Help Understanding Part Of A Turbo Compressor Map

[Moodles2]

Does anyone know the value a mild RB26's corrected air flow would be to apply to the pictured compressor map (and the formula you used to work it out):

(Garrett -9)

[Moodles2]

I think I still have alot more to learn, all I can gather is the -9's lose heat efficiency around the 24-25psi mark.. was hoping to plot a line that a stock rb26 would represent roughly, starting from the lowest pressure ratio all the way as far as where the heat efficiency peaks at 2.5-2.65ish (pressure ratio)

[simpletool]

Well it looks like 20psi would be the most you could use on those turbos if you want to stay in that map.

To plot the pressure ratio:

Take your desired boost pressure, add 15 (for sea level pressure) and then divide the total by 14 (which is an approximation of the intake pressure before the turbo compressor)

So (20 + 15) /14 = 2.5

to get 2.5 pressure ratio you would want

This gives the pressure ratio you can plot on the y-axis.

I chose 14psi as that is the absolute pressure before the compressor (after the air-filter) and it resulted from guessing a 1psi drop between the air-filter and turbo intake. This is what is in "maximum boost" by memory. Corky Bell may even use 2psi pressure drop which means the pressure ratio is 2.7 - which would mean you could only use about 18psi.

So then to get the x-axis value (for every thousand revs or so) you guess the delivered power, and approximate air flow values from there. I really have no good idea about what the ratio of air value to power on a tuned GTR would be, but my guess is about .11lb/min of air for every hp at the engine. This depends on tuning, it could be .12lb/min.

So say a mild GTR at 20psi at peak revs 8000rpm makes( umm I dunno), ~ 550hp at the engine. And at 7000rpm it makes 510hp (assumed less efficient after 7500rpm), and at 6000rpm it makes 440hp, and at 5000rpm it makes 360hp, 4000 it will be off boost. Then divide all these numbers by 2 for each turbo (actually should be a bit less than 2 as you need to assume that one of the turbos is making more boost than the other and you are really only calculating for the hardest working turbo, but we'll use 2)

So you convert those hp at the engine numbers to lb/min to plot on the x-axis at the y-axis level of 2.5, and label with revs for info.

So you have a line from 5000rpm which is 180hp x .11 = 20lb/min

To at 8000rpm, 275 x .11 = 30lb/min.

To plot before it comes on boost you need to approximate the boost level of the turbo as it comes on boost and the rev that it comes on full boost is the turning point on the map. so say 4000 it might be making 6psi boost and 180hp, so plot at (180x.11)/2 = 9.9 lb/min at pressure ratio 21/14 = 1.5

Now both these values are at the VERY edges of the chart at 2.5 ratio. So really these turbos are maxed at 20psi and can make about 550hp at the engine with a pair. Yes you can make a bit more but they'll really be cooking. Please remember I have made assumptions about hp to lb/min and all that so check those first.

Edited September 29, 2009 by simpletool

[discopotato03]

I'm not at home ATM so can't quote formulas to run the corrected pounds of air calcs .

I think the important bit is that there is a standard temperature that air needs to be at to be able to measure its weight or mass . The way I look at it is that temperature very much affects airs density and density is directly related to it's mass .

If it makes any difference , and I'm no fluid dynamics expert , I also tried years ago to work out air flow through engines in order to be able to spec the perfect turbo to a given engine . The trouble is that it's difficult to be exacting with any given engines volumetric efficiency let alone the temperature changes between the air filter and the inlet valves . 1001 variables and so many numbers entered into the formulas with fingers crossed - using "The Force" if you like . It doesn't promise an accurate answer and turbos are neither cheap or easy to R and R .

In the end I chose to look at how many pounds of air it takes to make X amount of horsepower and work that number into potential compressor maps .

Lets just say you want say 600 horsepower and want to explore twins for an RB26 , you have to be realistic about power goals here too .

The figure I see bandied about is that it takes 10 lbs of air to make ~ 110 Hp and I tame it down to 10 to the 100 so as not to be at the ragged bleeding edge of compressor maps . So I'd think of it as 60 lbs of air for 600 Hp and halve that because I'm looking at two compressor sections .

A significant thing to remember (IMO) is that the part of the map with the area of highest adiabatic efficiency (thermal efficiency) really needs to correspond with the air flow rate at the torque peak ideally . Generally an engines torque peak is at its point of highest volumetric efficiency so having the coolest on boost charge temperature should be of most benefit in this area .

I used to see people going cross eyed crunching numbers and planning to have this highest compressor efficiency occur at peak power air flow rates . I think all they achieved was using compressor sections way to big for the application .

One of Corky's quotes I like is that turbos make torque and that makes fun .

There is one thing that you cannot forget about when trying to size turbos - the turbine side . It doesn't matter how well or badly the compressor section is suited to the engine because it has to vent it's exhaust gasses through the turbine ends . If you get the hot side wrong it either won't drive the compressors properly or will choke the engine under load .

My 2c burnt , cheers A .

[Moodles2]

Awesome input, got a very good idea how it works now.

[StealthX]

I'm not at home ATM so can't quote formulas to run the corrected pounds of air calcs .

I think the important bit is that there is a standard temperature that air needs to be at to be able to measure its weight or mass . The way I look at it is that temperature very much affects airs density and density is directly related to it's mass .

