325i Turbo, Turbocharger Sizing?

[DmcL]

wanting to eventually boost my own 325i at some point. i know enough to handle the tuning/fuelling/ignition side of things but what im not sure about is the size of turbo i should look for whenever i take the plunge.

alot of people across the pond swear by the holset HX40 nad i think also the HX35 but there doesnt seem to be an abundance of those units over here and im wondering what sort of garret or garret knock off (read: cheap ebay turbo) would be suitable. would anything below a T3 be too small and restrict the engine at higher RPM? what is the smallest turbo i could run and still potentially go up to or above 1 bar of boost and not have issues with it restricting the engine or running out of its comfort zone at 1 bar or more? also is there any ideal compressor/turbine wheel sizes for the M20B25 i should be looking for?

my goal would be a maximum of no more than 1.5 bar, probably more like 1 bar before considering ARP head studs and a decent gasket and maybe some other small internal tweaks.

and last but not least.. would any of the turbo diesel turbo's out there be of any use to me? curious as there is an abundance of turbo diesel turbo's for sale on the cheap and second hand.

[Gunni]

Pressure is relevant so you should first mention what power level you want.

You can get more power with 1bar then 1.5bar .

So first things first.

How much power?
then
How do you want the power curve to be? Lotso torque and high end or not to much torque but high end, a wide maximum power curve?

I have access to a wide variety of Borg Warner turbos at good prices. They are becoming the choice turbo for fans of Holsets.

[DmcL]

basically as much power as i can run safely without toasting the engine.. would like to run the engine as it is with original headgasket and studs then maybe fit an uprated gasket or O ring the head and fit ARP headstuds at a later date and up the boost a bit more.

wouldnt want a turbo that falls short in the higher RPM either.

ideally quick-ish spooling and also making plenty of power from mid RPM onwards but that might be a bit costly so really as close as i can get with a cheap-ish (possibly no name) turbo.

[Gunni]

So what is your final power goal?

Like EXACTLY how much power do you want?

[DmcL]

dunno.. how much can i have safely without headstuds and on the original gasket?

say 300bhp for a ballpark figure.. maybe more if possible and it wouldnt cost alot more.

EDIT: forgot to mention, it could well be a rear mount turbo setup.

[Gunni]

You said that you would like the option of upping the boost later . So I´m kinda lost in what you really want. If you bring out a concrete maximum number you can envision yourself to ever hope to run then we can pick a turbo from there. Turbo prices aren´t that different for a slightly larger turbo so it´s more about selecting the one that will suit your current and long term needs.

In any case a 300hp would be something with about 45mm compressor. GT2860R sounds like a nice match. I can´t see a reason to EVER make a rear mount turbo setup. There is plenty room in a E30 for a turbo especially with the M20. Why would you complicate things like that?

The headgasket can withstand almost anything. Well above 500whp. The ARP studs
are to keep the head from lifting. Other then that 350whp is pretty easy to get on a M20 with stock internals, cams, intake and such using 8.8:1 compression pistons and 98-99octane fuel.

[DmcL]

its hard to say a highest ever planned number because i dont really have a maximum in mind. i plan on building the car up slowly over time as i have already been doing that to an extent by the level/type of servicing, maintenance and small tweaks here and there. its a long haul project so i dont really know where it will end up.. if a stock M20 bar eventually fitting a decent headgasket, O ringed head and ARP studs can make 500whp then 500whp ill shoot for ultimately.

for an initial setup to get me going id say about 300bhp (would 275-300whp be a bit optimistic?)

so for starters something in the T28 size range, maybe even a T3? also whats the deal with diesel turbo's? id say something around a 45mm compressor might be found on a diesel but i dont know if diesel turbo's have a smaller exhaust sheel or anything different from a petrol turbo?

fuel wise id want it to be able to run on the average stuff without detonation or problems. since i have a background in tuning motronic i could easily reflash the tune on my emulator for drag racing events or whatever with a nice high octane tune and maybe knock the boost up a little above the road setup.

im not decided on a setup but im not ruling out rear mount.. if i go for a traditional setup ill probably use one of the ebay manifolds which i think has a T3 flange so best to look at T3's or something with the same size flange im guessing.

[Gunni]

45mm can be found on petrol and gasoline.

If you plan on using the ebay manifold for starters then get a T3 type turbo.
Might want to see what Evo´s and Scoobies use as they might use T3 flanged turbos.

300whp and the usual torque of about 400lbs is stretching the limit of the stock headbolts by a decent margin, unless you keep the torque down while reaching 300whp. Which kinda negates fitting a smallish turbo to begin with instead of getting one that will do 500whp from the start and is laggier then a T28 and there for won´t create all that low end torque. But it would anyway have to be boost controlled to keep torque down.

