Turbochargers - turbine housing AR selection - how to work out?

[boxh34d]

Ok, I appolgise for the following essay, but it I found it hard to explain what I mean, so hopefully it all makes sense.


So I know (or at least have convinced myself I know) about matching a turbo compressor to an engine. However, choosing a turbine side to the turbo I am still not completely sure about.
I know that most manufacturers don't list their turbine housing efficiency details, so third hand knowledge or a rough guess is going to be required.

Garrett however do publish their data.
If we take a look at an example of say - a GT2860RS
http://www.turbobygarrett.com/turbob...S_739548_1.htm
Its compressor is capable of moving between about 13 and 33lb/min @ a PR of 2.25 without going into surge or choke. Having a look at the turbine effeciency map, there is a 0.64, and a 0.86 available for the turbine housing.

From what I've read, if a turbo has a good compressor to turbine ratio, (1.1 to 1.25:1) and the turbo is a reasonable match for a given engine, there is a bit of a rule of thumb for working out spool time. The lb.min flow when the turbine is at (or close to) its peak efficiency at the given pressure ratio, is the same as when the compressor should be capable of spooling hard enough to make the same figures.

Eg: For a fairly stock 2 litre engine that makes 300hp @6000rpm & 18psi: We can assume that at 3000 rpm, the engine will probably use around 15lb/min @ 2.25PR, and at 6000 rpm, it will use about 30lb/min.
The 0.64 housing for the GT28RS (link above) has got very close to peak efficiency at 2.25 PR at around 17lb/min turbine flow. Provided that the engine is capable of producing enough gas for this to occur, theoretically the turbine has enough power to drive the comp wheel to a similar figure - give or take a bit, so full boost would probably be reached just over 3000 rpm.
If this is compared with the 0.86 housing, the same sort of efficiency at 2.25 PR won't occur until around 21lb/min. The compressor probably wont be able to make full boost until after 4000 rpm.

I've seen this kind of guestimate figure work out somewhat close on a couple of different cars now, so I think it is within the ballpark.
What I don't understand, is what happens after this peak efficiency of the turbine is reached. Whilst we all know that a smaller AR housing will be more responsive, and have a lower overall power potential than a larger housing, is there any way to work out when this might occur?
In the example above, how do we know if the 0.64 turbine housing will flow enough for the desired 30lb.min of compressor airflow at 6000rpm without starting to choke, or if the 0.86 will be required?
Is there a rough guide for this one that anyone has heard? I once heard that if you go over double the peak turbine flow it will start to choke. So in the above example, the 0.64 would be ok for a bit under 34lb/min, and the 0.86 good for 42lb/min.

Does this sound even remotely close to what people have experienced, or is this another case of the interwebs providing some poor quality information? Has anyone tried a couple of different AR's with all else being equal and compared power / boost curves etc?

I guess I just really don't like the idea that choosing a housing always seems a case of either:
a) Listen to the bloke accross the road / on the interweb / behind the counter, because whilst they haven't built one, their second cousin's mate's brother had one and it made eleventy billion HP and had full boost by 2000rpm, or
b)Just buy one, hope that it's the right size for what you want and be prepared to try and trade it if it's not?

Look forward to your thoughts (and hopefully your insight) people.
Thanks.

[Supra967]

Can't really offer anything useful. This guy has put up some good info on his car which is top level and looks to have choked out his housings.
http://www.v-eight.com/tech_forum/vi...=1759&start=60

Don't know if the dyno can be used as a spool comparison cause there's 40-50km/h road speed difference in spool!

One thing that does work in the favour of the 'he said she said' argument is that there aren't a lot of combinations available and a lot of people make their findings public so if you sift through you could get a fair comparison. Maybe even start a thread for people to list mods and outcome?

[td42gq]

Just my 2c but I think the maths, in theory may be great but unless you know how to interpret all the maths accurately I think it might be easier and more effective to do as Supra967 suggested and take a generalisation from the population of people who have already done it.

My tuner has a computer program to calculate what your talking about such as boost times, pressures, lag etc but I don't have a copy so can't enlighten you much on it. So in theory the maths will work but unless interpretted exactly right you may get a better result from the population then from figures on a piece of paper. Unless of course you know a tuner with such a program who will let you use it or give you the information from it.

What motor are you putting what turbo on?

Aaron.