If it makes any difference , and I'm no fluid dynamics expert , I also tried years ago to work out air flow through engines in order to be able to spec the perfect turbo to a given engine . The trouble is that it's difficult to be exacting with any given engines volumetric efficiency let alone the temperature changes between the air filter and the inlet valves . 1001 variables and so many numbers entered into the formulas with fingers crossed - using "The Force" if you like . It doesn't promise an accurate answer and turbos are neither cheap or easy to R and R .

In the end I chose to look at how many pounds of air it takes to make X amount of horsepower and work that number into potential compressor maps .

Lets just say you want say 600 horsepower and want to explore twins for an RB26 , you have to be realistic about power goals here too .

The figure I see bandied about is that it takes 10 lbs of air to make ~ 110 Hp and I tame it down to 10 to the 100 so as not to be at the ragged bleeding edge of compressor maps . So I'd think of it as 60 lbs of air for 600 Hp and halve that because I'm looking at two compressor sections .

A significant thing to remember (IMO) is that the part of the map with the area of highest adiabatic efficiency (thermal efficiency) really needs to correspond with the air flow rate at the torque peak ideally . Generally an engines torque peak is at its point of highest volumetric efficiency so having the coolest on boost charge temperature should be of most benefit in this area .

I used to see people going cross eyed crunching numbers and planning to have this highest compressor efficiency occur at peak power air flow rates . I think all they achieved was using compressor sections way to big for the application .

One of Corky's quotes I like is that turbos make torque and that makes fun .

There is one thing that you cannot forget about when trying to size turbos - the turbine side . It doesn't matter how well or badly the compressor section is suited to the engine because it has to vent it's exhaust gasses through the turbine ends . If you get the hot side wrong it either won't drive the compressors properly or will choke the engine under load .

f**k, this guy knows what he's talking abut! lol

[discopotato03]

I don't know about that , it just seems logical to make the turbo/s capable of pumping just about enough air to reach the power ask .

There is never a guarantee that your engine can actually make the power you want but if the turbo/s are sized correctly it means you have to look at other things .

Same old same old , make the engine breathe well and use conservative turbo size and boost levels . Thats my theory anyway .

A .

[StealthX]

Pretty dam accurate theory if you ask me.

Too many people try to over complicate things when it comes to turbo selection, IMO unless you really know whats going on in terms of advanced physics (in particular fluid dynamics), or your are looking to make bullshit power from a smallish motor, don't bother with the advanced stuff concerning sizes. Just use something that's bin tried and tested (GT2871r for SR20, T51r for 2JZ etc), and go from there.

[discopotato03]

I think in simple terms make the compresor just a tad over the airflow requirements and match the turbine and its housing to the compressors air output .

I think better to have a balanced match of compressor and turbine than compromising turbine size to wind the thing up . If the balance is right and its still laggy I think it's trying to tell you that its overall size is too big .

In a way I see sizing turbos a bit like sizing cams in an NA motor , go too big and lose the bottom end torque . I suppose falling off the boost range is a bit like falling below a given cams power range .

I think in the end the trick is making forced induction boost torque and I wish people would forget all about Kw and Hp numbers . Those dyno graphs with the Hp/Kw lines like the side of Mt Everest mean jack to me , the torque plateau is everything .

To each their own , cheers A .

[tridentt150v]

Is there any chance of taking Simpletools answer and making it a tutorial or some such - so we have it for future ref???

[simpletool]

Explained another way:

Compressor maps are actually fairly easy to read. A google search will take you to the Garret website under "turbo tech" or to lovehorsepower dot com. Both pretty good.

The basics are you plot a y-axis and x-axis for a few rev points, at least 3 points: one coming into boost, one just on boost and one at maximum revs. On boost the y-axis we can assume will be constant - a flat line.

X-axis is the flow of air - which is almost directly related to power. On a well tuned engine you can APPROXIMATE will be 10lb/min of air to 100hp (AT THE ENGINE). My initial assumption of using 11lb/min per 100hp is a 10% increase that means you have some headroom on the compressor before running into choke and also that a turbo engine is usually tuned pretty fat - so it needs more air to make the same power.

Y-AXIS

Plotting pressure ratio on the y-axis.

Pressure ratio = (absolute pressure / pressure at mouth of compressor)

Absolute pressure = boost + atmospheric (or boost + 14.7psi)

You can use bar or psi or mmHg if you want - just be consistent since it is a ratio.

X-AXIS:

at each rev point you work out the engine power, and convert to air flow in lb/min. Remember and easy approximation is 10lb/min air for 100hp at engine.

And that's it. Just don't run into choke, which is on the far right. The line on the left is the surge line - the speed where the compressor starts to grab air properly. The line on the right is when the compressor is creating too much heat and going to fast to be efficient.

Then size this to a decent sized turbine (outlet a touch smaller than the compressor inlet) , and don't skimp on the A/R ratio so it doesn't either backup exhaust pressure (limiting flow through the engine head) or come on boost like a light switch and be too sensitive to throttle.

Disclaimer: All my thoughts are theory based with a touch of pracise - but at least I did thermodynamics in school

Edited October 1, 2009 by simpletool