Triple SUs

[John Coffey]

Why more is not always better when it comes to the L6:

A triple SU carburator setup might work if they are attached to a plenum. If you try to feed individual runners from triple SU (two runners per carb) you'll have trouble making as much power as dual SUs.

Why? The firing order of the L6 is: 1 5 3 6 2 4. Keep this in mind while we assume that front SU feeds cylinders 1 & 2, middle SU feeds cylinders 3 & 4, and rear SU feeds cylinders 5 & 6.

On a dual SU installation you get an intake pulse to each carb every 240 degrees of crankshaft rotation, and this is evenly spaced. You end up with a pretty even mixture balance between all the cylinders.

On a triple SU installation you get an intake pulse on each carb as follows:

Front SU: There are 3 cylinders (intake pulses) with 360 degrees of rotation between 1 & 2 and 1 cylinder (intake pulse) between cylinders 2 & 1 with 120 degrees of rotation. This causes an imbalance in the mixture (rich to lean) between cylinders 1 & 2.

Middle SU: There are 2 cylinders (intake pulses) with 240 degrees of rotation between cylinders 3 & 4 and 2 cylinders (intake pulses) with 240 degrees of rotation between cylinders 4 & 3. This creates a balanced mixture between these two cylinders.

Rear SU: There is 1 cylinder (intake pulse) with 120 degrees of rotation between 5 & 6 and three cylinders (intake pulses) between cylinders 6 & 5 with 360 degrees of rotation. This causes an imbalance in the mixture (rich to lean) between cylinders 6 & 5.

I may have my number of degrees off, but I think I've illustrated how it would be difficult to achieve a balanced mixture across all cylinders during one full power cycle (720 degrees of crankshaft rotation) with a triple SU carburator setup using individual runners.

Again, if you ran these triples to a plenum and then out to the runner, you could work up some power.

------------------

----------

John Coffey

johncof@veriomail.com

[This message has been edited by John Coffey (edited 05-04-2000).]

[John Coffey]

Another post on this topic from Kyle:

John-

Excellent post, very hard to <snip> down, so here's my attempt at

distilling the data:

Cyl Deg | f(2x) r(2x) | f(3x) m(3x) r(3x)

1 0 | 0 - | 0 - -

5 120 | - 120 | - - 120

3 240 | 240 - | - 240 -

6 360 | - 360 | - - 360

2 480 | 480 - | 480 - -

4 600 | - 600 | - 600 -

John's point becomes very obvious when you look at the "-" in the

chart; on a double carb (2x) setup, there's a pulse to each carb

every other cylinder in the firing order. In a triple carb setup

(3x) the pulses are not even, except for the middle (m) carb.

A very interesting point, to be sure. Seems that there was

someone around that ran triple SUs, was it Preston? I'd be

interested in hearing what was done to work around this issue. A

balance tube might do it, but I'd guess that the stock type tube

would be too small.

And if you have to run a common manifold (plenum), you'll get

into the same problem as with four bbl manifolds - long intake

path, fuel puddling. The rear carb in a triple setup would

effectively feed the other cylinders in the "dead" space between

#6 and #5, for example.

With a dry-flow MPEFI setup, this might not be as big a deal.....

but then again, why not just run a larger, single throttle body?

Hey! All you have to do to make this work is change the firing

order! A new cam and crank, and you're set! <> Looks like

1,5,3,6,2,4 might do it. 'Course, then the famous L6 smoothness

would go bye-bye;-)

Hmm... let's dig my grave a little deeper here... at 1000RPM, you

have 3000 intake pulses per minute, right? If that's right, then

you have an intake pulse every .02 seconds, on a 2x setup. That

means that each "-" in the chart above is equivalent to a .02

second dead spot in the intake flow. (jeeze I hope this is

right;-) Going further, that means that the biggest dead spot in

a 3x setup would be .06 seconds. And that's AT IDLE. At

3000RPM, there'd be 9000 pulses, one every .0067 seconds, the

worst dead spot would be .02 seconds. Does this mean that a 3x

SU setup would be as smooth as an idling 2x setup? And mightn't

velocity stacks or longer intake runners take care of this?

Bench engineering is almost as much fun as bench racing!

