MKIIT: Limits of Stock Turbo
From: uunet!Rt66.com!cal (Cal Smith)
Date: Thu, 2 May 96 16:44:35 MDT
Subject: MKIIT: Limits of stock turbo
Kostas wrote:
>The stock turbo will NOT provide more than 1.1 say max 1.2 bar even with a
>open filter K&N cone type no air flow meter and an HKS turbo exhaust. I
>think this is it for the stock turbo regardless of the type of control you use.
I have been making 18.5 psi boost (just over 1.2 bar) at will for
more than a year now. I have not taken it further due to several concerns;
the turbo and control system are not among them. My limitations are:
1) accurate test gauge to know just how high the boost truly is (over
18.5psi, that is),
2) Octane sufficient to keep the engine running at higher pressures,
3) Fuel-metering issues (both the factory protection/enrichment and
injector/pump capacity),
4) Intake air temperature (the factory intercooler is not efficient enough
to get the charge temperature down to a *safe* level at higher boost
pressures).
Additionally I want to balance and blueprint my injectors before
going further. After researching injectors for a major car manufacturer and
seeing what the Turbo Magazine project 5.0 Mustang w/ twin turbos did, I'm
not taking anything for granted. If you didn't see what happened to the
Mustang, Chassis Dyno pulls indicated power was down to about half of what
they expected. When the injectors were flow-tested, they had a huge spread
between individual flow-rates. The O2 sensor would attempt to correct the
problem and some cylinders would run rich while others _lean_. One injector
was so reduced in flow capacity that the most the engine could output
_safely_ as about 230HP. After flow-matching, over 400HP worth of fuel was
safely useable.
It would sure suck to run 20 lbs of boost with three cylinders at
12/1 air/fuel and have one cylinder at 9/1. Bye-Bye engine. At this point,
the protection system is shoving so much fuel through the engine, I doubt
_any_ of the cylinders are lean in the least. But while overly-rich may be
safe, it ain't fast either.
I am running a straight-pipe exhaust with no cats, a SuperChips
BoostGraphic controller, and an HKS PowerFlow air filter.
Brad Franklin wrote:
>wow! 19 psi? I'm no turbo expert, but from what I undestand from the mr2
>turbo experts I have talked with, 19 psi is indeed a bit too much to run
>on a non-modified stock ct26 turbo. I believe 19 psi is way out of that
>turbos efficiency rating, it works way to hard and creates more air heat
>and problems than it does good.
"Way out" of the efficiency rating, eh? Now there's quantitative
analysis! (Just kidding, Brad. :) ). First, let me say that I am no lover
of the CT-26 turbocharger--it simply is the one that we all have stuck on
our MKIIT's from the factory so let's find out what it will really do. It's
free, so how good is it _really_? If we are going to spend $"X" on
upgrades, is it better to put the money into a new turbo or spend it
somewhere else?
Granted, a lot of turbo shops don't really care for the CT-26, but
some do like it. Most of the turbo shops build units that are overkill for
what is usable with pump gas/octane booster and the stock internals of the
3S-GTE. What needs to be asked is: given the limitations of pump
gas/booster and stock 3S-GTE guts, will the CT-26 suffice or does it become
a performance bottleneck?
As for the answer and the efficiency approximations, see the below
section where I deal with assumptions of the limits of the CT-26. At this
point, suffice it to say that no, the CT-26 is not a significant limitation
in our quest for more power.
Brad continued:
>Also I don't think it will stay at 19 psi
>at high airflow (hi rpms), it just can't flow enough air to maintain such
>a pressure differential.
Again, the numbers will be discussed in detail below but I wan't to
deal with this train of thought here and now. The compressor _can_ flow
about 32-33 lb/min of air; enough to power the _3.0_ litre 1987-1992 Supra
Turbo to 345HP at 13.5psi, according to HKS. They leave the OEM CT-26 in
place through stage 6 of their upgrade chart. Why does HKS upgrade the
Turbo at stage 7 when the boost gets bumped up to 15psi? The turbo has run
out of mass-flow capacity at that point: running the CT-26 at 15psi will
still yield 32-33 lb/min of flow and no power gains (and the turbine shaft
will be at or near redline). They must go to a bigger compressor to flow
more air and make more power.