It´s not going to be nice trying to tune the motronic too 300whp without having metering all the way.

[eta]

300whp is ~ 400 flywheel horse power. I think Gunni understood when you said 300 hp you meant 300 flywheel horse power.
T3 turbo's come with a variety of compressor and turbine wheels, so T3 by itself says not alot. I think T3 refers to the housing/flange.

A turbo from a diesel will be fine they are still turbo's. What you need to know is your hp goal and then study compressor maps. As far as I am aware it does not matter if the turbo is made for diesel or petrol engine.

There is zoner who recently bought a TD05-HL-16T from a suburu imprezza and that is good for around 300 flywheel hp. Then again Gunni has Borg Warner turbo's atm for £530 new.

Can anyone realy tune motronic so that it can fuel a M20B25 properly to the kind of power you are talking about. I have been advised that it is far from wise and standalone management would be necessary even more a more modest 300 flywheel hp from a B27.

[DmcL]

the people who advised u probably dont know just what can be done with motronic tho.. the limiting factor of motronic is that u are stuck with an AFM, or with a lot of tuning, a MAF. i have been trying to get a MAP sensor working with it but no dice yet.. also standard motronic limits u to coil/dizzy and batch fire injection. an M50 motronic ECU and engine loom along with some sensors and bits from the M50 would give a factory MAF, COP, and sequential injection tho.

in the time i have been messing around with motronic i have found a host of maps and figured out what they are for, etc. that were previously unknown by anyone i have come across so far so im pretty much in no-mans land already with the standard ECU. i can even wipe out the multiple unused part and full throttle fuel/ignition maps in the raw code and re-write in my own custom sized maps with RPM and load points set wherever i want them to be. i should be well up to the challenge of pushing beyond the current sort of HP levels of turbo M20's still running standard management.

fuelling is essentially just a matter of measuring AFR and trimming fuel maps to suit how much air is entering the engine. sure motronic doesnt read boost but it doesnt really need to if the fuel maps are manually trimmed to increase fuelling as boost comes in and is maintained to redline. on 327ire's turbo setup hes running 8psi and i sold him a set of 30lb (315cc) injectors, since i have previously had the injectors running perfectly on my n/a B25 it took literally about 1 minute to cut back fuel maps to the right level for these injectors (not forgetting cranking fuel so as to not drown the engine in fuel and have starting problems). with those 30lb injectors, a walbro 255lph pump and the standard 3.0 bar fuel regulator i was running the 30lb injectors quite a bit below the standard n/a full throttle fuel map values even once the engine had come on full boost of 8psi. standard fuellling in one cell for an example was about 128 in the raw data range of 0-255, in the same cell with the 30lb injectors and a full 8psi of boost the value in the cell was down around 80-90 and running a fairly rich AFR of about 10 so theres quite alot of headroom with 30lb injectors to add more boost without farting around with an FMU or pushing the injectors too hard.

yea i meant 300BHP but after a few dyno sessions on my own car i really dont like the calculated BHP rubbish most places spit out by default so any further dyno runs will be done in WHP. if the standard M20 internals can handle 300WHP then id maybe shoot for that rather than 300BHP.

head studs wise.. i may replace the standard ones with ARP studs and just change them out 1 by 1 and retain the original gasket for a while or until it fails. if/when the head came off id fit a decent new gasket and ideally get the head O ringed with steel wire to stop the gasket from potentially pushing out under boost. also maybe have the water jackets welded up if cracking the head in those areas could be an issue.

also by T3 i meant a turbo in the T3 family. i wont be too overly picky on the details of the turbo as my primary concern is cost so if i found something a little bigger/smaller than ideal but at a lower price thats what id be going for. i envision that if i ever do go for mental HP numbers id be putting in a different turbo and probably spending a bit of money on the new turbo also, maybe even ball bearing.

[Gunni]

Any tuning must be done by metering actual mass flow or pressure. Anything else I consider personally as bad tuning.

Why?

Well the AFM maxes out rather soon and if your not flooring it but getting something like 10psi it´s maxed out and your not getting the 15psi you normally use when you floor it then you´ll get highly incorrect fuel and ignition values.

A ecu can only control the engine within its metering capabilites.

