Turbine Maps

[Dale FZ1]

Ok, in the past few weeks/months there has been quite some discussion about the products available from Garrett, and I have to say it sparked my interest in learning more about the science of matching them, and then the practicalities of fitting them.

The interest is in knowing how/why different configurations are offered by the manufacturers, and further variations by suppliers such as GCG and HP in a box.

The mathmatical guide that Garrett gives is enlightening, and tells a whole lot about matching compressor flow to the power requirements of the engine. So far, so good. Where it unravels a bit for me is in getting an understanding of matching compressor to the turbine (ignoring the problems of the availability of T3 flanged units to bolt-up to Skylines).

Running through the comments by many, it is clear that some units work quite well, while others are considered "laggy" and lacking a good mid-range (eg GT2540 on RB25) or having a boost transition point higher in the rev range than desirable. The concept of turbine acceleration and shaft torque sufficient to drive a given compressor was introduced, and this is what got me thinking... and confused at the same time.

Turbine maps appear much simpler to read, and those that graph multiple A/R capabilities clearly show the ability of larger A/R to flow greater air mass at the same pressure ratio - thereby achieving higher levels of efficiency. Simple.

The hard part is in matching and interpreting the capabilities of the turbine of driving the compressor, and knowing what sort of characteristics this will give. Obviously the experiences of others who have gone down that path helps, but how is it done?

I'd considered that the pressure ratio calculated and superimposed onto the compressor map would be of some importance, but regardless of that, the mass flow capabilities of the compressor do not coincide with the turbine for any given unit. I am sure that if I delve back into my grade 12 physics and chemistry texts I can rediscover the science behind that one.

What I would most like to know is how do you look at the two maps and logically reason how the two will combine.

cheers

[MattSR]

"The mathmatical guide that Garrett gives is enlightening"

Could you give us a link to this guide? Im interested in what they have to say.

[discopotato03]

Compressor and turbine maps are very different because of the environments they live in and what is expected of them .

The purpose of the compressor is to push a greater weight or mass of air into the cylinders than atmospheric air pressure can achieve on its own . The problem is that air mass or density changes with temperature so in order to keep the density up compressors need to be highly efficient . The aim is for the engine to inhale more air so we can burn more fuel to develop more horse power .

Desirable qualities in a compressor are low mass - less weight for the turbine to drive , mechanical strength - ability to withstand the centrifugal forces at high rpm without bending and breaking the blades , ability to slice cleanly through the air its pumping - minimum friction and less drive load to be shouldered by the turbine .

Compressor maps are invaluable for showing the capacity and limitations of a set compressor and housing combination . They show the all important surge line and efficiency islands plus mass flow generally in pounds vs pressure ratio , the speed lines are important too . With some formulas its possible to crunch the numbers to assess a compressor/housing's suitability for any engine .

The turbines purpose is to recover as much energy as possible from the engines exhaust gas flow whilst posing the absolute minimum restriction to its flow . I'm told that a third of the heat energy developed in the combustion process goes out the tail pipe , some of this energy can be used to drive our compressor .

Turbine maps look different to compressor maps because we are mainly interested in mass flow VS pressure , note the flow line rises to a certain level and beyond this point does not increase with pressure . This is called turbine inlet pressure and is a measure of resistance to flow . Also note that flow rises with ARR (Area Radius Ratio) for a set turbine trim , so its not possible to rate a turbines flow by merely looking at its exducer diametre or the housings at this point . I'm not sure as to how they work out the efficiency number , it seems logical that the shaft power generated vs mass flow and temperature drop across the turbine is how it works . Obviously higher is better meaning more energy rcovery for less flow restriction . The general consensus is low turbine efficiency = more turbine lag .

I believe correct turbine/compressor matching is about have equal pressure in both manifolds under boost so that the engine gets to scavange properly on the valve overlap period . Poorly thought out combinations that lead to ratios of 2.5 to 1 (typical and can be worse)exhaust manifold pressure (EMP) vs inlet manifold pressure (IMP) lead to reversion and pumping losses .

Turbo production engineers (of the petrol head variety) agree that keeping the turbine diametre within 15 - 20% of the compressor makes for similar tip speeds or mach factor . If one reaches its mechanical speed limit appreciably before the other then potential is wasted for no good reason .

Have to go will round this off later .

Cheers A .

[Dale FZ1]

Here is the link to Garrett:

http://www.turbobygarrett.com/turbobygarre...bo_tech103.html

Thanks Discopotato - it is very technical and takes a little digesting, but very good. Keep it going.

Do you have an engineering background by chance?

cheers

[discopotato03]

LOL no but I was a Fitter/Machinist till I worked out it didn't pay too well .

