> Has anyone had any experience with headers on an MGB?
I'm doing tbhat now (hi Chris!) and have studied it in GREAT detail,
both in MGB applications and in extrapolation from A Series knowledge.
> Looking at my catalogs I see prices on headers
> ranging form $99 (Victoria British) for a chrome plated 3 branch header to
> about $300 for a stainless steel one from Moss.
What do you want to do, just make the car run or make it better than
a stock pre-75 B? Any of these will make it run. Nothing you've mentioned
so far will make it run any better than a stock pre-75 manifold and
downpipe; about the most you can hope for is to reduce underhood temps
and weight by getting rid of the cast-iron lump. The factory did a
surprisingly good job of tuning the pipe lengths. (You *can* pick up
a little flow in the rest of the exhaust system, BTW, by going to a
low-restriction or single-box exhaust.)
> I notice that the stock pre-1974 manifold brings the front and rear
> ports together, with completely a separate branch for the middle exhaust port.
> This seems to make some sense. Since the middle port serves two cylinders,
> combining the outputs of the front and rear ports might bring the back
> pressure on those ports up to the level of that for the #2 and #3 cylinders.
Where to start, where to start...
You're right in your observation about branch geometry. This comes from
a type of design called "long center branch," or LCB, that is common in
Mini, Midget, and Sprite headers, and less so in MGB designs because the
factory's is generally good enough till you get into internal engine mods.
There's a lot of nonsense generally bandied about among the public concerning
back-pressure; the word back-pressure is so vague (positive or negative
pressure? what's the flow rate? what's the flow velocity? -- and yes, those
are different concerns -- what's the pulse timing?) that I'm not going to
use it because it's basically hopeless. Not everyone who uses the word
backpressure is an unreasoning savage who should be kept from using any
implement more complex than a flint adze, but you should suspect these people
as such until they prove themselves capable of complex reasoning and
articulate discourse. See if they're frightened of fire before you order
from them.
> >From the pictures in the catalogs I note that some of the header designs
> follow the approach used with the stock manifold while others provide
> three completely separate down pipes. Does one design have any advantage
> over the other?
I ought to apply for a service mark on the phrase "It depends." (SM) :-)
If all you want to do is duct exhaust gases out of the car in a more or
less nontoxic-to-the-passengers manner, then no. Buy the cheapest one.
If you want it to work better than the log manifold that British Phlegm-
sucking Leyland in their infinite pigheadedness used on post-75 MGBs,
then also no. Buy the cheapest one. But if you want it to work better
than what Syd Enever designed for the MGB back in 1962, there's a bloody
great difference between good designs and future recycling projects.
Basically, all 3-into-1 headers are either useless on any engine that
will ever be run below 7000 RPM, or just plain useless. The trick is to
see where the 3 tubes connect. If they connect above the level of the
steering column, they're plain useless. If they connect somewhere
under the driver's calves, they're only useless on all but full-race
E Production motors, and of marginal value there unless you run the
motor at 8000 RPM and up. Here's why.
My last epic on the subject of exhaust system design compared the
exhaust system to a sausage-making machine. It's still the best
analogy I know of: Imagine if you will four sausage grinders connected
to a bunch of pipes. Those are the pistons; the camshaft controls when
the sausage-casing is twirled, therefore how long the sausage is -- well,
the exhaust pulse, actually.
What the header has to do is take sausages from all four grinders and
duct them down the exhaust pipe. Each sausage pushes the one in front
of it a little farther down the pipe. Unfortunately, the force required
to do this comes out of the work done by each sausage grinder; if you
can reduce the amount of energy needed to push each sausage down the pipe,
you can push more sausages per hour. If the sausages get backed up,
the grinders bog down; if the sausages come out too fast, the grinders
can't push the sausages out the end of the pipe quickly enough. Get it?