[Rhyno]

Hi mate,

I'm no guru, but have been reading up a bit on this, as I am in the process of Turbocharging an old 18RG-U. There are calulators availabile (or programs as mentioned above) that takes a few figures and spits out the required AR housing. Was quite a while ago that I stumbled accross it, but from memory you will need:

Volumetric Effecincy of your engine:
Operating Air temp (ambient between x and y):
And the compressor map:

This will then spit out a map showing what the turbo will do given the air temps used and the volumetric efficency of the engine. I just googled Turbo Sizing Calulator on google and this was the first one that popped up... http://www.turbofast.com.au/TFmatch.html... not exactly what I was talking about, but do some digging and you will find the right one.

Perhaps some one can reccomend one?

Hope this helps,

Kind regards,

Ryan

[Supra967]

i think anything that calculates choking based on guessed flow conditions and area is going to be no more reliable than your cousin's mate's brother.

Given for example the GT28 above, having only two turbine housing options, that gives you a great chance of finding people who have used either. Do you go on with high level flow calcuations or hit google?

[jonathan yee]

and how does one accuratley work out the VE of an engine given the varing differnces in port design (shape/size/angle) valve sizes, cam sizes and timing of said cams, inlet manifold and length/dia of runners thru to the TB size?

dont all these things change the VE of any given engine?

[Rhyno]

Like I said I'm no Guru on this. but a little searching found : http://www.epi-eng.com/piston_engine...efficiency.htm

Kind regards,

Ryan

[boxh34d]

I have already worked out the VE accurately enough for sizing of the compressor side. I developed a spreadsheet calculator based on engine displacement, rpm, HP, BSFC, AFR, intake temp, and pressure. Whilst I'm positive that it is not the 'be all and end all' of calculators, I've tested it on a couple of different cars, (both NA and turbo) and it seems to be accurate enough for purpose - ie choosing a compressor. Does it really matter if my VE is out by a couple of % or my BSFC is a poofteenth off - not really.

I have trouble with the turbine section though.
Whilst I have no doubt that finding a formula for the exact flow available through a given turbine housing is never going to happen for a lay-man like myself due to being overly complex, that was never the intent. I did however think that there might be a rough guideline to give an indication on which turbine housing might be suitable for the amount of power you want to run through it without choking.

Yes, in some cases you can use other people experience as a guide. If I had an SR20 and wanted to put a GT28 on it, I'm sure I'd have no trouble at all finding someone with a similar setup, and similar goals. What happens when you want to: a) use a different engine, b) use a different turbo to most, or c) have different goals to most?
A lot of data around there is useless too. US dyno results aren't worth the paper they are printed on half the time, as somehow they have cars making 400RWHP, with a turbo that is only capable of supplying 35lb/min (350 engine HP) worth of air.

(start rant here)
The other problem with trying to find this info, which is incredibly common, is that a lot of people don't look too far into anything they do, or are just full of shit. How many times do you see people putting turbos on that are far too big because they read that the turbo is good for 800 hp, and think that it'll make their 300hp engine suddenly make 800? Strap it on, don't bother dynoing because they are still running on a stock ECU and some dodgy sensor benders - but "she'll be right mate, that's what the extra injector is for". They now think their car is the fastest thing on the planet "cos it really pushes them back in the seat when it comes on", but in reality they just think that its fast cos it has no power down low but doubles its power @6000rpm and flies to its redline at 7000rpm, but can't understand why they are getting beaten by stock 6 cylinder falcons.
(end rant)

Surely there are a few people around here that have tried a couple of different housings due to their first choice being too small/large for their goals, and have some dyno results or can remember the difference in power / boost / rpm from before and after. I'm sure that after having a look at a few examples, a rough trend would present itself so there is less requirement for relying on someone else having the exact same engine, turbo, power, boost, mods, etc as you beforehand.

I need a beer after that - sorry again for the essay. Keep any info, discussion or experience coming. Even though I probably am coming accross as ungratefull and a know it all I can assure you its not the intent (or the truth). I'm gratefull for the input, and looking forward to some peoples findings.

[NME308]

Hi there,
As for the question of what happens when you reach the peak efficiency point of the turbine the simple answer is that the back pressure will build fast and choke the engine.