Cheers A .

[Dale FZ1]

Well checking out the various models available, here is a run down of the relative compressors and turbines sizings, with the % differences:

HKS

GT2530 60.1mm; 53.8mm 111.7%

GT2535 69mm; 53.8mm 128.3%

GT2540 76.2mm; 53.8mm 141.6%

GT-RS 71.1mm; 53.8mm 132.2%

GT2835 71.1mm; 56.5mm 125.8%

GT3037 76.2mm; 60.0mm 127.0%

Garrett

GT2860RS 60.00mm; 53.8mm 111.5%

GT2871 71.0mm; 53.8mm 132.0%

GT3071 71.0mm; 56.5mm 125.7%

GT3076 76.2mm; 60.0mm 127.0%

Laid out bare, it is fairly clear where the model similarities between the two brands are, but all is not as it seems with variations in wheel trims, housing A/R, and flange sizes. HKS have evidently targeted certain niches with turbos that can bolt up to factory manifolds etc and give definite results for specific models, while Garrett offer greater opportunity to tailor various models.

For the sake of this discussion however, it is obvious where and why some models enjoy the reputations they do in respect of transient response. The “little” HKS 2530 and Garrett 2860RS (aka Discopotato) have relatively very similar wheel sizes (comp:turbine), and the write-ups about various cars running them suggest a linear power delivery and engine response like a bigger capacity naturally aspirated motor than a more “traditional” turbo job with massive mid range and top end.

The one aspect that does interest me from a technical viewpoint is the evident turbine:compressor drive efficiency that the larger GT3037/GT3076 “twins” have, giving the impression they would be damn effective and quite responsive power producers if you were targeting 450-500 hp. Whether the driver or driveline could cope with the results is a different issue.

[wolverine]

dale the GT3071R discopotato has been referring to in other posts is the one with the 60mm turbine not the 56mm turbine. so perhaps that is worth adding to the comparison.

[Dale FZ1]

dale the GT3071R discopotato has been referring to in other posts is the one with the 60mm turbine not the 56mm turbine.  so perhaps that is worth adding to the comparison.

<{POST_SNAPBACK}>

I picked up the possibility of the existence of such a beast - just could not locate anything in the Turbobygarrett.com catalogue, or the link to an older cat. (circa 2002) that Discopotato suggested.

When I digested his information about the two different housings (UHP, and NSIII), it appeared that the internally gated GT30 housing fitted to the GT3071R was a compromised beast, milled out to accept a cropped version of the 60mm turbine - done because everything was being made to go together when not originally designed to do so. Also note that there is no similar rotating assembly found in the HKS offerings.

I'll keep looking and if anybody can find the link to confirm the existence of a 71.1mm; 60.0mm combination please post it.

Also bear in mind that the original intent here is to investigate the relationship between turbine maps and compressor maps, so further comments please.

cheers

[wolverine]

disco has another thread here at forced induction devoted to the existence of said combo.

GT3071R link to thread

[discopotato03]

Dale FZ1 , if you divide the turbine major diametre into the compressor major diametre you get a more workable number ie GT28RS/GT2860RS is 53.8 divide 60.1 = .8952 x 100 = 89.52 . In other words the turbine major diametre is 89.5% of the compressor major diametre . For the real GT3071R (CHRA no 700177-23) substitute the numbers 60 and 71.1 to get 84.4% . The GT30R is 78.7% and the GT3040R is 73.2% - I think you get the picture .

With that mention of NS111 and GT30UHP its really a combination of turbine AND native exhaust housing that makes them work correctly .

In my opinion the good turbos generate a few pounds of boost at very low revs and don't do the massive torque rush when they spool up . I think its a case of low resistance to exhaust flow that does not go through the roof when the compressor gets up to speed . It also tends to be a bit more traction friendly and easier to drive close to the limit of adhesion .

Cheers A .

[Dale FZ1]

After a reasonable amount of research, I found a few pearls that are pertinent to this topic:

Compressor matching (to the required engine + power) is universally accepted as the first (and most important) activity. Plotting flow and pressure ratios on various compressor maps and interpreting them can provide some challenges, especially when trying to keep out of the surge area and get good performance under partial load/throttle conditions.

Compressor capabilities are established through a combination of wheel and housing specifications. Wheel diameter and trim must be considered – the larger trim has a bigger “grab” at the air, allowing greater air mass to be compressed. A larger diameter wheel generally pushes more air for a given turbocharger rpm. Housings with smaller A/R numbers are employed in higher boost applications due to their flow maps being extended upwards to the right (higher flows @ higher pressures).