The MR2 has 50% less displacement than the 3.0L Supra engine; to
achieve 345HP at 13.5psi boost would require a power-peak 50% higher than
the Supra (i.e. 9000rpm in the MR2 rather than 6000rpm in the Supra). The
dearth of affordable camshafts for the 3S-GTE pretty much eliminates this
possibility (rev-limits, injectors, accessory pulleys, etc. aside), so to
max out the flow capabilities of the compressor we must use as much boost
as we can stand to cram in more air at a lower rpm.
The tendency of any compressor is for the boost to droop at
higher-flow is natural. I postulate below, however, that the compressor has
enough cojones to sustain 19psi at maximum flow. The only reason the CT-26
unit might not carry 19psi up to 32-33lb/min flow would be if the exhaust
manifold or turbine housing/wheel needs to be opened up for greater flow
(so that the exhaust could sustain the "drive" forces needed to reach 19psi
boost). Simply put, the turbo must do more compressing (work) on the MR2 to
get the same amount of mass flow as on the Supra, and the additional force
to accomplish that work has to come from the exhaust.
Brad finished with:
>On a final note, I've also been told that trying
>to run over 17 psi on that turbo will cause it to overspeed, which is
>definitely bad for turbo life. I've read that if a turbo can overspeed
>enough, it can be very dangerous! (the analogy used was a grenade going
>off in your engine bay...think about all the kinetic energy stored in a
>turbine spinning 150,000 rpm!).
Yes, overspeed is bad and can ultimately result in catastrophic
failure of a turbo (read: BOOM!). These devices have internal power
approaching the flywheel power of the engine so we are talking the
potential release of a _lot_ of energy in a short amount of time.
Having said that, realize that turbo shaft speed has nothing to do
with psi of boost, be it 12, 15, 17, 20, or 100! Shaft speed is a function
of PRESSURE RATIO and mass-airflow. Consider for a moment the 3S-GTE at
about 4000 rpm: the CT-26 rotating assembly spins at the same speed at
14.7psi boost as it would at 11.1psi boost at 7500' altitude. The only
difference is the efficiency: how hot the compressed air is when it is
discharged from the turbo compressor vs. temperature of the ambient air.
In both of the situations above, the pressure ratio = 2.0; that is:
( Boost psi + ambient psia ) / ambient psia = Pressure ratio
where psia = psi absolute or ambient pressure in psi. If atmospheric
pressure were 30 psi, the CT-26 would be able to make 30psi boost all day
long with no problem (up to the mass-flow capacity of the compressor).
Likewise, if ambient pressure was 4psi, the CT-26 would have a real problem
making 12psi boost (at any flow rate).
Running any turbo near, at, or above its designed pressure-ratio
limit will shorten the life compared to running it at, say, half of the
designed limit. That is the nature of imperfect mechanical devices. So what
we have do figure out is _what are the CT-26 design limitations_? Thanks to
the lack of information provided by Toyota we have to derive it for
ourselves:
Almost all turbo compressor designs are designed to safely run at
2.6-2.8 pressure ratios for long periods at or near the mass-flow rating of
the unit. Assuming that the CT-26 falls into the above category (after much
consideration I believe it does), this translates into 23.5-26.5psi as a
maximum boost at sea-level (14.7psi ambient) and 17.8-20psi at 7500'
(11.1psi ambient). 17-20 psi should therefore not be a problem at virtually
any altitude, with more boost possible at sea level.
Remember, too, the amounts of time we are spending at these boost
levels. Maybe 10 seconds per 1/4 mile if it is a really good run. If we are
on the highway and decide to play around with the limits of the vehicle,
30-45 seconds if the road is really clear. Consider how many seconds/day
you spend _above 15 psi_ and I bet the time doesn't amount to much over the
course of a year. Even at 45 seconds/day, we are talking about 4.5 hours
per year above 15psi (and even then we are within the design spec of the
compressor as long as the above limits are observed). Road racing is
another issue, where a substantial amount of time is spent at max boost, as
a result they replace their turbos on a regular basis. Anybody spending
that much time on the wastegate is probably not using the CT-26.
Let's talk efficiency: Even a 100% efficient compressor will heat
the intake air by the very process of compression. The most efficient
automotive turbocharger commercially available peaks at about 78%
efficiency, and most really good auto turbos peak out at about 72-75%.
