compressor maps: the low down?

[ESEMES]

fellas (and ladies)

can someone help educated the 'less educated" here, please?


id really like to learn to understand/read compressor maps, but am struggling so far to do such.

Are there any easy-togoto- links, or sites that can help?

i apologize for my ignorance in advance

esemes

* Registered users of the site do not see these ads.

[Unabomber]

Turbo FAQ

[nhluhr]

For a crash course, lets look at this compressor map of the stock WRX turbo:



There are several features to understand but basically what this is showing is the relationship between air flow, pressure ratio, and compressor efficiency and it gives you a few more details like estimated rpm of the turbo, surge line, etc. Now, lets define some terms:

Absolute Pressure: Gauge pressure plus ambient pressure. For instance, at sea level at standard conditions, the air pressure is 14.7psia. The a in psia stands for absolute. if you see a pressure listed as psia, it means absolute pressure. If your car is boosting 14psi and you're at sea level, the absolute pressure in your manifold is gonna be 28.7psia.

Gauge Pressure: typically just the difference in total absolute pressure and atmospheric but what this really means is the difference between two things. Normally, a gauge is going to be immersed in standard atmospheric pressure (of 14.7psia) and measuring a relative pressure. The relative pressure (like that in your manifold) is the gauge pressure. It can be displayed as psig or just psi.

Pressure Ratio: The ratio of the airpressure coming out of the compressor to the airpressure going in to the compressor. If you are making the turbo create 14.7psi of boost, you end up with a total output pressure of 14.7psi PLUS whatever the absolute pressure going into the turbo is so the total pressure ratio is:

PSIatmospheric+PSIboost/PSIintake

Assuming there is zero pressure drop in your intake getting to the turbo (which doesn't happen in real life), the pressure ratio would look like this:

(14.7 + 14.7) / 14.7 = 2

but in reality, you almost always have a pressure drop of 2 psi in the intake which makes your pressure ratio look more like this:

(14.7 + 14.7) / (14.7-2) = 2.3

The reason it looks like this is because if you are boosting to 14.7, that means 14.7 relative to the atmosphere so the boost control system is going to be making the turbo create enough rise in pressure to compensate for the intake pressure drop, thus incresing pressure ratio. Important to note here is that turbos don't care what the intake pressure is. They operate on pressure ratios. If the pressure going in to the turbo is already at 30psia and it's set to operate on a 2:1 PR, then the pressure coming is going to double to 60psia. This kind of sequential turbocharging is useful only for diesel trucks and drag racers.

Compressor Efficiency: Relates how much of the work being done on the air by the compressor actually compresses it vs how much it heats it. If the turbo is 70% efficient at a given airflow and pressure ratio, that means that 70% of the work is being used to compress the air and 30% of the work is just heating the air. It is important to note that one turbo is not inherently more efficient than another turbo but their ranges of efficiency may well be in a different CFM/PR range and thus better for a given application. Pretty much all the turbos you can buy these days will have sufficiently well designed comrpessors that you don't have to worry about what brand it is, but only whether that turbo's compressor map is well-suited to your application.

Surge: when the pressure ratio approaches a level that is too much for the airflow through the turbo, air begins to reverse its direction of flow and thus "fall back through" the compressor. This is called stall and you typically find this happening in a repeated loop called hysterisis. Once this state is happening, you have surge. Surge is VERY DAMAGING to compressor blades and can destroy a turbo quickly if it is severe. For this reason, you can't ask a turbo to make too much boost at low rpm (where the cfm is low) or else it will surge.

CFM CFM on the compressor chart is how much air is flowing through your engine in its uncompressed state. A reasonable rule of thumb is that it may take about 4 CFM to make 3 HP. So if you're shooting for a 300hp engine, you need to flow roughly 400CFM of uncompressed air. This helps you figure out what range you'll be operating in on the compressor map. You can see that pushing 400CFM through the little WRX turbo is not gonna happen efficiently. If you plot any PR along the 400CFM line, you can see that you're outside of reasonable efficiency ranges and probably overspinning the turbo, asking for trouble. Solution? Look at a bigger turbo

Does this help a little?

[nhluhr]

Here is a lot more technical information, and they take you through the process of performing the calculations to determine what you need: http://www.gnttype.org/techarea/turbo/turboflow.html

Unfortunately, hardly any of the turbos we are offered for WRX's have compressor maps readily available.

[ESEMES]

great stuff, fellas..

if im reading this right, the sr55 will be capable of making 400 ieffeciently?

[nhluhr]

well you have to convert the air flow on that from lbs/min. I'm not too sure of the conversion on that right off hand, but you can probably find it on google.

