Port matching - how far to blend? (Peugeot mi16 head)

[AdamN]

I am currently working on a spare cylinder head for my engine, a Peugeot mi16. The head is virtually the same as the S16 that Guy worked on and detailed in the Cylinder Head Preparation forum.

However, one thing I have never come across in these forums or anywhere else, is how far to blend a port when port matching.

For example, my head and inlet manifold look quite good, but on checking them carefully, there is a mismatch.

The manifold ports have a very slight downstream mismatch near their roofs, which i'm happy to grind out.

Unfortunately the inlet ports have a rather surprising 2mm or so upstream mismatch, around the floor (see the red line on the pic below. That isn't my head, but shows where the mismatch lies).

My question is, in grinding this out, how far into the inlet port should I go?
The actual port shape looks very good, but it just doesn't match the inlet manifold/gasket.

Should I grind it out until the edge of the port has 'line of sight' down to the valve (as the standard ports are), or should I leave the port shape well alone, but blend in a 'ramp' to the existing port? How steep should this ramp be?

I appreciate that these are questions that only a flowbench could truely help to answer, but I imagine that there is probably a rule of thumb.

I hope the replies to my question are useful to other wanabee head tuners.

Many thanks,
Adam

[Guy Croft]

OK, but still need some photos that show me the relative entry axes and shapes.

GC

[AdamN]

Hello again Guy. I managed to get some pictures today. Hopefully they convey all the required information.

The ports are identical to the S16 head.

[Guy Croft]

OK, I have a handle on this now.

1st point is that a mismatch - even an upstream one - where metal in the head is standing in the way of the incoming airstream, can have relatively little effect.

For definition of upstream see:
http://guy-croft.com/viewtopic.php?t=7

If in doubt, remove it. With this head the airstream is going to be biased towards the upper half of the inlet manifold, because air flows faster round the outside of a curve, as does water. So a severe mismatch there might be an issue. You don't appear to have one. The mismatch apparent in the injector relief/region in head and manifold can be ignored completely, there is hardly any useful airflow in that part because of the intrusion of the injector itself.
As far as the lower port section is concerned, you indicate an upstream mismatch, metal would need to come out of the head to correct? In that case I'd play safe and remove it.
Go in about 16mm at 7 deg to horiz as shown red line below and then blend it into a smooth curve, green line. That curve would, typically give a fractionally lower flowrate than a straigh run on the port floor, but I am not going to suggest going line of sight from port entry because a) it's not that important with that particular inlet manifold, and b) I don't know how thin it would make the port floow section as you approach the splitter.

Hope that works out OK.

GC

[AdamN]

Thank you Guy, that advice is much appreciated.

Out of interest, how would your advice differ if the manifold was short and straight, as if carbs or throttle bodies were fitted?

This isn't a route i'm likely to go down, but it's good to know and could be useful for anyone else searching 'port matching'.

[Guy Croft]

Well, short answer is it wouldn't differ: if in doubt - match.

The manifold shown here was from Richard Longman & Co, for throttle bodies, good piece of kit. I swear, the head flowed more with it on than off, as-is and matched/prepped.

As you can see there is a curvature - as I outlined to you - on the port floor, where the manifold joins the port, there isn't line of sight.

As a footnote, perhaps, as you say, for others reading, I have to say, all in all, that despite its being a very high flow head, the ports are just way too big on that head to make it premier choice for competition.

The power results we got from the one featured in GC V/W were pleasing to the owner but not to me - 225bhp on cams with about 11mm lift or so, DTA Jenvey. There are those who might say, wow. Not me.
I guess you live and learn. What I finally concluded, on getting that dyno result, is that those big port layouts with cavernous entry volume - compared with say Cosworth designed heads like the Sierra Cosworth, Vauxhall HS, XE etc., (with smaller, near parallel run twin-barrel ports) have a real velocity & cylinder filling problem. This is, incidentally, a feature common to other production 16v heads eg: Honda, Suzuki, Toyota. They are examples of heads designed specifically for road cars and sure, the engines obviously do that job well or they'd never pass compliance tests. But they just don't have the tuneability of other designs.

GC

[sumplug]

Well in OE fitment, this engine is tilted back, so you need a curved manifold, but Throttle Bodies are set around 20 degrees upwards, so would allow a less curved manifold.

Andy.

[AdamN]

Interesting comments about production cylinder heads Guy.

One thing that I have heard numerous times (though I have no idea if it is true or not) is that the mi16/S16 head were almost identical in port shape to the Turbo T16 rally cars of the 80's.

Would this have any bearing on the port design if it were true?

[Guy Croft]

Turbo port vs n/a port, world of difference regarding what's needed there. Maybe they just carried over an existing casting to save cost and used it normally aspirated and it worked fine!

Turbocharging builds up pressure in the port and generates a high pressure ratio across the valve right thru the lift cycle, so you're far less dependent on air velocity.

GC