[DmcL]

the AFM is mainly used for part throttle as i understand it.. it does have some input at WOT but not as much.

if i ended up maxing out the AFM for any reason i can tighten the spring tension on the flap and re-tune the ECU to take that into account which would give more headroom over the point it previously maxed out at. i could also run an M30 AFM and do the same if i ever maxed that out.. im actually running an M30 AFM at the moment with custom tuned AFM maps to suit it and the car runs as if it was meant to be there.

[Gunni]

When you tighten the spring the AFM resolution at low load becomes worse

[DmcL]

true but with a hefty enough bit of power at WOT sacrificing a little low load resolution would be worth it, just means id have to either smooth out the lower load area or rework the load points that the ECU reads and have less lower load columns to avoid any jumpy areas down low. if i did the latter i could add more resolution higher up which might see a bit of an improvement in the higher RPM's and re-calculating the load axis wouldnt take very long to do at all. plus if im running the M30 AFM then id already have a bit of headroom over standard anyway so wouldnt have to tighten the M30 AFM as much if even necessary.

[GeoffBob]

[Gunni]

The AFM on a M20 has little resolution stock. Therefor the ecu is designed to switch over to fixed fuel and timing at certain load conditions. They simply assumed that the engine would remain stock so no issues with unknown airflow. As they knew it.

Later version is supposedly able to influence the WOT maps based on O2 readings during cruise and part throttle running. Basically they assumed if the engine has picked up airflow at a certain point in the part throttle range then so must it have done during WOT. And by the same amount.

[GeoffBob]

[DmcL]

that is the only downside to tuning the standard motronic.. i dont know if i can do much about it after the switch over point unless i were to maybe re-calculate the load axis of the part throttle maps to run beyond the standard 75% ish of load before the ECU switches over to WOT maps. id have to try it to see but i would be able to raise the switch over point (how high im not sure of yet) and in conjunction with tightening the AFM spring and re-tuning AFM maps to suit i may just be able to get the switch over point up in the 90% area or possibly re-work the maps so that part throttle will cover even full throttle conditions. wont know just how much i can do there until i try it.

short of doing something like that i would have to make do with setting the fuelling safe by finding the condition/gear that the turbo spools up a the lowest RPM and tune the fuelling to cover that area which would make it run a bit on the rich side in some instances but would prevent any damage being done or the possibility of any overly lean mixtures. it wouldnt be ideal but then again it would be far cheaper than going aftermarket so with budget in mind and never being one to back down from a challenge that might learn me a thing or two in the process i think retaining the standard motronic would be my best option. i wouldnt rule out the possibility of doing something like the above tho, i have already sucessfully recalculeted RPM and load axis data and also increased the size of maps for better resolution as well as altering the switch over point from high part throttle to WOT, however, i dont know if the mechanicals of the AFM/engine will be a limiting factor or if i could simply just run the high part throttle map out to 90-100% load. there may also be a small and currently unknown maximum/minimum type map relating the engine load and switching point that i have yet to figure out but i can check that out by extending the load axis to full 100% load and tracing the map to see whether the ECU continues reading the modified part throttle map up to 100% or whether it still switches over to WOT maps at a certain load amount below 100%.

[eta]

One of the problems Gunni had when he tried getting motronic to read a MAP sensor (on my car) was that after the car run for a little, after it was switched off it would not start again. It was like it was logging some sort of fault code after seeing it's load sensor (MAP instead of AFM) behaving very strangley. So maybe a way forward with motronic 1.3 would be wipe everything including its fault finding and logging capablilites and re-programme it completley so air can be metered with a MAP sensor.

I still don't fully understand how motronic can add the correct amount of fuel without being able to meter the air when running highish boost (beyond 1/2 a bar) or any boost pressure using an AFM. Although BMW did manage the 745i with L-jet or was it motronic 1.0 but that was low boost.

Given MAF conversions are easier would a boosted engine run well if motronic 1.3 read a MAF sensor or does that bring its own complications (I cannot see any).

[Gunni]

First idea of running motronic is to cut the WOT wire so that it never goes into the WOT mode.

[eta]

Gunni/Geoff Bob would a MAF conversion on motronic 1.3 allow a boosted engine to function properly? I'm just thinking aloud becuse in my car the motronic 1.3 loom plug would not fit through the motronic 1.0 hole in the bulkhead unless it was made larger. To seal it a rubbery/tar like substance has been used which has set. I think to pull the plug though the bulkhead to wire in VEMS or something similar might be alot of effort or impossible without damaging the loom.

So a MAF conversion might be easier is it would allow metering of air and proper fueling under boost.

M50 turbo's work well so I assume an M20 turbo on a MAF would work also.

