in a 4 valve per cyl head

Post by PersonaGrata » September 2nd, 2010, 8:56 pm

02092010179 (1).jpg
Strategies for increasing swirl in a 4v head
02092010179 (1).jpg (159.88 KiB) Viewed 4023 times
Whats perhaps not fully explained by the picture is the machining that I'm thinking of carrying out.

First I've bought some serdi compatible seat cutter profiles from Goodsons (3 in total) which my local engine builder can employ in his machine. The selection of the profiles is a story in itself, - they are varied, but not too much. Ultimately, the proof will be in the pudding. All profiles will maximise throat cross section, but not at the expense of profile. I've worked out the the throat diameter to valve diameter ratio will be about 85%

He is going to cut one inlet seat per cylinder such that the outer diameter of the 45 degree section will match the diameter of the valves (33mm). The seat will be from about 1mm to 1.4mm depending on the cutter. The valve will be back cut to reduce the 45 degree section to a value more in keeping with the seat width.

This is the standard seat and valve profile
StandardSeatAndValveProfile.jpg
StandardSeatAndValveProfile.jpg (14.31 KiB) Viewed 4023 times
And this is the proposed:
ModdedSeatAndValveProfile.jpg
ModdedSeatAndValveProfile.jpg (13.07 KiB) Viewed 4023 times
There will be a couple of different back cut angles used just for good measure.

All this work is on a scrap head - its just for testing out different methods, and the winning combination will be chosen after lengthy flow testing.

Ultimately, when I've got the best combination I will only get one inlet seat and valve machined per cylinder and leave the other standard. Neither ports will be machined unless it is required for blending in the throat cut. I might however machine the splitter so as to bias it to one side, and again favour the machined, high flow side.

Overall the cylinders flow figures will be better than standard (but not as good as they could be). However swirl will be a lot better than standard, so that's the trade off.

TomLouwrier
Posts: 331
Joined: July 28th, 2010, 4:09 pm
Location: Leiden, The Netherlands
Contact:

Re: Swirl in a 4 valve per cyl head

Post by TomLouwrier » September 2nd, 2010, 9:38 pm

hi Iain,

I have no direct experience with this approach, but it seems very much like the PolyQuad design as developed -and patented!- some 10 years ago by David Vizard. He wrote a fairly comprehensive article on it on a forum based in the USA, describing the thoughts behind the concept and the results from a number of heads he built. Both flowbench and dyno.

It sounded plausible to me. Fact is that several years ago some engines did have a secondary throttle valve in one of the ports, so that they would act like a '1 inlet - 2 exhaust' on low revs and switch over to the normal '2 inlet - 2 exhaust' behaviour at higher revs and outputs. This of course to improve low end torque (high air speed and swirl) and high end power (large flow). Very much like variable valve timing stuff does on most engines today. Probably cheaper and more reliable in mass-production.

Guy, can I post the link to that site / article directly?

Off topic and no offence meant: interesting to see the sort of discussions going on in the 'big iron v8'-world, regarding '2 versus 4 valve heads' and 'OHC versus OHV'. Stuff that was not built by any factory in Europe or Asia since the mid-eighties. Also shows how brilliant the F/L DOHC engine was in 1966.


regards
Tom
GC_29

Guy Croft
Site Admin
Posts: 5032
Joined: June 18th, 2006, 10:31 am
Location: Lincoln, UK
Contact:

Re: Swirl in a 4 valve per cyl head

Post by Guy Croft » September 3rd, 2010, 9:57 am

1. Thanks for the offer of the link to Vizard but - no!

2. I don't think the 'new' valve angles will work as well as the old, reasons explained in my new book, small valves (in all my exp) don't respond well to multi-angles in the throat region.

3.The port offset: does it exist on all 4 ports?

You need to do some bare-port flowtesting and let me know the results before considering any alterations at all.

I don't know the port configuration, assume semi-sidedraft? You should publish some more photos.

G

1NRO
Posts: 123
Joined: June 22nd, 2006, 5:46 pm
Location: Carlisle
Contact:

Re: Swirl in a 4 valve per cyl head

Post by 1NRO » September 3rd, 2010, 10:16 am

I am no expert and certainly don't know about the Proton cylinderhead but a trait in the Fiat 16v head I've noticed which I think assists in some swirl is the curve/bias in the inlet ports, one of the first things people loose when hacking at heads but for me I try to preserve this feature or even sweeten it a little. A touch of swirl to my mind complements the big feature in 4 valve heads of tumbling the charge over on itself. Just me looking with big eyes and not as I say with expert eyes.

Nik
259

Guy Croft
Site Admin
Posts: 5032
Joined: June 18th, 2006, 10:31 am
Location: Lincoln, UK
Contact:

Re: Swirl in a 4 valve per cyl head

Post by Guy Croft » September 3rd, 2010, 11:21 am

Hang on.

