Added - 2/19/01
Some people may say that this system is just nothing more than a
"Rev Kit". However, as you read on, you will see that a lot of thought, not to
mention a lot of time and money, went into the finished product..
First some background information on why we embarked on this
project in the first place.
Back in 1993 I read an article in the October edition of "CIRCLE
TRACK" titled "Trulink Rocker Arm System". This article dealt with a company
called DECUIR DEVELOPMENT and their development and adaptation of a new valve
In April 1994 "CAR CRAFT" also published a further article
featuring this system titled "Spring Ahead".
Back then "PETERSEN PUBLISHING CO" which is now called "emap
-petersen" published both of these magazines. For further insight on this valve
train system, request copies of these articles by sending check or money order
for $5, made payable to EMAP. Address to: "Emap. USA. Inc",
Attention Editorial Assistant, 6420 Wilshire Blvd., Los Angeles, CA 90048-5515.
Phone 323 782 2000.
What follows are the reasons for, and a brief description, of this
During the engine rebuild process it will be noticed, after
installation of the cylinder head and rocker gear, that turning the engine over
by hand has become increasingly more difficult, all in part due to valve spring
#s. Simply put, it requires horsepower to turn the valve train. The greater
these valve spring #s, the greater the horsepower required to turn the
All assembled valve spring #s are subject to a rocker arm
multiplication factor. This multiplication factor is exactly what the camshaft
lobe sees at the lobe/lifter interface.
For our example let us use an MGB engine valve train with British
Automotive's 1.55:1 roller rockers, camshaft P/N 6710-18) and Moss Special
Tuning valve spring set P/N TMG10707 (423-455).
First, we need to observe the following data:
Now let us look at what happens if we move the inner valve spring
37# to the pushrod side.
We now have to calculate the pushrod inner spring #. Let us
presume that we have the same valve spring seat 37# at the same installed
height 1.575". Since we know that this spring rate is 76# we can simply
multiply this number by the camshaft lift of .289 (76# X .289") = 22# then add
the seat # for a total 59#.
In one complete camshaft revolution we would have realized a
savings of 392#s. (2440# minus 2048#). At 1000 engine RPM (500 camshaft
revolutions) we would see a savings of 196,000#s.
NOTE: This is just an exaggerated example and not a working
model. In this case the pushrod spring seat# would have been far too excessive
to a point where the lifter, in all probability, would not rotate, resulting in
rapid camshaft lobe and lifter wear.
The forces present at the cam/lifter interface are the sum of the
inertia force and the valve spring force. At low engine speeds the inertia
force is very small so almost all of the force on the cam is due to the valve
spring. At maximum RPM, the valve spring force is just equal to the inertia
force before the valves float. This means that friction savings would decrease
as the engine speed increases.
Reduced valve spring #s is not the only advantage with this
system. We continue with the following:
During the valve closing sequence, with the valve fast approaching
contact with the valve seat, we have linear velocity and momentum of the valve,
pushrod and lifter. Likewise, we have angular velocity and momentum of the
rocker arm. When the valve contacts the valve seat, it's stored energy is taken
up by the valve spring. The stored energy in the rocker arm, pushrod and lifter
is however not taken up.
This energy is stored mechanically in these 3 components,
especially in the pushrod. The faster the valve is moving, the instant before
it makes contact with the valve seat, the more energy will be stored in the
valve train. When this energy is released the rocker arm tends to "dance" on
the valve stem tip. The more energy that is in the system, the more force the
rocker will have to "tap" the valve with.
The "Decuir Development Company" found a way to absorb the stored
energy of these 3 components. This was done by positively linking the rocker
arm and pushrod assembly as a single unit (by way of a special rocker arm
adjusting screw and a threaded female pushrod sleeve). Also by preloading the
lifter with the installation of an inner valve spring on the pushrod itself.
The above text was edited from the original format with changes
and additions made where necessary.
Let us continue with the trials and tribulations of attempting to
adapt this system to the MGB engine (June 1995).
First, this system was not going to work unless we could find some
way of straightening up the inherent "inward" pushrod lean common on all "B"
series engines. Manufacturing the following components partially solved this
problem. The maximum amount we could offset the rocker pedestals was found to
Next up was to order the components for one valve only. This
consisted of the following components:
These components were assembled, without the pushrod spring, as
per illustration "A".