So a little imagination should get you to the understanding that there
is a "sweet spot" (no, not left over from making chicken-apple sausage)
where things work best: where the inertia of each sausage helps not only
to push the downstream sausages out the pipe more easily, but actually
helps suck on the sausages that are still upstream and take a load off the
grinders. That sweet spot is going to be dependent on three things:
1. The speed of the engine
2. The characteristics of the camshaft
3. The configuration of the exhaust tubes
Configuration includes shape, diameter, and length between connectors.
For the MGB, the LCB design works best because of the timing of the
engine and the siamesed ports, as you suspected. What you want is for
the #1 and #4 ports to be connected about 15" down from the exhaust
face, and for the middle port to be about 34" from exhaust port face
to the collector. Longer gives you better low-end torque, shorter
gives you better high-end operation. MGBs don't really require more
than about 1.5" of internal diameter in the header tubes, but 2" to 2.5"
in the main exhaust system is a good size. If the tubes are too small,
you restrict flow (trying to push too much sausage through too small a
pipe); if it's too large, you don't get inertial effects till you run
8000 RPM or up (the sausage just sits there until it comes in so fast
that each slug pulls the next one after it). The best systems are
designed to provide a boost in efficiency over the widest range of
RPM, but the whole system is like an organ pipe tuned to resonate
best at a particular frequency.
What this gives you, at the sweet spot whose exact location (or actually,
range) varies with the characteristics described above:
The MGB firing order is 1, 3, 4, 2. So the #1 cylinder pushes its slug
of exhaust gas (its sausage) out the valve and down the short forward
tube. It gets about to the junction with the #4 pipe when the #3 cylinder
pushes its slug down the middle pipe. The #1 slug gets down to the main
collector about the time the #4 cylinder pushes its slug down the rear short
tube, and here's where the magic happens: if it's all the right sizes and
dimensions, the #1 pulse is just going out into the collector about the
time that the #4 pulse comes into the 1-4 pipe, and there's a negative pulse
(what's a negative pulse! you didn't tell us about any damn negative pulses!)
that reflects back up at the exhaust face of the #4 cylinder just in time
to help suck the exhaust out and take some of the load off the engine.
Likewise, when the #2 exhaust valve opens, if it's timed right, the #3
pulse has gone from the center pipe into the collector, the negative
pulse travels back up, and this low-pressure wave sucks the burned gases
out of the #2 cylinder, reducing the amount of power required by the
engine to get the exhaust out of the car.
Okay, negative pulses: when gas goes out of a pipe into a larger, less
restrictive container (such as a bigger pipe or a muffler), there's a
reflection of low-pressure gas that travels back up the pipe at the speed
of sound (hey, it's kinda like TDR with gases!). This negative pulse can
help increase the pressure differential across the exhaust valve face,
and any increase in this pressure differential makes less work for the
engine to evacuate the cylinder, meaning that more power is available to
turn your expensive high-performance tires into expensive high-performance
skid marks at the starting line. This is the main point of exhaust
tuning; the part about matching the pulse velocities to the port volumes
is science, but fitting this to the rest of the engine is an art.
(For the record, yes, Mazda in their racing cars have experimented with
varying the tube lengths under electronic control, so that the tube
length varies with engine speed and load. It's easier with intakes than
with exhaust, because of the heat issues, but it's still way flipping cool.)
> Is there one brand which is distinctly better or worse
> than another?
Janspeed (in the UK) generally know what they're doing; you can get Janspeed
headers from Seven Enterprises in, um, Virginia or some other Piedmont state.
C'mon, somebody must know their number. I'm going to order mine from my
usual supplier, because it's been worth my while to have a local shop that
I can go to to buy stuff when I need it.
The Moss stainless steel LCB header looks decent, certainly no worse than
the stock setup. But really, the factory setup is pretty close to spot on
for all but heavily modified engines. And you can get some tangible gains
with the factory header by grinding the inlets to the header -- the parts
that butt up against the manifold gasket -- to just *larger* than the
exhaust ports. This helps forward flow, makes everything work just a
little better, and was something I could actually feel when I did it to
my stock MGB. For my next act, I'm getting a Janspeed (one of these days...)
--Scott "Hey! Only eight to ten pages to go in the last chapter!" Fisher
|