Decoding compressor maps

[free5ty1e]

OK, I'm staring at this compressor map (last page of the '92 FSM posted on the board) for our turbos and am trying to make sense of it. I see that each line represents characteristics of the turbo at a certain... load, I'm guessing? The crossing lines represent how fast the turbo is spinning. Then the two axes are somehow pressure and volumetric flow?

How do I extrapolate from this the efficiency range, in PSI, of our turbos? Little help?

[vrg3]

From what I understand, you do have it right.

The vertical axis is pressure ratio. That is, the ratio of compressor outlet pressure to compressor inlet pressure. So, for example, if barometric pressure is 14.7 psia and boost measured at the compressor outlet is 6 psig, the pressure ratio is (14.7 + 6)/(14.7), or about 1.4.

The horizontal axis is compressor inlet (I think) airflow. It's usually measured either as volumetric flow (volume per unit time) or mass flow (mass per unit time). On this particular map, it's volumetric, in cubic meters per minute.

The "islands" drawn on the map are isolines along which compressor efficiency is the same. Each line is labeled with its efficiency value.

The horizontalish lines do indeed represent lines of constant compressor rotational speed.

The surge line on the left indicates when the compressor will surge. Any condition to the left or above that line will result in failure of the compressor to properly convey flow. There's also often a choke line on the right, past which the compressor chokes as air velocity exceeds the speed of sound.

It's not really clear to me how best to make use of the map. There are a few other significant variables: the engine's volumetric efficiency, pressure drop and temperature drop across the intercooler...

[free5ty1e]

Hmm... so the percentage values I see on the lines are not load but in fact efficiency? Interesting.

The surge line is an interesting thing... is this some sort of lower limit for the turbo's airflow? As in.. the minimum airflow necessary to actually spin the turbo or something? And a choke line sounds like the maximum, even though this map doesn't have one - as I doubt our turbos can force air to approach the speed of sound. Although that would be an interesting problem to have to deal with

Anyone ever made use of one of these things in the practical world in which we live?

[vrg3]

Yep. They're efficiency numbers. The term "load" doesn't really mean anything specific as far as I can tell, anyway.

I don't fully understand compressor surge, but I think it's generally just an unstable condition. Air doesn't flow smoothly along the paths in the centrifugal compressor; it just kind of slips through unpredictably. I don't know whether or not that corresponds with the wheel ceasing to rotate though.

I haven't actually seen a good clear analysis using a compressor map... but let's try a rough simplified one... I'm gonna kinda just ramble on in some math and see where it goes.

Let's assume we're at RTP:

Atmospheric pressure = 14.7 psi
Ambient temperature = 20 degrees C = 293 degrees K

Let's say we want to run 12 psi of boost.

Pressure ratio = (14.7 psi + 12 psi) / (14.7 psi) = 1.816

Let's see how hot the pressurized air would be if the compression were adiabatic (100% efficient):

Ideal outlet temperature = 293 degrees K * (1.816^0.286) = 347.5 degrees K

That means an ideal temperature increase of:

Ideal temperature increase = 347.5 degrees K - 293 degrees K = 54.5 degrees K

If we assume the compressor is going to be 70% efficient, then we'd get a higher actual outlet temperature:

Actual outlet temperature = 293 degrees K + 54.5 degrees K / 0.70 = 370.9 degrees K

Okay... Now, hmm. Density is proportional to pressure divided by temperature, so we can compute the density ratio:

Density ratio = 1.816 / (370.9 degrees K / 293 degrees K) = 1.434

Now, it's a 2212-cc motor with a 6500-RPM redline. Since it's a four-stroke, only half the motor actually goes through an intake stroke each rotation.

Ideal flow = 2212 cc * 6500 rpm * 1/2 = 7189000 cc/min = 7.189 m^3/min

Okay, and let's assume the engine's volumetric efficiency is 80 percent.

Actual flow = 7.189 m^3/min * 0.80 = 5.7512 m^3/min

If that's the volumetric flow at the compressor outlet, dividing by the density ratio should give us volumetric flow at the compressor inlet:

Inlet airflow = 5.7512 m^3/min / 1.434 = 4.01 m^3/min

Okay, that gives us enough to look up a point on the compressor map. A pressure ratio of about 1.8 and a volumetric flow of about 4.0 gives us a compressor efficiency of about 71% and a compressor speed of about 125000 rpm.

Now I guess we should take that 71% number back through the calculations to refine our guesses:

Actual outlet temperature = 293 degrees K + 54.5 degrees K / 0.71 = 369.76 degrees K
Density ratio = 1.816 / (369.76 degrees K / 293 degrees K) = 1.439
Inlet airflow = 5.7512 m^3/min / 1.439 = 3.997 m^3/min

Looking that back up in the map puts us in almost exactly the same position.

