Selecting a Compressor

[Mafdark]

I noticed that the stickies at the top were missing some more technical information which can help in selecting a compressor. This writeup was done by a friend of mine, its a perfect walk through for beginners who want to get a better understanding of some of the more technical aspects of compressors and compressor maps. When I stop being lazy I'll post up my own map for my GT28RS and then the GT2871r that I switched to.

[Mafdark]

Before you start peeking through the Garrett catalog or play Add to Cart with a multi-thousand dollar turbokit, you should know what you're buying and how it's going to work. Moreover, you should know whether or not it'll mesh with what your plans are.

First things first: Kill the ricer within you now.
Everyone wants a strong motor, but a bigger turbocharger is NOT necessarily better unless you know how to select it.
We will use high-school algebra to figure this all out, so grab a beer and start reading.

First, let's look at a compressor map from something smaller, such as a Garrett T04E-40 trim.

http://not2fast.com/turbo/maps/t04e-40.gif

In case you want to follow along with a different compressor, here are a bunch more maps to play with on this website:
Not2Fast: http://not2fast.com/turbo/maps/

I can already see some of you squirming behind your monitors. Relax. If you had the mental capacity to beat Super Mario World, you can nail this.

Before we start, take a good look at the graph. The island shape in the center is the region of values that this compressor is capable of creating. We'll learn how to find out where we are on this map, given a certain set of conditions that most of us usually see while driving. From this map, you'll be able to figure out the minimum RPM a compressor can support, the maximum RPM it can sustain a certain pressure, how much horsepower you're going to make and how hot the intake air is going to get, both before and after the intercooler. These bits of information are enough to know what parts you'll need and how fast you're going to go.

I'll break down the compressor map into four sections:

* Compressor Efficiencies: These are the concentric "islands" within each other on the map. The more to the center of the map you go, the higher the compressor efficiency. Efficiency means the compressor heats the incoming air less. Compressor maps will usually give a general value for each efficiency island, such as "65%" and so on.

* Compressor Wheel Speed lines. These lines run across the map and give you an idea of how fast the compressor is spinning when you are in a certain efficiency zone. Measurement is in RPM, so yes - 98,000 is really an RPM value. Most compressors spin at least 10 times as fast as your vehicle's crankshaft. Start thinking about using Synthetic Oil.

* Surge and Choke Zones. (No-Man's Land). Outside of the map, the left side represents the surge region. Surge means your engine is not producing enough airflow to operate the compressor at a certain pressure ratio. The right side is the Choke Zone, or overspeed area. This is where the compressor will spin well beyond its designed maximum speed, but never create the pressure you're looking for. In short, these are bad regions. You do not want to be there. If you find yourself outside the map, find another compressor. That's the core of this whole article.

* X and Y Axis. These axes represent Airflow (in CFM, lb/min or CMS - Cubic Meters per Second) on the X-Axis and Pressure Ratio on the Y-axis.


Pressure ratio is simply the ratio of what boost you're planning on running, with respect to the current atmospheric pressure. At sea level, that pressure is 14.7psi. At higher altitudes, it goes down. If someone is running 1 bar of boost at E-town in Jersey, they're running about 14.7psi. If they start drag racing in Denver, 1 bar of boost is 12.2psi. However, both of them are running a Pressure Ratio of 2.0. I'll explain this shortly.

Airflow is a measurement that's easy to understand. It can be measured by volume or by weight. You can have air at normal, sea level pressure, which can be measured by how much fits in a single cubic foot. Alternately, you can measure air by how many pounds pass through a certain region per minute. Compressor maps can use both of these values, depending on the map. It is important for us to know what altitude the compressor is working at, because air is thinner and lighter at higher altitudes and our values will change.

