DeTomaso Mailing List: October 98, Message #104

> Subject: Yet another head porting opinion (part 2: port flow)
> 
> Continuing the series I feel like I have to talk about actual port flow
> before hitting the CNC topic.  Like I said just a few moments ago, I
> would rather look at low lift flow and the shape of flow curves than
> peak flow numbers anyday.  This is analogous to comparing horsepower
> numbers.  WE have seen over and over the effects of cars with different
> peak horsepower numbers that result in non-logical ET's at the
> dragstrip.  What is usually at work here is a car with lower peak power
> that gives quicker ET's has a little better torque curve than one with a
> higher peak number that doesn't ET as well.  It's relevant that ET is
> all 'made' in the first 1/4 of the track and mph in the last 1/2.  MPH
> is usually a very good indicator of how much peak power an engine has,
> but doesn't tell much about how the car works (tractionwise) or how good
> the ET's will be.  A car with a smooth torque curve that is high will
> typically build speed early in a run and result in good ET, but may not
> necessarily make big MPH.
> 
> A little of the same logic is useful when talking about cylinder head
> flow (which is merely a measurement) and cylinder filling (which
> ultimately makes the horsepower).  If the low and mid lift flow rates
> are very good with no anomalies (like a dip in the plotted flow curve),
> the air will get moving faster sooner in a port and can ultimately
> produce more power than a port that has a high peak number and poor low
> lift.  Just like the car with a meaty torque curve makes good ET's.  I
> hear it over and over and over "well, this is a high revving engine, so
> it needs more flow at high valve lifts and we don't worry much about
> what happens at lower lifts.  Low lift flow only affects low rpm
> driving."  That statement can't be more wrong IMO.  The valve opens and
> closes at low lifts every time it opens.  If the low lift flow is good,
> the air will start moving better a little sooner in the cycle which is
> tough to do since it is essentially stopped and the pressure
> differential is not that high. Then, good low lift flow will also let a
> little more in late in the cycle when the valve is almost closed and the
> air is still trying to get through (only this time the air is already
> moving and the closing valve is trying to slow it down, so the shape
> needs to lend itself to very high velocities at the seat, often
> supersonic.
> 
> Of course high lift flow is important as well and like 1/4 mile trap
> speed is a good indicator of applied horsepower, maximum port flow is a
> good indicator of the  potential power an engine will make with those
> heads.  The old saying "an engine is nothing more than an air pump"
> still holds true.  The more air that you can move through, the more
> power it will make.
> 
> Problems arise here when trying to compare different shops' flow numbers
> and determine what the 'best' deal is going to be.  Often, shops that
> report very high peak flow numbers are either outright lying about the
> measured flow or haven't looked after the low-midrange flow in the
> ports.  I suggest getting all the shops you are interested in to fax or
> send you a copy of a flowsheet for their portwork on a certain set of
> heads.  It's best if they flow them at every .050" lift, but most only
> hit the .100" numbers.  Whenever I flow heads I do them at every .050"
> since there can be trends in the plotted curve that you won't see if you
> only take the .100" readings...especially if you are only measuring
> through .600" or so (that is only 6 data points).  The most modern
> flowbenches will actually do a sweep of a lift range and give you the
> actual flow curve rather than just a couple of data points.  Me likes. 
> Anyway, take these flow sheets and map them into a spreadsheet (hell,
> you can even plot them by hand, but it's a pain) and overlay the
> different curves to see what they look like.  Provided they didn't lie
> about the numbers, you will quickly get an idea of what heads will make
> more power by looking at how meaty the curves look.  If they are smooth
> and rise rapidly at the low lifts the engine will make good power. 
> Bonus if the curve keeps on climbing above .500" since that means they
> didn't hit a flow barrier.  If there are ANY dips in the curve, call
> them up and ask them about it.  If they resent you calling and asking
> questions, become wary.  Remind them that you are the potential customer
> and are looking at spending what you consider to be a considerable
> amount of money.  Answers like "these are the best money can buy" won't
> cut it at this point.  Especially if they have something bad to say
> about someone else's work.
> 
> It is important to keep in mind that when you are dealing with a certain
> port configuration (in this case, a standard 23 degree small block
> chevy), all of the better port jobs will be almost impossible to
> differentiate from each other.  What happens is that with the 23 degree
> SBC, the ports have been around a while and EVERYONE that ports heads
> has done a little development on them and people that do a lot of heads
> have worked a LOT of them.  There is a lot of data out there on this
> configuration and once they are that highly developed, they will almost
> all be exactly the same unless someone starts welding and moving the
> valves and ports around.  I am talking about fully ported heads and
> combustion chambers here, not partly ported heads.
> 
> Now, what probably 95% or more of us need to get our cars to hit that
> price/performance point, is partly ported heads.  Again the concept of
> diminishing returns pops it's head.  I strongly stand by the statement
> that 90% of the performance of a set of heads (and consequently the
> engine itself) can be realized with just a good carefully selected
> partial port.  This means porting the combustion chambers and the valve
> bowls and not worrying about the ports themselves.  This is because the
> ports are not the problem with unported heads...the bowls and chambers
> are.  If you want to think in terms of importance, you will work from
> the actual valve seat out both directions.  The areas within 1/2" of
> each side of the valve are the most critical in the entire head.  You
> will save a ton of money getting a bowl port and great valvejob and will
> have damn near all the performance of a full port (at probably twice the
> cost).  If you were to run that partially ported set of heads for a
> while and then get them fully ported, you might be surprised that you
> would feel little to no difference (let alone the difference you felt at
> getting the bowl port for the same cost).  At the track, depending on
> your combination, you might see a tenth or two at the most with the full
> port over the bowl port unless you are seriously port limited in which
> case you should have LT4 or AFR heads anyway.  For a mild 383 or even a
> strong running 350, it seems like to me that the extra $6-700, maybe
> more, saved on getting a bowl port is better spent somewhere else (like
> a really good PCM program or Grotyohann headers) than getting that last
> 5% of port flow that you will never notice anyway.  Hell that is the
> cost of re-gearing and putting an S-10 converter in the car.  Just
> something to think about.
> 
> Notice that there has been no discussion on velocity yet...this only
> covers raw port flow.  Velocity is very important and will be in another
> writeup.
> 
> To be continued ...
> - -- 
> Scott Knight  mailto:sknight@mich.com
> http://www.mich.com/~sknight IRC:SS396man ICQ:8353618
> '95 Black Impala SS
> '94 Ducati 900SS CR
> 
>