It is important to differentiate between changes of port axis that exist naturally as a function of trying to get a large port area to converge to two valve throats that must fit_the_cylinder and ports that are designed with a deliberate offset to generate swirl as the charge enters the cylinder.

A port setup that directs the charge straight to the center of the cylinder does not have induction swirl. A port where the axis is offset (either in a straight line or via curvature) will create induction swirl (it's called a swirl offset) and the X19 head is an example as is the Peugeot 106 Rallye or Ford CVH. That swirl offset naturally causes a reduction in flow but this loss may well be (can be) offset by faster and more complete burn. This is one reason the SOHC Fiat can produce fantastic power with only 28 deg of peak advance whereas the TC needs 34/35 deg. For induction swirl to be effective the charge must be forced to rotate around the circumference of the cylinder during filling - which means some pretty aggressive offsetting.

Does it matter if there is no offset? Certainly not - you can improve the burn rate by other means eg: high compression - more densely compacted charge burns quicker. Should you induce it if it's not there? I wouldn't.

The 16v Fiat (Tipo/Integrale type) has no inbuilt swirl but like many heads with a long splitter it does generate horseshoe vortices in the barrels and they (in my view) no doubt do something to help the charge homegeneity. It's easy to get the heads flow over 150cfm @ 10" (enough for over 270bhp atmo for sure) but big-port types like Hondas flow close to that or more with no mods at all (ie: 'out of the box') and despite flowing less BPF than some more modern heads with comparable valve sizes and stubby splitter - on the 16v Fiat head of that genre the slightly lower flow is definitely no impediment to power. A 16v 1600 conversion Fiat with head flowing only 143 cfm will can definitely match or better a 1600 Honda Vtec head flowing way more than that and you have to ask yourself why.

As for tumbled air motion the only head I have worked on that might conceivably have had that was the Rover V6 and because of the losses in intake that can result (not to mention impact on valve-in flow distribution) it's not a concept I take overly seriously on SI units.

G
Attachments
CVH offset.jpg
CVH offset.jpg (111.72 KiB) Viewed 3969 times

1NRO
Posts: 123
Joined: June 22nd, 2006, 5:46 pm
Location: Carlisle
Contact:

Re: Swirl in a 4 valve per cyl head

Post by 1NRO » September 3rd, 2010, 11:39 am

Handy to have an expert on hand :-)

I like the comparison of higher flowing heads not being any better than those with lesser cfm numbers, velocity? I've a vtec honda (the better halfs car) which is just horrible to drive, nothing until the rpm is silly high and then it's time to change gear. Big isn't always better it seems!
259

Guy Croft
Site Admin
Posts: 5032
Joined: June 18th, 2006, 10:31 am
Location: Lincoln, UK
Contact:

Re: Swirl in a 4 valve per cyl head

Post by Guy Croft » September 3rd, 2010, 11:47 am

velocity for sure, if you don't run variable valve timing/event on those 'big port' designs they can have very poor torque at low revs; conversely varying that cycle can make a huge difference as the comparatives betw the aforementioned engines in my new book show.

There is an inherent conflict between big port = high bhp (ie high torque at high rpm) and small port = good mid/lower torque on all engines but I think in the case of the Honda and several others they went a bit 'mad' on port size. They probably deduced that a very big port meant viscous losses at the port wall could be ignored - and they were right - you can 'polish up' the ports as much as you like and it makes no difference to the BPF at all.

G
Attachments
VTEC.jpg
inlet port - ferkin massive.
VTEC.jpg (117.65 KiB) Viewed 3966 times

1NRO
Posts: 123
Joined: June 22nd, 2006, 5:46 pm
Location: Carlisle
Contact:

Re: Swirl in a 4 valve per cyl head

Post by 1NRO » September 3rd, 2010, 12:06 pm

That is "ferkin" massive, looks to be the making of lazy air. I've seen those (or similar) ports full of weld/epoxy prior to porting in some peoples builds, they certainly didn't build a dog when all was finished. They seemed to focus on velocity and charge manipulation without a care for big cfm numbers. Makes sense to me when I cast thought to how much time air gets to fill a cylinder.

Nik
259

Guy Croft
Site Admin
Posts: 5032
Joined: June 18th, 2006, 10:31 am
Location: Lincoln, UK
Contact:

Re: Swirl in a 4 valve per cyl head

Post by Guy Croft » September 3rd, 2010, 4:18 pm

Yes, I should be interested to know if detailed computational analysis was performed on that design (and it's not confined to Honda) and if pressure waves (even if attentuated to an extent by the huge volume) still overcame the obvious net reduction in charge velocity thru such a big port.

Maybe the pressure waves work better? I just don't know.

Obviously the two work hand-in-glove but I figure there must be more than guesswork involved here.

Anyone with knowledge of that field - I'd like to hear from you.