Engines create vacuum when they run, even if they have a turbocharger bolted on. When you boost, the 15psi you see on your boost gauge doesn't mean your turbocharger created 15psi worth of work. In reality, most stock engines will pull in the ballpark of 20 inches of vacuum at idle. That means your turbocharger has to compress the vacuum to zero, then keep working until the positive 15psi registers on your gauge. Mathematically speaking, the compressor has put nearly 30psi-worth of work into that airflow. This is known as "Absolute Pressure." Absolute Pressure is like driving 10 miles south of your house, then turning around and driving 20 miles north. You may only be 10 miles north of your house when you're done, but your car drove 30 miles to do it. Absolute Pressure is heavily affected by atmospheric pressure. We must determine how hard the turbocharger has to work, so let's take altitude into account by using the Pressure Ratio:

Determining the Pressure Ratio for 15psi of boost at sea level [0 ft] looks like this:
Pressure Ratio = (Target Psi + Atmospheric Psi) / (atmospheric psi)

Therefore:
Pressure Ratio = (15 + 14.7) / 14.7 = 29.7 / 14.7 = 2.02

Had you been in Denver, you would've replaced all the "14.7's" with "12.2's". If you still want to run 15psi in Denver, things change considerably. You can see the turbocharger runs a higher pressure ratio (2.23) to run the same boost you did in E-town where your pressure ratio is 2.02. That means it's spinning faster to compress thinner air to the same pressure. What happens when something spins faster in an engine? You create more heat. Anyway, back to New Jersey.

Draw a horizontal line across the map, starting at the "2" on the compressor map. You can see it cuts pretty nicely right through the middle of the map. That's good.

On the X-axis (bottom scale), we must convert the Lb/Min airflow value (or Cubic Meters per Second, depending on the map) to Cubic Feet per Minute to work with our equations.
To convert Pounds per Minute or Cubic Meters per Second into CFM, you need to take the air temperature into consideration, since the ideal gas law tells us that the hotter a gas gets, the more it expands and the less it weighs per cubic foot. It's the same thing as watching how much cash your wallet holds during the holiday shopping season.
The typical compressor map assumes an 85 degree outside air temperature. Most maps also show the temperature formula on the bottom of the map, if you decide to fiddle with making your numbers absolutely exact.

SCIENCE: One Cubic Foot of air at 85 degrees weighs 0.07282 lbs.
Therefore, convert lbs per minute to CFM by multiplying by the lb/min value by 13.73.

SCIENCE: One Cubic Meter per Second = 2118.64 Cubic Feet per Minute, assuming above temperature.
Therefore, convert CMS to CFM by multiplying any CMS values on the X-axis by 2118.64

Now look at that line you drew on the map. The entry and exit points on this map correspond to 15lbs per minute and 35 lbs per minute of flow. Just look down at the X-axis to see what I mean. Multiply those two by 13.73 and you get 205 CFM and 480 CFM. That means if you want to make 15psi with this compressor at sea level, you will need airflow between 205 and 480cfm to do it. As you rev a motor, you increase the CFM of exhaust it spits out. So, in essence what you're learning now is "where does the turbo spool up and where does it start falling off boost at higher RPM's?"

Let's figure out a few things. An engine is basically a fancy air pump.
First, how big is the pump? (Displacement)
Second, how fast is it spinning? (RPM) Pick an RPM to find out what's going on at that speed.

Now plug that RPM into this equation:

CFM for 4 stroke = ((Displacement in Cubic Inches) / 3456) x RPM x VE

So for my 3S-GTE engine, my stock displacement is 1998cc's, or 121.9 cubic inches. At 6000 rpm, it flows:
CFM = (121.9 / 3456) x 6000 x VE = 211.6 CFM x VE

VE is volumetric efficiency, which is a percentage measurement of how much air ACTUALLY makes it into the cylinders on each stroke, compared to how much can THEORETICALLY make it. It's almost always less than 100% unless you've got serious mods or lots of money in headwork. For now, assume your average near-stock engine at 6000rpm has a 90% VE.

Therefore:
CFM = 211.6 x 0.9 = 190.5 CFM

Now, we already know we need 205 CFM for that turbo to make the boost we want. The good thing is, you've just calculated how much air the motor would move if it were a naturally aspirated motor.
Now you need to figure out how much it'll do under boost. That requires something called Density Ratio.