Peaks of 65%-70% are still pretty good, and most compressor maps are drawn
out down to about 50% efficiency. The compressor still functions well as
the efficiency goes down below 50%, the air just gets heated a _lot_ more
and is less dense than it would be if the unit were more efficient.
Even when the compressor is 100% efficient, at 15+psi an
intercooler is necessary to keep intake charge temperatures below
preignition (detonation) levels. Once an intercooler has been added, its
efficiency far outweighs that of the turbocharger compressor in terms of
relative importance. This works in favor of the CT-26, as an increase in
intercooler efficiency yields much more power and lower temperatures than
the same increase in compressor efficiency alone.
Why don't we run a model of the CT-26 Turbo/3S-GTE Engine/SW-20 Chassis?
Given: 20psi boost, 14.7psi ambient, 100 degree F ambient, 6000rpm engine
speed @ 95% volumetric efficiency, Stock intercooler efficiency at 60% and
a pressure drop across the intercooler of 1.5psi.
Now, let's assume a worst-case scenario with respect to CT-26
compressor efficiency: 40%. The CT-26 should be more efficient than this at
this flow-range and pressure-ratio (somewhere between 45 and 55%), but just
to show how dominating the effects of the intercooler improvements are vs.
compressor improvements, it will suffice.
For comparison, let's assume HKS or GREDDY (or whomever) has a
turbo upgrade that is 75% efficient at this flow-rate and pressure-ratio
(75% is generous unless the turbo is way too big for good throttle
response. 60-65% is more likely). Also, consider the GREDDY intercooler
which, according to test data provided by the company, has an efficiency of
about 89% and a pressure drop of about 0.7psi at 20psi (The efficiency
figure of the intercooler seems high to me, but it came from valid test
data...).
Now, for the moment you've all been waiting for...
Turbo Intercooler Mass-Flow Intake-Charge Temperature
1) CT-26 stock 24.5lb/min 254 degrees F
2) HKS, etc. stock 27.3lb/min 182 degrees F
3) Ct-26 GREDDY 29.8lb/min 142 degrees F
4) HKS, etc. GREDDY 30.8lb/min 123 degrees F
To convert mass flow to horsepower, multiply mass-flow by about 11. Each
engine is a little different, but that accurate enough for DTR (DeskTop
Racing :-) ).
Note that the $1000 intercooler kit with the stock turbo (example
3) outperforms the $2000-3000 Turbo upgrade with a stock intercooler
(example 2). Also, adding the $2-3K turbo to the $1K intercooler (example
4) only increased power 3% vs. the $1K intercooler alone (example 3).
Remember too, that these are pretty conservative figures--the CT-26
efficiency has been fudged to look worse than it really is and vice-versa
for the aftermarket turbo. The point I'm trying to make is that a change of
intercooler makes a much bigger difference than a change of turbo.
Here's the same test but at 15psi boost:
Turbo Intercooler Mass-Flow Intake-Charge Temperature
5) CT-26 stock 21.8lb/min 223 degrees F
6) HKS, etc. stock 23.8lb/min 166 degrees F
7) CT-26 GREDDY 25.7lb/min 134 degrees F
8) HKS, etc. GREDDY 26.4lb/min 118 degrees F
Note several things here:
A. Example 5 is also the HKS Stage 3 upgrade. They say 247HP and my
prediction yields 240HP. Like I said, close enough for DTR.
B. Example 5 = 262HP predicted, and Example 2 = 300HP predicted. I
didn't run the model at 17psi (HKS stage 4 adds their turbo upgrade kit
running at 17psi) but extrapolation between the two above models to 17psi
predicts 281HP. HKS says 286HP.
C. Note the intake-charge temperatures: Example 3 is lower than
both examples 5 and 6. Add the Greddy intercooler and crank the boost to
20psi and your intake-manifold temperatures will be lower than either the
stock turbo or an aftermarket turbo at 15psi with the factory intercooler.
Can you say less detonation? I knew that you could. ;-)
D. HKS did not offer an intercooler upgrade because, according to
their engineers, the stock unit was pretty good and an aftermarket unit
would not add much performance at these boost levels. In that conclusion I
think they were mistaken. Even at the HKS Stage 3 level (15psi), Example 7
yields 283HP predicted, vs. 240 predicted with Example 5. The GREDDY
intercooler reduced intake-charge temperature 89 degrees F vs. the OEM
cooler, too.