Hm.

Is that all right?

The thing is... even if that is all correct, it all goes out the window the minute you install an intercooler. An intercooler causes a pressure drop, so manifold pressure is less than compressor outlet pressure. Simultaneously, it causes a temperature drop, meaning higher density at the intake manifold or intake valves.

And if it's a top-mount intercooler, the amount of temperature drop (and consequent density increase) can vary widely depending on heat soak.

I dunno.

[free5ty1e]

Well, it was a valiant effort V. The compressor map makes a bit more sense now, and the math works out. But yes I can see how everything just gets entirely insane to calculate once an intercooler and other modifications are in place. Thank you though, as I no longer go cross-eyed when I look at the map

[NICO]

seirous stuff on this topic, gives me ideas.

[morgie]

i made a quick calculator 1 month ago.

http://www.clubsubaruquebec.com/calculp ... irflow.php

you can use it to give you an idea

[vrg3]

Interesting, morgie... I couldn't get it to work with my browser but I tried reading through the code to get some idea. Does it just compute volumetric flow through the engine at all RPMs, assuming a fixed volumetric efficiency?

I've been thinking more about this, since doing the initial calculation...

It seems to me that if you have a really good intercooling setup, it pretty much shouldn't matter what kind of compressor you're using. Right? I mean, all the engine knows or cares about is the pressure of the air it's ingesting, and the temperature of that air. Since volumetric efficiency is fixed for any given manifold pressure, and air density is determined by the pressure and temperature... So, 12 psi is 12 psi, whether a VF10 or a TD05 is producing it, right? That is, assuming the temperature is the same?

Of course, we can never achieve perfect intercooling... But maybe we should try harder. A ridiculous pressure drop across the intercooler would be okay if it resulted in very good cooling, too, since all we care about is the temperature and pressure at the outlet.

I guess it's risky to rely on the intercooler, though, since heat soak can happen even with the best designs, and then you're open to some dangerous prospects.

Okay, I'm just rambling again...

[WRXdan]

Without reading thru many many threads.....Has anyone been able to change the code in the ECU? Or are people in the reading stages? Too bad there are not companies supporting this idea. Kinda like www.tunercar.com, they make a good program to reburn EEPROMs for GMs.

edit: tunercat might be gone now??

double edit:

Found them:

http://www.tunercat.com/tnr_desc/do_tc.html

[tris91ricer]

Thank you though, as I no longer go cross-eyed when I look at the map /

No longer cross-eyed?? i went cross-eyed looking at Vikash's math! geez... now i know why i don't want a turbo... too complex! ack!

-=tris

[tris91ricer]

let me try that again...

free5ty1e wrote: [quote] Thank you though, as I no longer go cross-eyed when I look at the map[\quote]

No longer cross-eyed?? i went cross-eyed looking at Vikash's math! geez... now i know why i don't want a turbo... too complex! ack!

-=tris

[tris91ricer]

damn. oh well.

[THAWA]

i believe you mean:

Thank you though, as I no longer go cross-eyed when I look at the map

No longer cross-eyed?? i went cross-eyed looking at Vikash's math! geez... now i know why i don't want a turbo... too complex! ack!

-=tris

also you can edit your posts

[NICO]

i got my ecu retuned 3 weeks ago and yoish smoothed out the hole car it drives so perfect, he siad my car was leaning out in the middle rpms and massive rich at idle.

[free5ty1e]

morgie: your calculator is written in some strange language... it frightens me

seriously though... Nico... please 'splain where and how you got your ECU tuned. Inquiring minds want to know.

[bignose]

morgie: your calculator is written in some strange language... it frightens me

seriously though... Nico... please 'splain where and how you got your ECU tuned. Inquiring minds want to know.

afaik, yoshio is a japanese guy in toronto that does some discustinglin good ECU mods on subaru's. whta i've heard is that he doesn't speak any enlglish, and if you start telling him what you want, he doesn't pay attention to you. he does it "his way" and no one ever complains.

but this is all here-say.. i could be way off.

[free5ty1e]

Damn... toronto is a bit far for me. Cmon wheres the orlando Legacy ECU tuners?

[NICO]

no that is right about yoisho, he is good i just told him FASTER FASTER hahah, then he drove my car came back smiling he new it was fast.

i have to go back when i get the td04 turbo and my headers back on so he can tune it agian for max power.

if you have the same stuff i got he can make a chip an you can stick it in your car its easy for legacy turbos he siad. i helped him take my ecu apart.

[free5ty1e]

Wish I had the same mods you do Nico... heh... front-mount intercooler and td04 would be sweet. Oh well, i'll figure out something. Still waiting for the t-bird turbo to get here and have to find a rebuild kit at a decent price... and an exhaust shop that would be willing to make me a 3" turboback including downpipe. If only I lived in the Great North