Maximum valve lift readings were recorded and compared against
previous engine rebuilds that included the same camshaft and the OEM
rocker/pushrod set-up. Something was drastically amiss; we could not achieve
anywhere near the valve lift of an OEM 1.426:1 * rocker arm. Remember we
are using a 1.55:1 roller rocker arm.
*Most camshaft regrinders use the OEM ratio of 1.426:1
(1.358" divided by .952") to quote anticipated maximum valve lifts. These
numbers represent the center line distances of the valve contact anvil and the
valve clearance adjusting screw relative to the rocker arm bush center line. It
is quite possible that you may come up with a ratio somewhat less than 1.426:1.
Simply put, OEM rocker arms never give you this ratio. The 1.358" number has
been verified over and over again, however, the .952" number can vary as much
as +.030" between various rocker arms.
Since we were using a 1.55:1 modified roller rocker arm (1.443"
divided by .931") we should have seen increases in maximum valve lift.
Apparently, the critical point is in the initial pushrod inward lean, valve
closed, and the pushrod to rocker arm attachment method. The engine was turned
over slowly and the following observed, as the pushrod moved upward it began
outward lateral movement to a point before full valve lift. No more lateral
movement of the pushrod was observed until the same point was reached on the
valve closing cycle, at which point the pushrod began inward lateral movement
back to the original starting position.
Another pushrod was manufactured that would allow the use of the OEM adjusting screw, male and female threaded adapters as in illustration "B". Once again we went through the checking procedure mentioned above. On this particular check we ran into the same problems but to a lesser degree. Pushrod lateral movement was certainly less than "A" however, we were still short on valve lift.
At this stage we simply decided to fit an OEM style re-profiled
rocker arm as per illustration "C" with the same pushrod as
used in "A" without the threaded female adapter.
Hey Presto! Problem solved. Very little pushrod lateral movement
with anticipated valve lift.
The OEM rocker arm with its ratio of 1.426:1 allowed the pushrod
to attain a more vertical position at rest. I doubt whether or not this was the
only reason, especially when look at the two dimensions, for each rocker, from
the rocker arm adjusting screw ball center line to the center line of the
rocker arm. 1.55:1 roller rocker arm .931" 1.426:1 OEM rocker arm .952" a
difference of .021".
Having the pushrod and rocker arm assemblies independent of each
other ("C") has more geometrical advantages than either "A" or "B".
During the valve opening and closing sequence the rocker arm is
constantly changing angle. In "C" the rocker arm adjusting screw ball and
pushrod cup work independently of each other, this allows the adjusting screw
ball to establish any unrestricted angle during valve opening and closing
without any great amount of pushrod lateral movement or loss of valve lift.
In "A" and "B" the adjusting screw and the pushrod are an
assembly. Although there is full ball and cup articulation within this
assembly, I am at a loss to explain why we could not achieve greater valve
lifts. I must admit I don't quite understand the geometry at work with this
system, but obviously it has something to do with pushrod installed angles and
the accompanying rocker arm angles during the valve opening and valve closing
Decision time. Where do we go from here? We decided to continue
with the project but leave the rocker adjusting screw and pushrod as
independent components. We now concentrated on developing the rest of the
system. Before valve spring and pushrod spring analysis could be carried out we
needed further information.
All the valve train component weights were recorded. Our intended
valve spring set P/N TMG10707 (423-455) was sent to "Elgin Racing Cams" to
establish "natural frequencies" (vib/min) and "harmonic numbers" as was a
roller rocker arm to establish the "moment of inertia" scale reading.
What came back was a full analyst of the valve train requirements
to allow the engine to operate safely in the 6500-RPM (30% safety factor) to
7000-RPM (15% safety factor) range when using this new system.
Our original intention was to just simply install the 423-455
valve spring set and use the outer valve spring on the valve and the inner
valve spring on the pushrod. Steve Gruenwald of "GRUENWALD SOFTWARE" in
conjunction with Dema Elgin came up with some interesting conclusions. If we
had split up the valve springs, as I intended, we would have had the valve
springs vibrating out of control well before 6500-RPM with subsequent valve
float. Apparently, the "natural frequency" and "harmonic number" range numbers
at 6500-RPM of both springs was too low when coupled with the higher "moment of
inertia" (more weight centered over the roller rocker tip than the OEM rocker)
of the roller rocker assembly.