Heck, in comparison to the HKS Stage 4 kit, you get equivalent
performance without having to buy a new turbo, intake-manifold temperatures
are reduced over 40 degrees F, and the CT-26 is running at 15psi vs. 17psi
for the aftermarket unit, reducing stress on the turbo and engine and
increasing the life of each.
E. All of these power predictions and quotes from outside sources
are at the ENGINE and not at the WHEELS. Wheel horsepower is less than
engine horsepower because of frictional, slippage, rotational, and drag
losses.
In conclusion; in the MKIIT, here's the most cost-effective route
to high-performance mods:
1) High-flow exhaust
HKS, GREDDY, etc. or just a simple straight-pipe after the
catalytic converters. Note that the cats are optional ;) depending on local
emissions laws/enforcement and performance goes up with the guts of the
cats gone. Boost creep has not shown itself to be an issue up to 18.5psi
boost and can be dealt with if it arises.
2) High-flow intake air cleaner and high-performance ignition system
Intake filters abound, from the HKS element replacements which are
legal in SCCA stock-class competition, to various foam or paper-covered
velocity-stack type units which are sold by HKS, GREDDY, etc. The latter
units flow more air than the former, but the intake turbo whine is much
more subdued in the former. As to whether you want to hear the whine or
not; that's personal preference--I like it, you might not...
The ignition additions _are_ necessary to insure reliable, complete
combustion at increased boost levels. You will make more power than stock
by simply reducing the airflow restrictions and cranking up the boost. Add
the ignition in addition to these things and you will make even more power,
get better mileage, have a smoother and longer-lasting engine, and extend
the intervals between oil changes. Sources include: Jacobs, HKS, and MSD.
3) HKS Fuel-Cut Defencer (or similar unit from other manufacturer) and some
form of boost control
A control is needed to adjust the point at which the turbine
wastegate opens, limiting boost. The controls are used to divert or block
some of the pressure"signal" that would otherwise open the wastegate, thus
raising boost pressure. They range from the mundane bleeder T-valve (from
aquarium/pet stores) to complex devices like the HKS EVCIII. Cost ranges
from $10-$1000, and units are made by HKS, GREDDY, SuperChips, etc.
A Fuel-Cut Defencer is used to delay the onset of an
overboost-protection feature designed into the MR2T's ECU by Toyota. At a
preset mass-flow or boost point, the circuit engages and shuts fuel off at
the injectors when you try to make boost thereafter (turning the engine off
and back on resets the circuit, FWIW). The Fuel-Cut is a very good feature
to have, but it comes in so early (about 12psi) on the 3S-GTE as to limit
our fun considerably. The HKS FCD unit delays the fuel-cut circuit to
somewhere around 22-25psi boost (this has ramifications for nitrous users,
but those will be discussed at a later date), so we can have fun with our
raised boost without fear of fuel-cut. Manufacturers include HKS, GREDDY,
and SuperChips. The units are about $100 alone but are often bundled with
boost controllers (for obvious reasons).
4) High-efficiency, low pressure-drop intercooler
All the reasons expounded upon in this post illustrate that the
intercooler should be the next item upgraded. GREDDY makes the
highest-performance unit which is easy to simply buy and bolt-in. Spearco
also makes a nice, inexpensive unit which fits but you have to do the
fabrication/fitment of intake tubes yourself. HKS also makes a unit for the
MKIIT. It is hard not to go with the GREDDY at about 1K.
5) Vein-Pressure Converter
Any speed-density converter (of which the HKS VPC is the most
common and easiest to install) will get the restrictive mass-flow sensor
out of the way and raise power/reduce turbo lag. This mod is probably
essential to get the full benefit out of the intercooler or to get the
power up above 280HP. The HKS VPC is about $800. I don't know who else
makes a similar unit. I am getting speed-density from an aftermarket ECU
and thus doing an end-run around the VPC.