The good thing is there is an answer called the wastegate! If you are not aiming to absolutely max out the compressor - which in itself is not a particularly great idea from a heat and efficiency viewpoint - then a small A/R housing which gives better response need not be the cause of choking the engine. Obviously once you reach desired boost pressure the wastegate is opened allowing exhaust gas to escape without going through the turbine. As long as the wastegate is large enough and positioned advantageously the turbine will be given the amount of gas it requires to keep the compressor powered and the rest is bypassed. If the basic turbine wheel to compressor wheel geometry is correct then the turbine housing size can be kept nice and small to give better response.

I had to port the internal wastegate on my GT2560 with .64 turbine to allow the engine to consistently run 20psi. Previously it would hit 20psi for an instant and then drop back to 16psi or so and the engine note would go a bit 'dull'. On our HQ my brother and I dropped the exhaust turbine size from 1.06's to .63 housings even though most advised against it. Due to running twin 60mm wastegates with good postion on the manifolds we picked up full boost 1000rpm earlier on the transbrake with the housing downsize and havn't choked the engine. We are currently running the turbo's at the edge of the garrett hp rating and no signs they are out of puff yet. I say running the turbo's at the edge of the garrett hp rating but not at the upper pressure ratio.

There is always many ways to skin the proverbial cat. In this case I happen to like this one as it gives me great engine repsonse!

Cheers,
Jason

[Supra967]

Boxhead - for a guess/guide, you need to do an energy balance between the compressor and the turbine.

Steps would be:
1 - work out engine power (mass flow) and required boost level
2 - work out compressor power needed to maintain flow/boost - p=m delta h (take h from air property tables, don't forget compressor efficiency)
3 - guess turbine efficiency for given flow (if not able to take directly off the map)
5 - work out turbine entry conditions based on compressor power input required, mass flow, guessed egt before turbo and guessed turbine outlet conditions - work back from p = m delta h (this is the big one with the most unknowns.
6 - work out maximum flow through orrifice based on AR and turbine inlet conditions - formula is a mofo
7 - compare 5 and 6.

[boxh34d]

NME308: Very interesting with the 308. Any chance of some basic specs on displacement, horsepower, boost, turbo sizes etc? Very interested to have a look at the numbers on this one. The idea definately sounds like a good way to try and have your cake and eat it too.
+ rep

Supra967: Funnily enough, I thought that the 'real' way to do it would be something along those lines - bloody near impossible. Thanks for the insight though.
Apparently I must spread love around before coming back to you.

[NME308]

Haha, sorry boxh34d I have a 350 chev on board these days... I couldn't keep the heads on a stroked holden V8 naturally aspirated so didn't dare force feed one!

Specs are 5.7 liters displacement, forged bottom end with 7.2:1 compression, iron eagle 230cc 23 degree heads, Garrett GT3582R turbo's, and a 850cfm blow through carb. Horsepower numbers vary with the medium they are measured by. So far - 20psi boost gives 975 moroso drag strip horsepower which gave 786rwhp on dyno dynamics device. With 4psi extra boost it went 940rwhp with 1300ft/lbs (1760nm) torque and at the strip next outing it spat the powerglide transmission. Currently saving for a mega dollar trans made entirely of unobtanium...

Cheers,
Jason

[boxh34d]

Sorry Jason, but what RPM was max power?

[NME308]

Max power was at 6500rpm. The max power rpm figure goes up higher as we feed more boost in. When we were running in on 12psi it was making 600rwhp at 5000rpm. The power run from earlier in the morning to the 940rwhp was 786rwhp on 16psi boost all in at 6000rpm.

Cheers,
Jason

[boxh34d]

Jason,
Sounds like a fair animal thats for sure.
Looks like the turbo's are a perfect match for your combination though. Very interesting that you can almost achieve the max out of the compressor (I assumed 940RWHP = 1175engine HP, ~ 120lb/min, so 60lb/min each) without having troubles with too much turbine pressure. I bet the 60mm gates are well and truly screaming by then though!
Thanks for the info - this certainly gives me something to go by if / when it comes time to upgrade my current package.
Sorry but I can't rep you any more.

Just as a side note though, you must be getting some pretty impressive VE with your head / cam choice. Whilst my calcs have previously stacked up well with lightly modded engines, I couldn't get your package to add up in terms of boost pressure, flow (assumed HP) and the revs stated until the VE was up around 110%. Whilst I have no doubts that my calcs aren't 100% perfect, I have so far found it good enough for guestimates, but this one had me stumped. I've never had anything to do with trying to work out numbers for something quite in this league. Interesting stuff.