New valve springs were sourced resulting in the following:
As previously mentioned, this set up allows us to operate at an
engine speed of 6500-RPM with a 30% safety factor. This will also allow enough
spring pressure to account for increasing valve train flexibility as the engine
speed increases without having excessive pressure that causes premature
Let us look at the following:
When compared with our example of the OEM dual valve spring set
423-455 we realize a savings of (2440# minus 2360#) = 80# per crankshaft
revolution. At 1000 engine RPM (500 camshaft RPM) we would see a savings of
40,000 #s. However, as we stated previously this savings would not be linear.
Let us presume that this figure of 40,000 #s is our maximum
savings at 1000 engine RPM and using 6500 engine RPM as maximum engine RPM
"valve float" condition with zero #s load on the camshaft lobes, can we just do
a little math and draw a straight line graph representing theorized loads at
various engine RPMs?.
1000 RPM = 40,000 #s/min savings (Idle Speed)
This should give us some approximation of spring # savings at
various engine RPMs.
Another benefit of separating the inner and outer valve springs
lies in the elimination of friction between these two springs. This friction
would have increased with engine speed. Not having two springs rubbing against
each other (or a damper) reduces the wear of the springs.
Yet another benefit is provided by the pushrod spring which
permanently preloads the camshaft. This preloading dampens the camshaft cyclic
loading and unloading forces, resulting in a more stable valve train with
reduced distributor drive gear "chatter" and stable ignition timing.
Less valve train loads equals less valve train component wear.
With this set-up on my 1979 MGB "Limited Edition" the most
noticeable thing, upon engine start up, was how much quieter the engine
On June 3rd 1996 we drove the MGB out of the shop (mileage
104,730. As of November 11th 2000 (mileage 110,429) with only 5,699 miles the
AVT system is still working fine.
The engine in the MGB is my own developed 1924cc (actually 1927cc)
with 9:1 GCR. This engine was previously dynoed at a whopping 137ft/lbs of
torque @ 3500 RPM and 115BHP @ 5000 RPM. Although the AVT system was designed
to peak at 6500 RPM with a 30% safety margin, I see no need to rev the engine
above the maximum BHP RPM. Occasionally, when the situation deems necessary, I
jump on it, with no complaints from the engine.
Did I notice any improvements in the cars performance after
fitting the AVT system? I can honestly say I don't know. With such a flat
torque curve 128ft/lbs @ 2500 RPM. - 127ft/lbs @ 4500 RPM prior to the AVT
installation, significant increases would have to have been realized in both
torque and BHP figures for a "seat of the pants feel".
Very few MGB owners actually understand that valve spring #
selection should be matched to the intended engine operating RPM range.
Camshaft dynamics require that low engine speeds be avoided when using heavier
than stock valve springs. That is why camshaft regrinders require that you run
the engine, upon initial start up, at 2000 to 2200 RPM for 20 to 30 minutes to
aid with the break-in period of the camshaft lobes.
I have a set of valve springs used for racing which come with a
warning that the engine should not be operated below 3000 RPM.
ENGINE BLOCK PREPARATION
Prepare your engine block as you would do in a normal engine
rebuild, however, have your machine shop bore down the pushrod guide tube holes
(using a cutter the same size as the hole) to within 1" above the lifter bore.
This will ensure that the pushrod spring will have sufficient wall clearance.
Also, the engine block deck height needs to be measured at the
front and rear then recorded per the illustration below. OEM engine block
height was approximately 9.941"
Where possible, but not essential, all the camshaft lobes should
be of the same heel diameter within .005". Reground camshafts are notorious for
having large variances in heel diameters. For your information the OEM camshaft
88G303 used from 65 thru 74 had a heel diameter is 1.120".
Our engine block deck height was recorded at 9.921" .020" below
the OEM 9.941"
We used the above information to establish where the lower pushrod
spring retainer groove should be machined on the pushrod to establish a seat
pressure of 20# at an installed height of 1.240".