GC

kpsig
Posts: 35
Joined: May 10th, 2008, 7:41 pm
Location: Greece

Re: Swirl in a 4 valve per cyl head

Post by kpsig » September 3rd, 2010, 10:08 pm

1NRO wrote:That is "ferkin" massive, looks to be the making of lazy air. I've seen those (or similar) ports full of weld/epoxy prior to porting in some peoples builds, they certainly didn't build a dog when all was finished. They seemed to focus on velocity and charge manipulation without a care for big cfm numbers. Makes sense to me when I cast thought to how much time air gets to fill a cylinder.

Nik
If I understood well, some people reduce the opening (x-x and y-y) of the ports on Honda engines? I would like to think that they do this only if they remove the Honda plenum and do some ITB's installation.
Kostas, Greece

PersonaGrata
Posts: 52
Joined: October 8th, 2009, 7:42 pm

Re: Swirl in a 4 valve per cyl head

Post by PersonaGrata » September 6th, 2010, 10:04 pm

There is no intrinsic bias in the Proton ports.
03092010189.jpg
03092010189.jpg (130.27 KiB) Viewed 3901 times
The manifold has very definite bias in the outer cylinders
03092010191 (1).jpg
03092010191 (1).jpg (69.3 KiB) Viewed 3901 times
And very little in the inner cylinders.
03092010192 (1).jpg
03092010192 (1).jpg (69.92 KiB) Viewed 3901 times



I’m not sure how significant that would be ultimately because I reckon that I can improve a port/valve combo by about 14% over standard. How much extra bias would be added by the manifold is an unknown – and the relationship between port imbalance and swirl is not likely to be linear. Its all unknown right now, but needn’t remain that way because swirl can be estimated on my bench with some modifications.

Ultimately, an imbalance in the power output per cylinder is less likely to lead to rough running than physical imbalance of the reciprocating machinery. ?????? Didn’t you say Guy that you had run engines with one odd sized cyinder bore and it was alright?

The BPF of these heads per cylinder (standard) is about 125 CFM at 10 inch. The BPF per port is about 67 CFM at the same depression.

In valve figures are about 122 CFM at 10mm lift, 10 inch per cylinder and 61 CFM per port/valve combo.

The previous head I modified was passing about 135 per cylinder after modification.

There are quite a few threads of thought that need to pursued right now before a decision can be made....

Like, what is the effect of improving one port, valve only where that improvement is just on the valve and seat? (ie port dimension standard otherwise and identical on both sides) I’m guessing that an increase of flow velocity on one side only will lead to pressure imbalance at the splitter – and subsequent cross flow across the face of the splitter. What will be the effect of this???

What effect will increased velocity on one side be on the inertial charging effect at the end of the induction stroke? Increased velocity must ultimately mean more momentum, but could the timing of that charging be out of phase ?– and/or could the stronger flow lead to charge being pushed out of the port with the weaker flow?? Could it even extend the range of rpm that the charging effect was significant at??

I looked at the Vizard article on polyquad and he talked of biasing the porting – was this to avoid the imbalance to the charging effect perhaps?

This whole business of port sizing and gas velocity is a tricky one. Instinctively, I keep thinking that there must be some instances where big ports are better than smaller ports but can’t work out any physical principles that would suggest that.

It seems to me that big ports would allow charge to be drawn with fewer losses in the first part of the induction stroke – giving a greater mass of charge. In the first part of induction cycle.
In the latter part, its probably more about velocity and timing to get the charging right, and this effect seems to be dominant in many cases.

So running conditions where induction length, or cam timing or rpm are not favourable for induction charging then the big ports would win? I’m thinking turbo’s would suite big ports?

Where conditions do favour induction charging then velocity is king, but of course that has to be balanced with port size because high velocities would never be achieved if the ports were too small...

Hmmmm..

1NRO
Posts: 123
Joined: June 22nd, 2006, 5:46 pm
Location: Carlisle
Contact:

Re: Swirl in a 4 valve per cyl head

Post by 1NRO » September 7th, 2010, 11:42 am

Iain,

To my knowledge there's no way an argument for big ports in forced induction, like in NA a % of the valve diameter is set as the controlling section, only a very small difference in MCSA between the two as far as I can tell. I look at it as the volume being the same (ish) with mass being the difference. Once a relevent % (no absolutly correct % it depends on application but ball park is within a few % either way) is set the focus thereafter is shape that assists the charge on it's route both before and after the smallest cross section.

Nik
259

Guy Croft
Site Admin
Posts: 5032
Joined: June 18th, 2006, 10:31 am
Location: Lincoln, UK
Contact:

Re: Swirl in a 4 valve per cyl head

Post by Guy Croft » September 7th, 2010, 1:20 pm

..except to say that if the controlling section is overly small for given boost it can choke (go transonic) and thereafter raising the boost will not give any more flow, merely raise the manifold pressure.

G

Post Reply

Who is online

Users browsing this forum: No registered users and 9 guests