[NWWagon]

10lbs/min is good for about 100hp give or take some. To convert CFM to lbs/min mutiply the CFM by .069 to get your lbs/min so.... stock Turbo 14.7psi 360cfm is 24.8 lbs/min so around 250hp on a stock turbo.

[bboy]

A much glossed over aspect of a compressor map are the speed lines that run (generally) perpendicular to the "efficiency islands". While may people say that such and such compressor is more efficient than another, they don't really say how this is measured.

The compressor has two notable efficiency, one is the efficiency with respect to adiabatic heating (the heat of compression). When reading a compressor map you will see little number is every other "island" that say 75% or 68%. These are the efficiency of compression with respect to the adiabatic heat. The turbo is said to achieve 78% of the adiabatic heat. In a perfect compression (the closest would be a syringe) a certian amount is created as the molecule are force closer to one another. Compression without any heat creation is called isothermal. Everyone pays attention to the "compression efficiency" because its important to operate turbos when they compress air efficiently. The reciprocal way of viewing efficiency is to say that outside of these "island" the turbo is either not compressing air or it is heating the air "more than" it is compressing it.

The second "efficiency", the overlooked one, is there on the compressor map in the speed lines. Some compressors are capable of compressing air at a lower speed than others. If a compressor can compress air 2 fold (pressure ratio = 2) at 100,000 RPM while another similar compressor (similar in the sense that their maximum air flows at the highest RPM close to equal) can compress air 2 fold at 120,000 RPM, then the first compressor is 20% more efficient because it can compress the same amount (2 fold, or PR=2) at 20,000 lower RPM. 20% is a major gain in "efficiency" and amounts to one step larger turbo. In another thread I posted compressor maps of the Mitusbishi 20g wheel versus the presumptive FP Green wheel (T04E 50 Trim). The 20g-TD06H and the FP Green are identical turbos, except for the compressor wheel. The Garrett T04E 50 Trim wheel in the green compresses air to a PR of 2.2 (about 17.5 psi of boost) with 15% fewer RPM than the 20g. I would say that the Green's wheel is 15% more efficient over most of it most "efficient" range (i.e. the highest adaibatic compression efficiency islands).

Compressor maps are hard to read because they are 4 dimensional maps: PR, Air Flow, Compression Efficiency, and Shaft Speed. The APS site has a few 3D compressor maps and that is a good way to grasp the "efficiency balloons"--not the 2D "islands".

Compressor maps are a great tool for comparing compressor wheels in similar or the same housing. Put the same wheel in a smaller or larger housing the the map will look similar is shape, but it will be shifted from left to right on the Air Flow axis.

Turbine housing and wheels seem more unpredictable, generally bigger more flow, but the combination of wheel and housing can produce unpredicted results. I don't think anyone knows very much about how the exhaust moves through the turbine. Garrett publishes some great turbine "efficiency" graph that depict a line that asymtoptically reaches a maximum flow rate. Beyond this flow rate no more gas goes through the turbine, and must be shunted off (waste gated). If we had more flow testing of the kind that Garrett depicts (BTW we don't know how these measurements where made--for instance what was the temperture of the gas moving through the turbine) we could begin to make better guesses as to the appropriate turbine (hot-side) to select for out 2 and 2.5L engines. I've never seen another manufactuer even hint that they have flow numbers for thier turbo's hot-sides.

Hot sides matter a lot, but are often ignored. They basically controll when a turbo comes on boost (except as noted above where more RPM efficient wheels can boost earlier), becuase the hot side spins the shaft up to speed. The Mr. Hyde of the turbo turbine is back pressure. The more the exhaust builds up behind the turbo the more energy is expended to push it out. This is lost horsepower and is to be avoided. Wastgates can help, but sometimes even the wastegate plus the turbine cannot keep up with the flow rate. The result is the exhause back pressure quickly rises and kills torque and thus power. The stock STI VF39 turbo is a prime example of this. Port all you want, but unless you put a cannon ball through there its not going to flow enough.

This will be the most controversial thing I write tonight. Subarus seem to have a particular problem with exhaust back pressure (EBP)more so than say Evo's or DSM cars. The hot-sides of turbos need to be larger on Subarus than there inline 4 cousins. I believe, base on almost not data, that the reason for EBP problems on Subarus is the long exhaust path of the horizontal block and the convoluted path to the turbo charger high in the engine bay. If a shorter path were available, we could use smaller hot-sides on our turbos. The down side of large turbine housings is slower spool up. A shorter exhaust path is not reality in the bolt-on world, so why cry about it. I think there are two solutions, size the hot-side larger for subarus, or run the exhaust hotter. I say run the exhaust hotter, because I believe that it is the cooling of the exhaust gas that is giving rise to the EBP. Cooler gas is less fluid and clogs the uppipe and turbine with thicker gas. This is a testable hypothesis, by just tuning to higher EGTs at high RPM (typically EGT is low with subaru's rich anti-det tuning). Secondly, measure EBP with a gauge tapped into the uppipe (I have yet to find a pressure gauge that can withstand high temp exhaust, I continue to look).