In our situation we machined the pushrod groove 2.800" from the
base of the pushrod. Arriving at this measurement was very time consuming,
however, it established a base from which to work from for future AVT
adaptation to other engine blocks.
It is safe to assume that most MGB engine blocks are going to be
redecked during the engine rebuilding process. The amount of material removed
from the engine block directly effects the pushrod spring installed height. For
example: if we removed .010" from the surface we will have decreased the
pushrod spring installed height by the same amount, which then would increase
the seat #. When using the 105# rated spring we would see an increase of
approximately 1#, raising the seat # to 21#. So, for every .010" removed from
the engine block will increase the pushrod seat # by 1#.
Reduced camshaft lobe heel diameters have the adverse effect. Any
diameter smaller than the 1.120" OEM must be subtracted from the OEM diameter,
then divided by 2. Our reground camshaft measured 1.030" a .090" difference,
when divided by 2 we have .045" difference. Here in our example, if we had not
compensated for this difference, we would have a reduced seat # of 4.5#, for a
total of 15.5#. So, for every 020" reduction in lobe heel diameter will see a
reduction in seat # of 1#.
We understand that many MGB engine rebuilders are not too
concerned with the uniformity of matching camshaft lobe heel diameters. We
decided it would be more convenient to standardize where the lower retainer
groove is machined on the pushrod. All pushrods will now have this groove
machined 2.800" from the base of the pushrod.
Changes in the spring seat # will now be controlled by the height
of the upper pushrod spring retainer above the engine block deck. In our
previous example, to attain a pushrod spring seat 20#, we needed the retainer
to protrude above the deck by approximately .250".
What we have now done is to convert all the calculations back to
the OEM engine block deck height (9.941") and the OEM camshaft lobe heel
diameter of 1.120". We have also made allowances for the compressed cylinder
head gasket of .023".
Using this information, we would require that the upper retainer
protrude .214" above the engine block deck. Example: 214" minus .023" = .191"
multiplied by spring # 105 = 20#. For an installed height of 1.240".
The upper pushrod spring retainers will be supplied in sufficient
length for machining purposes. This would then become a simple matter of
measuring each individual retainer, without the cylinder head gasket in
position, at .214". Remember to mark the correct location (1 thru 8) of each
When assembling the lower pushrod spring retainer to the pushrod,
we recommend that the retainer be seated on the 2 split collars by inverting
the pushrod and, while the retainer is supported, tap the lifter end of the
pushrod with a non-metallic hammer.
Increasing rocker arm ratios, whether they be OEM rocker arms or
roller rockers, are established by reducing the center line distance between
the valve adjusting screw ball and rocker arm bush. This only holds true when
using OEM rocker arm pedestals.
The roller rocker arms that were used in the AVT system were
offset a further .085" at the valve side of the rocker arm. The rocker arm
pedestals were offset .110". This .025" difference had little effect on the
rocker arm roller to valve stem centerlines. We would have liked have offset
the roller rocker arm .110" but unfortunately, this .085" was the maximum
offset the rocker arm design would allow.
Prior to completing the installation of our AVT system, we checked
out various roller rocker arm ratios. Due to pushrod lean we found that the
maximum ratio we could use was 1.59:1 (1.443" X .905"). Using 1.59:1 roller
rocker arms over the 1.55:1 set would result in a maximum valve lift difference
of approximately 2.5%. However, this 2.5% would be applied throughout the
entire valve lift.
Depending on the CFM flow rates of the intake or exhaust valve,
increasing the valve lift maybe worthwhile.
Having spent around $1200 to develop this system for the MGB 5
main bearing engine, I have to be realistic as to the marketing potential of
the AVT system to the MGB owner. Since an investment in a 1.55:1 roller rocker
kit will set you back $375, we need to determine what other costs will be
incurred to purchase the rest of the kit.
With all the prototype production costs taken care of, we can now
estimate what it will cost to put together the remaining components based upon
a quantity of 10 sets. We will then be in a position to establish a selling
price to the MGB owner.
So, if you are interested in possibly purchasing the AVT system
let me know. Also, if you have any other comments I will be glad to hear them.