[GeoffBob]

I can say little or nothing on the Motronic side since I am unfamiliar. A MAF (hot wire anemometer) is an improvement over an AFM (swinging gate air flow meter), however, in that it measures mass-air-flow as opposed to volumetric-air-flow. To convert volumetric-air-flow to mass-air-flow requires temperature correction (to account for the density of the air) which is no big deal TBH, but will result in tuning errors if not done correctly. Barometric pressure must also be taken into account or tuning will adjust with changes in altitude.

FYI, MAP based systems are generally preferred by tuners for their simplicity and the fact that they present no obstruction to the inducted air, but even they cannot measure mass-air-flow like a MAF can, and must be temperature compensated by means of a MAT sensor. For the record, the only way an ECU can determine exactly how much fuel to inject is to know the mass-flow-rate of air inducted into the engine. Whether a tuner chooses to acknowledge it or not, even a MAP based ECU is effectively "calculating" the mass-flow-rate of air into the engine by means of the manifold absolute pressure and temperature. Interestingly, this is not actually calculated by the ECU, but is inherent in the tuned 3D fuel table (which, whether the tuner likes it or not, is an expression of the ideal gas law and the VE of the engine). It is interesting to note that if the ECU were to receive an accurate MAF signal as its input the 3D fuel table would condense to single 2D line of data that effectively relates injector duration to mass-air-flow. One could then add a percentage modifier to this to account for desired AFM (at specified RPM) as a function of a feedback signal from an O2 sensor. Above a certain throttle position the O2 feedback would be ignored.

TBH, I like MAF's. They are still widely used on GM (Opel, Vauxhall) and Nissan, to name a few.

[DmcL]

yea MAF is a good thing.. i have had my M20 running better than on either the M20 or M30 AFM. only catch is theres a mysterious little hiccup at/around full throttle maybe 30% of the time, its a single split second of hesitation before running fine again. sometimes it happens in the middle of a full throttle pull up a gear towards the higher revs and other times it will happen after shifting up to the next gear and flooring the throttle again, doesnt happen under normal driving or at part throttle and happened so far on 2 seperate tunes of the AFM maps as well as with different fuel/ignition settings in all the main part/full throttle maps. im wondering if it may be to do with the part and full throttle maps having slightly different RPM points in some places more recently because i read a thread about a rich spike in MS that looks to be caused by differing RPM values between maps so i might have to stick the MAF back in and re-code the full throttle maps to match part throttle RPM points exactly and see if it still happens or not.

as for MAP sensor.. have only managed to get the engine to idle and free rev on it but as soon as its revved it doesnt sound right, happened on 2 occasions with different tuning on the AFM maps also so might need some sort of signal converter between a MAP sensor and the ECU to run a MAP sensor..

@ gunni, do u mean the wire from the TPS that triggers the WOT maps?

[Gunni]

yea

[DmcL]

interesting.. never thought of that haha

i wonder if i could even just change the WOT map pointers or whatever theyre called in the index of map addresses to direct the ECU to the high part throttle maps instead of the WOT maps and extend the high load side from the standard 75% ish load up to near 100% without the ECU throwing a wobbler..

[Gunni]

Thats been done before and works.

[eta]

The throttle switch in the TPS is "on" at a certain throttle angle, that control the switch over to WOT maps, so even if the change the "indexing of map addresses" would that really make any difference. Surely it would be a whole lot eaiser to disable the WOT function all together.

[DmcL]

Thats been done before and works.

ah cool so no need to actually cut any wires then.. just redirect the ECU back to the high part throttle maps instead of the WOT maps and it will just run the last load column as if it were the WOT map?

this gives me an idea.. i have resized and reworked maps in motronic before. if i set it up so that instead of 2 smaller low and high part throttle maps i had a single part throttle map and then re-directed the ECU back to it for WOT i could keep everything in the one ignition/fuel map and not have to keep going between maps to trace what cell the ECU is reading at low, high and full throttle.

[Gunni]

No it´ll keep measuring from the AFM the whole time now.

You can simply make one 16x16 map or any size for that matter. that incorporates the part throttle and WOT throttle, the load axis just goes to maximum(maximum Ti, 255) while you figure out how much load your highest axis point should be.

Doing this you´ll actually see when the AFM reaches it´s flow limits.

[DmcL]

yea so basically make a large single map for fuel and another for ignition and run the load axis up to 100% or 255. id imagine it will run up the load columns till the AFM maxes out then just jump up to the full load end.. would make adjusting the spring in the AFM easier as it could be adjusted to run further up the load axis, only catch i can see even before trying that is to gain resolution up top with the AFM id be giving up resolution lower down. M30 AFM might be a good idea there..