To calculate Density Ratio, we use our Pressure Ratio and figure out how much the compressor's going to heat the air up.
The equation is:
Temp Out (in F) = (((Temp In (in F) + 460) x (Pressure Ratio ^ 0.283)) - 460)

So 15psi of boost at sea level, on an 85 degree day looks like this:

Temp Out = (85 + 460) x (2.02)^0.283 - 460 = 205 degrees F

This assumes a 100% efficient compressor... which is ideal but not realistic. Take a SWAG (Scientific Wild Ass Guess) as to compressor efficiency by averaging the values of the map. On this map, 70% looks like a decent neighborhood to start with, so let's use it.

That makes our compressor's outlet temperature:
Actual Temp Change= (Ideal Temp Change) / Efficiency

For our example, the Ideal Temp Change is 205F - 85F or 120F:
Actual Temp Change = 120F / 0.70 = 171F

So the compressor is going to heat the air 171 degrees above the outside air temperature. Add 171 to 85 degrees and we get 256 degrees coming out of that turbo, going toward your engine. This is where an intercooler comes into play. Speaking of which, what happens when that air hits the IC?
First the temperature drops a bunch and second, the pressure drops a little. Your average pressure drop for a smaller high quality side mount IC is around 0.5psi. For a larger front mount it could be over 1psi. For the IC, we will assume a 65% efficiency, which is reasonable for a decent sidemount. For a larger front mount, you could assume perhaps 70 to 85%. If you're using water spray or fancy nitrous foggers on the IC's surface, you could increase efficiency up to 100%, maybe even more. The main factor is airflow. If your front mount is mostly blocked by the bumper and isn't properly ducted, it isn't going to work. If you're driving a car that looks like it has a metal mouth and none of the IC is blocked, it's going to work a hell of a lot better.
To determine efficiency, you can measure the temperature at the compressor's outlet and at the throttlebody inlet. If the compressor heats up the air 150 degrees above ambient, but the intercooler takes all the added heat away and drops it back to the same 85F temperature as outside, then you have a 100% efficient intercooler. In the end, you can usually ballpark your figure by assuming 65% for the average sidemount and 80% for the average front mount.

The formula looks like this:
T IC drop = (T IC in - T ambient) x IC efficiency
T IC drop = (256 - 85) x 0.65 = 111F

That means the IC drops the turbo's outlet temp by 111 degrees. That transforms our 256 degree temp into 145 degrees and drops the pressure from 15psi to 14.5. Don't forget the pressure drop. Pressure ALWAYS goes down when temperature goes down, all other things being equal.

So what does this do for our "normally aspirated engine"? Well, density of the air is increased by a certain ratio:
Density Ratio = ((Temp In + 460) / (Temp Out + 460)) x (Pressure Out) / (Pressure In)

Where "Pressure Out" is your Pressure after the IC, plus the Atmospheric pressure
And "Pressure In" is simply the Atmospheric pressure.

For example, we get:
Density Ratio = ((85+460)/(145+460)) * ((14.5+14.7)/(14.7)) = 1.79

That means you're packing 1.79 times the DENSITY of normal atmospheric air into the engine with this compressor and IC combo as you would if the engine were operating without any forced induction. Add the appropriate amount of fuel and you're off to make more power.

Go back to the original 190.5 CFM value we got. Multiply that number by our density ratio of 1.79 and now we have 341 CFM (or 24.8 lbs per minute). Look at the compressor map and you'll see that 24.8 lbs @ 2.02 lands you right in the middle of the "island" to make the boost we want. Now you know you're making 24.8 lb/min of airflow at 6000rpm at sea level on an 85 degree day, with 15psi of boost (pressure ratio of 2.02). Excellent!
If instead your calculations landed you outside the compressor's islands, you simply wouldn't be getting 15psi out of the compressor at that rpm. At lower RPM's, surging the turbocharger would sound like pops and backfires coming out the intake. At higher RPM's scale, boost would fall away as you kept revving the motor and the compressor would overspeed.
This is exactly what happens when smaller turbos tend to lose boost at higher rpm's.

The 341 CFM value falls within the highest efficiency range (dead center island) on the map, so that means the actual temperature at the throttlebody will be a little lower than we'd calculated and our density ratio a tad higher. It's close enough to give us a good idea of what we're working with. If you REALLY wanted accuracy, you could go back and redo the calculations with the new efficiency you've deduced, to get a more accurate CFM value. You'll find the values can change upwards of 5%.