Evaluating turbocharger matching to a road car engine is generally accepted as a measure of its ability to deliver the required power at WOT, as well as relatively elastic (read: quick) boost and engine response at low to middling engine rpm and throttle.

This is where understanding turbine maps and establishing the capabilities of the compressor and turbine combination is essential if you want something that is a capable all rounder - rather than excellent in one area but falling rather flat in another. The concept of “average power” and reviewing the total area under a curve rather than peak values has previously been discussed on this board.

I struggled for a while to advance my understanding of transient turbocharger response (ie. what happens in on/off throttle driving conditions) until I took notice of the compressor speed lines on a compressor map. Plotting your points required also tells what speed range the whole rotating assembly has to operate in so that the engine delivers the desired power. Taking note of that speed range, the turbine selection/evaluation process becomes less difficult when the way it operates is appreciated. The turbine wheel/housing assembly presents a restriction to free exhaust flow when under load, thereby accelerating the wheel. Energy is transferred to the turbine wheel (causing acceleration) as the pressure drops when gases dump through the vanes and into the outlet.

A properly detailed turbine map should be produced as a multiple series of discrete flow curves according to the turbine speed and pressure ratio. Go to Borg Warner’s site www.turbodriven.com/en/turbofacts/design_turbine.asp to see what the words here are attempting to describe. What is generally done in fact is that a single curve is drawn as a representation of the mean mass flow over a specified range of turbine speeds. Think of it as a “smoothing” of the maximum flows recorded at various rpm to simplify the map.

Turbine maps that are released by Garrett differ in what I would say are two important ways – firstly the maps have the smoothing characteristic as described, and therefore omit information about the speed ranges are required to produce a certain level of efficiency. Secondly, only the maximum efficiency figure is quoted rather than producing a second y-axis scale for efficiency. Go to www.turbobygarrett.com/turbobygarrett/catelog/Turbochargers/GT30/GT3071R_700382_3.htm for an example to compare the detail. The issue is that the turbine will only be operating in its maximum efficiency range some of the time.

The match of turbo-engine entails matching the compressor to the engine airflow requirements, and the turbine must match the compressor’s drive requirements to achieve targeted boost response characteristics. This means you need to know the compressor’s expected operating speed range – and by implication the turbine’s required speed range. With turbine rpm controlled by a wastegate, it may be too slow to be driving efficiently, leading to slower than desired shaft acceleration, and thereby produce laggy boost response. Maintaining and/or accelerating to the required compressor speed is critically important, with another site at www.automotivearticles.com/Turbo_Selection.shtml giving some good information.

Apologies for the long post.

[discopotato03]

No probs and good points raised . With the bit about the wastegate and turbine acceleration the gate should not be open at all as the turbo "spools up" . The wastegate is opened normally by compressor outlet pressure to give constant boost pressure as the speed and air flow rate rises . The whole thing about well matched turbos is having low exhaust restriction so we can develop more power from less or sane boost levels .

You are quite right saying that turbine flow maps are all about maximums in flow and pressure but they are the critical points with turbines . When your turbocharger is windmilling , light throttle cruise no positive inlet manifold pressure , turbine efficiency is not terribly important . Exhaust flow rate is important and minimum or no restriction is very important . The manufacturers are only interested in what happens when it starts to boost . Most current turbochargers owe their development to the world of Diesels with good reason . Next time your on the highway near an interstate semi listen to the way they load up and the turbo sound . They are designed to work over a narrow rev range (engine) and the turbo can stay on full boost all day and all night . They are designed to be efficient at full boost and survive it for long periods of time . The locomotive diesels we operate literally run flat out at 150kpa boost pressure for hours churning out 3000 Kw .

Transient response is not just about spooling a turbo up from low but rising engine revs (and exhaust gas flow) . You could be rolling down a hill doing 4000 rpm with the throttle shut and turbo only idling (very little exhaust gas flow) and boot it at the foot of the next hill , if the turbo had heavy internals the transient response can be not real flash because of the innertia to be overcome . Very rapid transitions can make a modern turbo engine package feel like a larger NA engine torque wise .

Cheers Adrian .

[Dale FZ1]

Thanks for the comments, since I wanted to keep the thread going and get the whole picture.

Turbine maps and size matching has significantly more science to it than it first appears. Getting the turbine speed range “right” ie. So that it is running the compressor in its required speed range to deliver sufficient airflow into the engine to produce the target power output seems easy. Turbine size selection is simply such that it will generate the required shaft torque to push the required mass of air – correct? Not entirely as it turns out.

Checking my own understanding, I had to review what characteristics seem to point towards a good engine-turbocharger match. Quick acceleration of the rotating assembly is a prerequisite to good engine response.