6) Nitrous-Oxide
A small kit, ala Kostas' 50HP unit will decrease turbo lag
tremendously and increase HP across the board without any internal engine
mods required. Any more than 50HP and you will probably have to do internal
engine mods as insurance (forged pistons, metal head-gasket, etc.). The low
installed cost of a nitrous kit make it the bang/buck leader over a turbo
upgrade
7) Turbo upgrade
A turbo upgrade will have a small impact relative to all the
aforementioned upgrades, but hey: it will yield 5-10HP and a lower intake
air temperature. It could mean running a couple more psi on pump gas
without detonation. Remember, however, that most of these "upgrade" turbo
kits will spool more slowly and your street performance will suffer for the
sake of race-only performance.
Any upgrades beyond this point are really beyond the scope of this
posting and require special attention. Things like full-synthetic oil in a
low viscosity, a clean air filter, and properly-gapped spark plugs are
always important and you should have these things already.
Boy, can I talk or what?
Cal Smith
From: uunet!Rt66.com!cal (Cal Smith)
Date: Mon, 6 May 96 21:42:28 MDT
Subject: TURBO: Greddy Intercooler specs
Firstly, sorry I killed Majordomo. It seems like the first ones to go are
always the least deserving...
Brad Franklin wrote:
>A lot of this is based upon the numbers given with greddy's intercooler..89%
>efficiency..wow..are these numbers actually proven?
As I am writing I am digging through my pile of car stuff...I know
it's here _somewhere_...YESSSSSSS! I didn't mark the date when I got this
data, but I think it was about 6 months to a year ago. The source was
Millard Cook of GREDDY. He had to search and search to come up with the
data; I think he even called Japan and got them to FAX the dyno results.
This came from an otherwise stock MR2T on a chassis dyno. Ambient
temperature was 70.0 degrees farenheit, and the humidity was low. The data
they took was:
Air temperature into intercooler: 262.4 degrees F
Air temperature out of intercooler: 89.6 degrees F
HP gain: about 22HP
The data needed to solve for intercooler efficiency are intercooler
temp in ( Tin ), intercooler temp out ( Tout ), and ambient temp (Tam ).
The equation is:
Intercooler efficiency ( ni ) = ( Tin - Tout ) / ( Tin - Tam )
which works out to:
ni = ( 262.4 - 89.6 ) / ( 262.4 - 70.0 ) = 172.8 / 192.4 = 0.898
or 89.8% efficiency. I guess I was just plain mistaken, listing it as 89%
rather than 90% or 89.8%.
How accurate are these numbers? It depends how much faith you have
in GREDDY's test methods. Until one of us ponies up the bucks to try it, we
aren't going to get better data. The one thing that bothers me about this
test is that it is really hard to simulate accurate airflow over the
intercooler face when taking dyno numbers.
An air-to-air intercooler's efficiency is _highly_ dependent on
airflow over the core. I understand GREDDY used something like a
squirrel-cage blower fan to blow air onto and past the intercooler. Their
airstream velocity may have been accutate--or it may not have been. It is a
very hit and miss art at this point in intercooler development, and only a
few companies (like SPEARCO) give you accurate face-velocity efficiency
curves. Even SPEARCO can't tell you what face velocity to expect in a given
installation; you simply take a guess of 10mph, lookup that value on a
chart, and then compare that to actual data taken in road-tests after the
installation is complete.
GREDDY's values could be right on the money for 60mph cruising. Or,
they could be values representative of what the intercooler "sees" at
110mph road-speed. I doubt the efficiency will be 90% in the 0-30mph
road-speed range--but that is what the intercooler fan is for anyway; low
vehicle-speed intercooler efficiency-boosting.
I am not really worried about whether the 90% value was derived
with a high face-speed or not. It is the best and easiest to install
aftermarket kit available for combined street and race duty. It will still
yield more performance gains that a turbo upgrade at less than half the
cost. Finally, I am investigating a heat-loss coating (sort of like "air
wetter" :-) ) for the intercooler exterior that should raise the unit's
efficiency across the board. For approximately $75.00 US and shipping, it
should make the GREDDY setup irresistable.
Cal Smith
P.S.- For those of you thinking about coating your stock intercooler; I
wouldn't worry about it. It is still pretty restrictive and the coating
doesn't help that a bit. The GREDDY is sooooo much bigger/better that it
will still be the performance champion far and away (of course; if you are
on a reeeeeeeal tight budget, it might be the best $100 you could
spend...after the FCD. Who knows, it might be worth doing just out of
principle and posting before and after results to the group).