Now the fun stuff: Approximating Horsepower Output

The basic crank HP formula is:
Crank HP = Manifold Air Pressure (in absolute psi) x Compression ratio x (CFM / 228.6)

The compression ratio for a Gen II 3S-GTE such as mine is 8.8, so we plug in the real numbers into our HP formula and get:
Crank HP = 29.2 x 8.8 x (341/228.6) = 383 HP

Toss in a rough 20% drivetrain loss and you'll have 306whp at 6000RPM.

Remember one thing. The Basic Crank HP formula is exactly that: Basic. It doesn't take into account all the frictional losses from oil, spinning engine accessories, octane limitations or piston friction. It has no idea what Base Specific Fuel Consumption is or how awesome your race-prepped head flows. It's simply a one-size fits-all equation. Because of that, it's reasonable to take away another 5% and claim it as "frictional losses" from the reasons above.

Frictional Losses = 5%, Therefore:
306whp x 0.95 = 290whp.

Given my exposure to this compressor's real-life results, not too far from what I've seen on built motors with good tuning.

Volumetric Efficiency isn't a static number. It's constantly changing as you rev a motor and modify it. Because the turbine and wheel will affect the volumetric efficiency's behavior, it will also put a limit on overall HP output. For example, my stock CT-26 turbine and its housing are so restrictive that it easily drops the engine's VE well below 90% at the 6000RPM we looked at. The compressor is capable of plenty of power, but the rest of the plumbing is not. A stock MR2 turbine housing mated to this 40 trim wheel would create a maximum of 260-270whp. That's close to 30 horses shy of what we'd calculated. Big difference, eh?
Turbine sizing and A/R really DO make that much of a difference. On a T3/T4 turbo using the same compressor wheel but a larger turbine outlet, you could be closer to 290whp at 15psi. Realistically, 290whp (based on the additional 5% frictional loss) is a reasonable expectation. Anytime energy is transferred, you're going to lose energy in friction, so don't get disappointed if your compressor needs you to pull out the stops and run racegas to get you where you calculated it could.

One other thing we should check now that we have the numbers, is whether the compressor will be forced into the surge line. Surge is caused when the engine cannot ingest enough air to keep the compressor inside its map and the symptoms we listed above, will happen. Surge KILLS your turbocharger's bearings, so it's something we want to prevent. We saw that at a 2.02 pressure ratio, the surge line is around 15 lbs per minute, or 205 CFM.

Let's mathematically rearrange the CFM equation with respect to RPM, to find out exactly how low on the tach we can make 15psi before we start backfiring out the intake and ruining our turbocharger. Let's use a higher VE of 95% due to the lower RPM range we're operating in:

Remember:
CFM for 4 stroke = ((Displacement in Cubic Inches) / 3456) x RPM x VE

Remember to first divide your CFM value by your density ratio of 1.79, then plug it into:
RPM = ((3456) x (CFM)) / ((Displacement) x (VE))

So:
RPM = ((3456) x (114.52)) / ((121.9) x (0.95))

If you did your calculations right, you'll end up with 3417 RPM.

Now you know this compressor won't make 15psi below 3417 RPM. The VE might be better than we assume, but don't play around when you're that close. Basically, compressor supports 15psi @ 3500rpm and that's it. 3500rpm is your surge limit and you can easily plug in your CFM value for the other end of the map and find out at what RPM it begins to choke. (Hint: it's 8002 RPM).

Now let's translate what we have learned into English. This is a compressor that will give you 15psi from 3500 to 8000rpm, on a 2.0L engine with a decent ebay intercooler, on a hot summer day, at sea level. Peak charge efficiency (the "center island" in the map), will occur in the ballpark of 5000 to 6500 RPM. This is a hot street turbocharger. It can hit 15psi just after 3500rpm, carry power past 6500, and wheeze out over 8k. This is the type of turbocharger that loves on-ramps, twisty mountain roads and autocross, because it supports a mid to upper range RPM band, yet gives you enough breathing room if you need to stay in gear a little longer for that slalom. For a drag racing application, you would want a turbocharger that is in peak efficiency closer to redline, with the widest possible efficiency island to support your rev range and with headroom for more boost at higher RPM's.