Discopotato had earlier bandied the suggestion that sizing the turbine and compressor wheels within a “rule of thumb” 15% would yield a responsive turbo. This led to some posts about relative wheel sizes in various Garrett and HKS models – ignoring for the moment the actual specs of the housings they run in. The figures showed that the GT2530 / GT2560 twins had the least difference in relative size, and they enjoy a reputation for exceptionally quick boost response combined with power potential of about 320hp. Lots of happy owners using them in RB20, RB25 and SR20 applications too.

The figures also showed the larger, higher flowing models commonly fitted to RB25, 26, and 30 had a fair deviation from the 15% “rule”.

One model, the GT32 (note no R suffix indicates a plain bearing construction) fell inside the 15% relativity of matching, and I found it listed by turbomaster.com as a recommended fitment for a heap of different European diesel engines. Nowhere could I really find anyone who had fitted it to a performance type engine despite a power rating up to 400hp. So how do you make sense of that?

Firstly the primary difference between compression and spark ignition engines is that conventional spark ignition engines are throttled. That means airflow into the engine is controlled by the throttle valve(s) and the engine will not ingest the full volume it is capable of until at WOT. The diesel engine cops the full amount with every inlet valve opening moment. Power is controlled by the volume of fuel introduced into the combustion chamber. As Discopotato indicated, the effective engine rpm range of a diesel is limited – giving the engineers a very good opportunity to size their turbos to work within a narrow range of efficiency. Quick wind-up is comparatively easy, because a change of (for example) 500 engine rpm can have them going from a loafing cruise and into maximum load/torque production. And they are further improving that with the introduction of VNT – variable nozzle turbine.

Slightly off track, but the point is our beloved RB engines work over a much wider rpm band. Not only do we find a quick wind-up desirable, but we also want the flow and efficiency to stay high when the revs are high.

Next I went into a bit of high school physics research to relate back to sizing and transient response and the 15% idea. A couple of terms came up that should be remembered – centripetal force, centrifugal force, and polar moment of inertia. They are worth doing a Google search just for your own knowledge, but the most significant finding was confirmation of an assumption regarding wheel diameter, wheel mass, and acceleration of rotating assemblies.

The mass of a rotating object is a determinant (along with velocity) of the amount of kinetic energy it stores when turning. Diameter does not impact stored kinetic energy. Also the forces acting on a rotating body act inwardly towards the rotating axis. ie. A spinning turbine will have perfectly balanced forces acting towards the centre of the shaft it is connected to. This explains why they can spin at up to 180000 rpm without failure.

The big one is that a larger diameter body of the same mass will offer greater resistance to a change in rotational velocity than one where the mass is more concentrated (ie closer) around the rotating axis. This means a smaller diameter turbine will generally offer a significant advantage in acceleration. The trade off is in a likely decrease in overall flow capability of the turbine unit vs a larger, less accelerative turbine. This is something not unexpected if you were attempting to compare (say) a 280hp unit with a 600hp unit.

The lesson emerging from this is that the turbine unit to suit a relatively small, quick revving RB engine must also be relatively small in order to quickly accelerate the compressor – overcoming the polar moment of inertia of the WHOLE ASSEMBLY. Being a “slave” to the compressor, the turbine must simultaneously be big enough to generate the torque to drive the compressor which shifts the required volume of inlet air, while being small enough to have a low polar moment.

Getting these ideas straight ends up giving long posts, so apologies again.

[discopotato03]

Yes Diesel engines have the advantage of no air throttle so get a better shot at better cylinder filling ie higher volumetric efficiency . When diesels run lean (being compression ignition with lower flashpoint fuel) they drop heat/cylinder pressure/power rather than becoming a self consuming oxy torch like petrol engines .

Back to turbines , this is from an early Hugh Macinnes book - Turbochargers .

Total time required to reach maximum speed is a functionof overall turbocharger efficiency and the polar moment of innertia of the rotating group . Moment of innertia is the resistance of a rotating body to a change in speed represented by the letter I .

I=K(squared)M

where K is the radius of gyration and M is the mass of the body . Radius of gyration is the distance from the rotating axis to a point where all the mass of the body could be located to have the same I as the body itself . In other words a turbine 1 foot in diametre might be represented by a ring with a diametre of 7 inches . In this case K = 3.5 inches .

For good rotor acceleration it is essential to have the lowest moment of innertia . Turbine wheels are designed with a minimum of material near the outside diametre to reduce K because the moment varies as the square of K .

Always remember that the turbine while being light needs to be free flowing enough to pass the gas . A forced induced engine gives the torque AND exhaust flow of larger NA engines so it critical not to choke it . A big NA engine with a tiny tail pie is a similar kind of dog .

Cheers A .