Kyle

[John Coffey]

Another post on this topic from Marc Sayer (I think he sent this out a couple years ago):

From: Marc Sayer

Subject: Triple S.U. Setup - not so good.

Well I see I really opened up a can of worms with this. Before people

start going out to buy up triple SU bits, let me run this down for you so

you understand what's up.

With the dual SU manifold the air / fule distribution is poor because the

runner lengths are different. Datsun worked their butts off to try and

compensate for this and really did a pretty good job (you should see how

bad some of the British stuff is). The advantage of the dual manifold is

that the pulsations come in even periods. Let me try to put this visually;

1---

2------- 103020103020

/

3---/

A "0" represents a "dead" time slice, a time when none of the cylinders

fed by this manifold are in the intake phase. Any other number indicates

the cylinder which is in the intake phase.

4---

5------- 050604050604

/

6---/

As you can see the pulse pattern is very even, being one time slice on, then

one time slice off.

This means that each cylinder is pulling on a column of air/fuel (what is

contained in the manifold) with about the same velocity, because each

cylinder is preceded by the same length of "dead" time in the manifold.

This promotes all sorts of positive effects, and keeps air speed in the

manifolds fairly constant. It limits the amount of "flow reversal" that can

take place by limiting the length of "dead time" in each manifold. These

effects are mitigated to some degree by the differences in individual

runner lengths and by the distances from the main plenum of each manifold

to each intake valve. But all in all this manifold and firing order do work

fairly well together.

Now here is the pattern for a triple SU set up;

1--

--- 100020100020

2--/

3--

--- 003004003004

4--/

5--

--- 050600050600

6--/

You can see from this that the front and rear manifolds have uneven

pulsations. 3 time slices off, 1 on, 1 off, 1 on. This means that the cylinder

that opens after the 3-time-slices-off "dead" period, will be pulling on a

much slower column of air/fuel than the other cylinder will.

This will cause those cylinders (#'s 2 & 5) to run a little leaner and their

counterparts (#'s 1 & 6) to run a little richer. The reason for this is that

air reacts to changes in velocity more quickly than the fuel does.

Since the air/fuel mix in the manifold has 3 time slices to slow down

before cylinders 2 & 5 draw on it, it will be slower for them and as it

tries to speed up again, the air will respond more quickly and get to the

cylinder slightly sooner that the fuel. By then the intake stroke for these

cylinders will be ending, the extra fuel meant for these cylinders will

stack up in the manifold and then cylinders 1 & 6 will open.

They will be opening to a fast moving, extra rich, mixture. (On some 4 cyl

motors this phenomenon is so bad that the engine idles on only two of the

cylinders if the mixture is not within range).

The center manifold is different, however, its pulsation pattern is even,

with 2 off time slices followed by 1 on time slice. each cylinder will be

pulling on a column of air/fuel with about the same velocity, but

thatvelocity will be fairly low due to the fact that twice as much time is

spent "off" as is spent "on".

So the triple SU manifolding has three basic problems;

First, the pulsing patterns are not the same for all manifolds and,

Second, the outer manifolds will have distribution problems caused by

uneven pulse patterns and,

Third, the center manifold will suffer from reduced velocity.

The third problem is the least important of the three and is true for

triple Mikuni's as well (in fact they suffer from a 5:1, off to on, ratio). I

would not be concerned about this except as it relates to the behavior of

the outer manifolds.

The solution for the outer manifold's distribution problems is probably

going to be similar to what was done for the British 4 cyl, siamesed

engines. The runners of the manifold were "aimed" at the lean running

cylinders slightly. This helped to overcome the distribution bias that was

created by the uneven pulsing. I am not sure how well this would work on

the Z motor, but on MGB's it helped quite a bit. It didn't eliminate the

problem, but it did significantly reduce it.

The last problem, that of the manifolds not all having the samepulse

patterns is not soluble except by altering the firing order. But by

ameliorating the negative effects of this, we can to some degree cancel

the problem out.

If the manifolding is done well it will be equal to, or slightly better than

the stock manifold for distribution, but will increase flow. If runner and

plenum size is kept small enough to keep air velocities up, the drivability

and throttle response should be better than a dual set up.

Anyway, there's something for all you triple SU freaks to chew on for a

while. You could solve all these problems by going to 6 of the 1.25" (32mm)

SU's you know. :-)

Marc Sayer