Choosing turbo is not any rocket engineering but with stupid choises you can ruin your engine's capabilities. I did wrong choises first, but now it seems that my currenct turbo is matching well to base engine.
Compressor is one of 2 most important factors of choosing turbo. Compressor inlet bore and compressor A/R can tell some things about flow of the item but best is if you can look compressor's efficiency map and determine if it matches your use. If you haven't any info of properties of certain item you can do some very rough assumptions about inlet bore. Usually 40-42mm bore is capable to flow 200hp, 49-50mm to 300hp. But how to do it right ? Find out your max revs R, engine displacement (cid) D, boost B, and volumetric efficiency V. Ok here we go:
Corrected engine max (cfm = cubic feet / minute) flow is: R * D * 0.5 * V * B/ 1728
Note ! This formula is valid only at sea level, if you're located higher you have to use altitude correction
When you're having corrected airflow, you can plot it on the compressor map and explore which would be the efficiency of this turbo at your max revs.
|I use my engine as example: max revs = 6000, cid = 120.6 (1979cc), volumetric efficiency is hard to find out but I think you can use 0.85 for normal cih engine. Boost 1.7bar absolute (0.7 relative to atmospheric). So we'll find out corrected airflow (cfm): 302.5 cfm. If map uses lbs/min as these maps you can convert by dividing your result by 14.285. So we'll get 21.2 lbs/min. Now if I had used Garrett T03 with trim 40 -compressor, efficiency of turbo would be enormously low at high rpms. And low efficiency means extra heat... maybe detonation... engine crash.. No good. In the picture there is red plot at the point of max rev and max boost (0.7bar) situation. As you can see plot is located far away from even 65% efficiency area. Reasonable minimum for efficiency at max revs is 60% specially if you haven't intercooler. Actually this compressor would not flow 21 lbs/min. If you try to get it with this comressor you have to try to go over max revs (turbo's rev lines are marked on map) and that can lead to very disastarous effects.|
|What about if I choose compressor from another end of scale: Enormous T70 for race engines up to 800hp... it has to flow enough. I plot again the max rev situation on turbo map. Damn ! Then I notice that plot is located left from compressors surge line ! This compressor won't be able to work so low flow levels and low boost.|
|Finally I find some suitable compressor = T03 50-trim... maybe 60 would be even better but I let's assume that this was cheap ;-). Efficiency at max revs is near 70% and when I plot through engine's flow range I notice that compressor suits well to my engine. How do I plot other situations than max revs? explore the rpm versus boost curve and plot counted points to map. Boost curve is hard to estimate, but when you have correctly chosen turbine section boost starts from 1/4 - 1/3 of rpm range and it is at it's maximum near 2/3 -3/4 of rpm range. I found out later from my engine management's logs that boost starts at 1750rpm and it reaches 0.7bar near 3200. So I can count flow at these 2 points and draw straight line between the points. If I wanted to speculate how this compressor would suit to higher boost rates I can plot new curves on map and I'll notice that this should work up to 1.2bar.|
It's hard for amateur like me to give any reasonable hint about choosing right turbine. Thumb rule is that smaller A/R turbine gives little more boost at low revs and lower turbo lag but higher back pressure and heat at upper rews. Bigger A/R turbine allows more flow at upper rev and produces so less extra thermal burden but disadvantage is bigger lag. Anyway I have heard rumours that Garrett turbines are not so sensitive. You can choose little bigger A/R turbine without so serious effects to lag. I have no experiences about this. If you want to hear one of my assumptions, I would choose 0.48 A/R for 2.0 litre and smaller cih's (200hp or under) and 0.63 A/R with 2.2E cih and more powerful 2.0E cih's (and when talking about popular Garrett T03 series chargers).
If you can find turbo with watercooled bearing section, get it. It gives little insurance against oil cooking (but is no reason to use unsuitable oil in turbo engine). many Garrett T03 are fitted with water cooled bearing section. Of course water cooling adds little complexity to hoses but it's no reason to choose oilcooled unit.
There are T03/04 turbos at markets which have internal wastegates and are rated to 450hp flowrates. So I would guess that in standard street use you don't gain anything with external gate... but complexity. Ok, if you're pro... go ahead.