[DmcL]

right.. another question for the turbo M20 gods..

i just picked up a pair of cams, will they be ok to use with boost? my logic is to put in a cam now and enjoy the extra power till such times as i boost the engine. at that point if the cam wont hinder performance with a bit of boost id leave the cam in. might it even increase power on lower boost compared to using a standard B25 cam?

the specs on both cams i picked up today:

Kent "Sports-R" 277

duration - 277 degrees
cam lift - 7.01mm
valve lift - 10.66mm
valve timing - 27/70 70/27
valve lift at TDC - 1.57mm
valve clearance - 0.25mm/0.25mm
full lift - 111 degrees

Schrick 272

duration - 272 degrees
cam lift - 7.0
valve lift - 11.0
valve timing - 25/67 67/25
valve lift at TDC - 1.6mm
valve clearance - 0.25mm/0.25mm
full lift - 111 degrees

[GeoffBob]

Longer duration and higher lift cams are not suited to turbocharged engines due to turbine pressure in the exhaust manifold forcing exhaust gasses back into the combustion chamber and inlet tract during the period of valve overlap. OEM turbocharged engines typically use shorter duration cams than the stock normally aspirated equivalent OEM engine to avoid this problem. With the exception of those fitted with large turbines (where back-pressure never exceeds inlet pressure) the same rules apply to an aftermarket fitment.

Straight from Graham Bell’s Forced Induction performance Tuning:

”aOnce we move away from the stock cam, turbo engines are the most fussy purely because more cam duration and lift takes away low speed torque. Therefore, when a broad power range is desirable I would say that in the case of road cars we should be looking at 10’ less duration, and for competition engines it would be closer to 20’ less, also maximum valve lift would be around 5% and 15% less, respectively, than for an equivalent naturally aspirated engine.”a

Also, with regard to exhaust back flow:
”aThis problem is much greater on engines where a smaller turbine wheel and housing is fitted to improve bottom-end power and reduce turbo-lag.”a

FYI, stock cam is best on turbocharged M20 for most (other than biggest) turbochargers. Longer duration/higher lift cam has some advantages on supercharged M20 provided fuel consumption is not an issue.

[DmcL]

was thinking it may not be ideal.. would either of those cams suffice for running up to maybe 1 bar of boost on a T3 turbo? id be fitting the cam with a vernier pulley so i could advance/retard timing. would there be a way to eek around the overlap issue with either of these cams and adjusting the cam timing via the vernier pulley?

my logic here is to potentially make more power with less boost than the typical M20 turbo with standard cam if possible.

suppose if its a definate no i could always swap the standard cam back in come turbo time.

also there seems to be a bit of scaremongering going on in the US E30 forums about installing an aftermarket cam. alot of people are now saying its near mandatory to install an aftermarket cam with a vernier pulley otherwise valves will meet pistons. i dont know if i would actually believe that or not tho.. any input on the matter? im thinking a few eejits put a cam in on an engine thats had the head skimmed or block decked and turned it over without checking or something similar.. couldnt see a mild cam like a 272 or 277 causing valve/piston contact with a standard pulley and set at the standard markings when timed up with a new belt after the cam install.

[ross2009]

you could alter the cam timing but youll never alter the lift duration of any cam as they are what they are a 272 wont hit anything over i believe a 288 will hit the best bet for your turbo engine with out spending money will be the standard cam ive got a turbo grind cat cam in mine but that was near on 400 quid

[GeoffBob]

was thinking it may not be ideal.. would either of those cams suffice for running up to maybe 1 bar of boost on a T3 turbo? id be fitting the cam with a vernier pulley so i could advance/retard timing. would there be a way to eek around the overlap issue with either of these cams and adjusting the cam timing via the vernier pulley?

Definitely not on a T3! Remember, a longer duration cam gets you more top-end torque on a normally aspirated engine by exploiting resonance in the inlet and exhaust runners to improve the volumetric efficiency of the engine. Bolt a turbine to the exhaust manifold and all that resonance in the exhaust goes out the window (or at least shifts to an altogether different/pointless resonant frequency). Fitting a longer duration cam (or at least one where valve overlap occurs) to a turbocharged engine with a small turbine (such as a T3) serves no good purpose whatsoever.

Vernier pulley on a SOHC won’t help you either since it is the LSA (lobe separation angle) that would need adjusting to reduce exhaust back-flow.

couldnt see a mild cam like a 272 or 277 causing valve/piston contact with a standard pulley and set at the standard markings when timed up with a new belt after the cam install.

All depends on the cam lift and what they did with the rocker arm ratio.