Diatribe complete, flame me at will--I've got thick skin. Sorry, I got a little off topic.

[SaabTuner]

Don't forget also that the "boxer rumble" so coveted by Subaru enthusiasts is the result of several exhaust-gas pulses "cramming" into the turbine housing at once. If the flow were more evenly spread out, and if the manifold/header were sufficiently thermally insulated, the exhaust housing could probably be smaller.

The length of the headers/manifold will always present some amount of trouble; it's much like a larger aircraft has different aerodynamic forces. (Has to do with Reynolds' Number, and more specifically, the characteristic length when calculating the number.)

Don't forget as well the role that the speed of sound has to play in all this. Once the flow of exhaust gas out of the engine is such that the minimum cross-section of the exhaust housing has reached the local speed of sound (roughly 1,500 mph @ 1,600*F) the volume per second through the orifice will reach a limit. Any additional mass of exhaust must be further compressed to pass through after that point.

Keeping the speed of sound high by keeping the exhaust gas hot will raise the flow-limit and reduce back-pressure.

With respect to Adiabatic Compression and turbocharger efficiencies ...

The heat of compression is sometimes difficult to understand. What you must realize is that the temperature of a substance is the average energy of each of the mollecules within. In the case of a fluid like air, the energy is represented by the average velocity of the particles within. As the fluid is compressed some velocity is imparted into the fluid from the walls which are shrinking inwards towards it at some velocity. That movement inwards increases the average velocity of the particles in the gas, and thus their temeprature. The average speed of the particles within a gas is more or less the speed of sound in that gas. Or at least closely representative of it.

The case of the turbocharger is more complicated. From first glance you might think the HUGE accelleration done by the compressor wheel should heat the gas way more than say a cyllinders direct mechanical compression, no? But the actual SPEED of the particles inside the gas is already at the speed of sound. When the wheel spins each particles vector componant is alligned in the direction the wheel is forcing the gas. The sum of all the vector componants is still more or less the same (we'll get to that difference in a second) but they are now pointing towards the outer edge of the compressor wheel. When the particles hit the diffuser section their vector componants are forcefully mis-alligned and "diffused". (Which is why it is called a diffuser section.) The diffusion of the gas is not perfect, and they do not lose all of their initial velocity, therefore they have a higher velocity, and since velocity is heat, they are hotter than they originally were. Since some of this heat is due to friction between the gas, the gas is hotter than the increase in pressure alone can account for. This is what causes the less-than-100% efficiency.

A 100% efficient compressor would heat the gasses such that the increase in average velocity of each particle is equal to the total force required to compress the gas divided by the average momentum (read: ratio of specific heat) of the gas.

For n00bs who just want a calculator to do it for them (or lazy people like me) here's a lovely one: http://www.stealth316.com/2-turbotemp.htm

This stuff is why vector sums were so important in Multivariable Calculus. Bet you wish you'd payed attention now!

Adrian~

[Physics Junkie]

http://bimmerforums.com/forum/showthread.php?t=258035

Bimmerforums FAQ on how to read compressor maps

[SilverSurfer04STi]

This is a great thread. Thanks all.

[ESEMES]

a big THANKYOU to the above participants.

this is solid old info.

im very grateful to have it too, as i am replacing my turbo very soon (hopefully).

thankyou again; i have ALOT of reading (and RE-reading!) to do, and alot of great sources for getting it.

you guys REALLY know your stuff.

happy holidays

esemes

[bboy]

I love you Adrian!!

Even at my geekiest you take it another step, multivector calculus.
Guys (gals) listen and learn, the man knows his physics. He's the only guy I trust to talk about boundary layers and sonic disruption.

Hey Adrian, I'm hoping for the Hydra this week. I can finally test out some of the pre-turbo WI stuff to see what happens over the holidays. I'll post it up on the WI board. Hope you are well.

[mbiker97]

Nice write up guys.

[Tobey]

Originally Posted by mbiker97

Nice write up guys.

ditto.

[aaronyoung]

Originally Posted by nhluhr

well you have to convert the air flow on that from lbs/min. I'm not too sure of the conversion on that right off hand, but you can probably find it on google.

cfm*.0691 = Lb/min

[midnightsti]

wow, that had some wonderful info.

[rottenspam]

This stuff is why vector sums were so important in Multivariable Calculus. Bet you wish you'd payed attention now!