You can control spool on most turbochargers through the turbine housing's size, or "A/R." A smaller A/R would focus the exhaust energy more onto the turbine blades, encouraging faster spool at the expense of becoming a restriction at higher RPM's and pressure ratios. The larger A/R delays spool and allows exhaust to flow with less restriction and thus improve VE at higher RPM's. Another option is a turbocharger with an Anti-Surge compressor housing like a GT30R. This housing actually extends the compressor surge line more to the left, allowing more stable pressure at lower airflow numbers and helping make the compressor's behavior more flexible on response-based applications.

Now it's time for questions. Are these numbers what i want? Do I want more power? Do I want to have more power at a different rpm? Do I like where the surge line is? Is this amount of boost pressure okay or can I run more? Will my car always be at sea level or am I moving to Denver in a month? Can I actually enjoy this turbo or will i have to rev the piss out of it just to get the car moving?"

Now you can tailor your turbo to fit your needs.

Altitude Pressure List:
Sea Level -- 14.7 psi
600 feet ---- 14.4psi (Lebanon Valley Speedway)
1000 feet -- 14.2 psi
2000 feet -- 13.7 psi
3000 feet -- 13.2 psi (Taconic Trail Peak - Rt.2 into Mass)
4000 feet -- 12.7 psi
5000 feet -- 12.2 psi
6000 feet -- 11.8 psi
7000 feet -- 11.3 psi
8000 feet -- 10.9 psi

[dallasb84]

if i had a cookie id give you two

[CaTcH22]

Good info. I use these equations and formulas for choosing my turbos, and our customers turbos.

[Mafdark]

Thanks, all the compliments will be forwarded to my friend who wrote it. Its rare to find someone who has such a strong grasp on the subject but is also able and willing to write it up in a way that others can learn it too.

[SPOOLIN]

just a note:

A/R doesnt always affect spool. I went from a 1.0 to a 1.10 T4 large frame housing and it spooled faster. And then i moved to the 1.25 a/r and it didn't slow down at all.

[Mafdark]

just a note:

A/R doesnt always affect spool. I went from a 1.0 to a 1.10 T4 large frame housing and it spooled faster. And then i moved to the 1.25 a/r and it didn't slow down at all.

as the write up says:

You can control spool on most turbochargers through the turbine housing's size, or "A/R." A smaller A/R would focus the exhaust energy more onto the turbine blades, encouraging faster spool at the expense of becoming a restriction at higher RPM's and pressure ratios. The larger A/R delays spool and allows exhaust to flow with less restriction and thus improve VE at higher RPM's. Another option is a turbocharger with an Anti-Surge compressor housing like a GT30R. This housing actually extends the compressor surge line more to the left, allowing more stable pressure at lower airflow numbers and helping make the compressor's behavior more flexible on response-based applications.

I'd refrain from the typical "This happened to me so it must be true for everyone" type of stuff without providing the physics to back it up. In your particular case there could have been hundreds of variables which played a part it what you were observing or what you thought you were observing.

[SPOOLIN]

I'd refrain from the typical "This happened to me so it must be true for everyone" type of stuff without providing the physics to back it up. In your particular case there could have been hundreds of variables which played a part it what you were observing or what you thought you were observing.

Dude your not an instructor here so i won't refrain from SHIT. my results are Hardly a rare occasion for big turbo and small engines in my applications and type of racing... There are not 100's of variables when nothing changes but a housing. I don't know how much shit you read or how much ask Jeeves and google you love to use but ive owned and used more turbos and turbo configurations than you can't count on 4+ hands. You can throw a whole book out there but I've seen cars make spool increases from jumping from a little gt35r to a 72mm precision gt turbo. I've got all the inside turbo design contacts you could want... I'm not a fucking retard.


www.MSSRACING.com

[Mafdark]

Don't get all butt hurt on me. The write is an intro for people new to the subject and if you get into every single variable or tiny exception then its going to be tougher for people to learn the concepts. Regardless of what you think your special case was, the physics of the subject as explained in the write-up are dead nuts accurate.

[Sneezin]

good info,,,going to help ppl,,,but i don't want into that^^^^