Adrian~

Yeah he's right this crap does come back to haunt you! I'm finishing up a MS in Systems Engineering, and Matlab is my friend for vector calc.

[bboy]

I use lbs/min / 0.075 to get CFM.

A cubic foot of dry air weighs 0.0807 lbs at sea level and 25C.

Air is never dry. At 75 F (~25C) 50% saturated air contains 0.001 lbs water/CF or about 0.5 grams/CF. This means about 1% of the weight of the air is water, changing our value to 0.080 lbs/CF

Most air is greater than 50% saturated with water at any given time, and most of us are not driving at sea level all the time. I picked up the 0.075 figure somewhere along the way. Anywhere in the range is going to give you a decent estimate, after all we have no idea what the turbo manufactuers used as standard temp and pressure.

[SixSick6]

68º for everywhere besides sea level at which point 80º is used.

[flycaster]

Excellent posts, guys. Well done.

[SaabTuner]

Damn. I could have worded some of that a LOT better. In the interest of making sure this thread remains usefull (I hope), here's something that might come in usefull to someone searching for a thread like this:

The speed of sound in a gas = sqrt (gamma*R*T) where gamma is the ratio of specific heats (1.4 for air at STP), R is the gas constant (286m^2/s^2/K*) and T is the temperature of the gas in Kelvins.

The ratio of specific heats must be used for a gas because a change in temperature from a constant pressure test results in a different requisite energy than a change in temperature in a constant volume test. For a constant pressure test, the volume of the container in which the test is being carried out must expand to lower the density and maintain a static pressure.

Good page on the ratio of specific heats here at NASA's Glenn Research Center website: http://www.grc.nasa.gov/WWW/K-12/airplane/specheat.html

Maybe I'll think of something more usefull later ...

-Adrian

[ride5000]

Originally Posted by bboy

The second "efficiency", the overlooked one, is there on the compressor map in the speed lines. Some compressors are capable of compressing air at a lower speed than others. If a compressor can compress air 2 fold (pressure ratio = 2) at 100,000 RPM while another similar compressor (similar in the sense that their maximum air flows at the highest RPM close to equal) can compress air 2 fold at 120,000 RPM, then the first compressor is 20% more efficient because it can compress the same amount (2 fold, or PR=2) at 20,000 lower RPM. 20% is a major gain in "efficiency" and amounts to one step larger turbo. In another thread I posted compressor maps of the Mitusbishi 20g wheel versus the presumptive FP Green wheel (T04E 50 Trim). The 20g-TD06H and the FP Green are identical turbos, except for the compressor wheel. The Garrett T04E 50 Trim wheel in the green compresses air to a PR of 2.2 (about 17.5 psi of boost) with 15% fewer RPM than the 20g. I would say that the Green's wheel is 15% more efficient over most of it most "efficient" range (i.e. the highest adaibatic compression efficiency islands).

bboy, how do you figure that slower shaft speed translates directly into overall (ie thermal) efficiency?

[TheMadScientist]

Originally Posted by ride5000

bboy, how do you figure that slower shaft speed translates directly into overall (ie thermal) efficiency?

They are not related. Look at a compressor map, the speed lines cut through different eff islands. A common compressor(Small16G) can run a shaft speed of [email protected]@65g/sec and be on the 65% island, then run that same compressor at [email protected]@200g/sec and run on the 76% island.
So that blows the shaft speed theory away.
My data:
http://www.stealth316.com/images/td05-16gsmall-raw.gif

TMS

[duncangrant]

Originally Posted by SaabTuner

Don't forget also that the "boxer rumble" ... is the result of several exhaust-gas pulses "cramming" into the turbine housing at once. If the flow were more evenly spread out ... the exhaust housing could probably be smaller.

Key point!

Originally Posted by SaabTuner

Once the flow of exhaust gas out of the engine is such that the minimum cross-section of the exhaust housing has reached the local speed of sound (roughly 1,500 mph @ 1,600*F) the volume per second through the orifice will reach a limit. Any additional mass of exhaust must be further compressed to pass through after that point.

Keeping the speed of sound high by keeping the exhaust gas hot will raise the flow-limit and reduce back-pressure.

{and in a later post}
The speed of sound in a gas = sqrt (gamma*R*T) where gamma is the ratio of specific heats (1.4 for air at STP), R is the gas constant (286m^2/s^2/K*) and T is the temperature of the gas in Kelvins.

So for a given mass flow with higher egt the volume flow increases proportionally with temp but the speed of sound increases as the square root of temp (if I understand you right) and the max attainanable volume flow increases directly with speed of sound or sqrt of temp.

If I've understood you correctly then I don't understand how increasing temp improves the maximum mass flow.