turbocharger upgrade with roller bearings

Good luck with that.........


Snag a ball bearing turbo with updated aero and a good map for your application, and be done.

 

turbo lag is a function of compressor efficiency and integral to the volute design and impeller trim. I.e. bearings are not going to solve it. The turbo's of that era were not anywhere near the efficiency of what is available today, coupled with the fact that more then likely it was designed for high RPM and peak power the turbo would be 'bigger' then needed and lag is a result.

The lag comes from the time it takes the engine to create enough exhaust pressure to spin the turbo and pump compressed intake air into the engine and is longest when the engine is in a low-rpm, low-load cruising situation.

If you’re wondering why somebody doesn’t make a turbo that makes full boost from idle on up to redline, there’s a very simple answer for that: like all engine parts, a turbo has to be matched to a specific rpm range to function correctly. A turbo that’s small enough to make a significant amount of boost under low-rpm situations would overspeed and possibly explode as soon as full throttle was hit. The opposite is also true—a turbocharger that makes the most peak power will make virtually no boost until well into the engine’s powerband. Most turbo setups are a compromise between these two scenarios.

 

Alberto,
no experiences with turbocharger upgrades, but with high speed bearing applications.
Like you probably know, the high revs like on a turbocharger turbine wheel are the main lifetime limiting factor of ball or roller bearings. One of the reason for preferring friction bearings. Make a lifespan calculation and you will be amazed.
I once developed a bearing design for an application with similar revs. We doubled the ball bearing in pressing a small roller bearing into the inner race of the bigger outer bearing. This means a relative movement, or better said, a partition of the rim speeds.
Of course it depends on how much space you have for something similar.
Maybe needle roller bearings could find enough space.

Have fun designing
Martin

 

Smg,
what do you think about less friction of roller bearings vs. slide bearings on high speed applications. Somehow high revving japanese motorcycle or tiny ultra-high revving car engines with roller-bearing crankshafts come to my mind.

Best Regards
Martin

 

I dont think you would eliminate lag in the 208 as the design is a massive compromise.

A V8 should always have 2 turbos but they piped the front manifold around into the rear with a simple T piece before the turbo.

 

If you are going to change the turbo, then put the most sophisticated turbo you can and measure the difference.

 

Alberto,
do not compare apples with oranges. On your Mercedes you have a slow revving Diesel engine with high low rev torque designed 'from zero' for turbocharger appliction.
On the Ferrari 208 you have small high revving low torque engine on which they grafted a turbocharger much later. I agree with Andy. That's a massive compromise.

Best Regards
Martin

BTW; I love technical disputes

 

Funny you ask at the moment I'm designing a scavenge pump housing for helical spur gears. Normally the oil pressure pump side can be used to feed the galleries and that tends to solve the issue in a relatively simple way, provided that the oil doesn't cavitate, aerate or run low. Sheer is also a concern.

I've looked at PTFE, PTFE impregnated bronze, graphite 66 nylon etc.. while also using the pressure side of the scavange to feed the galleries for a quasi splash lubrication. needle bearings are the other choice and I'm leaning towards that approach.

For loads that don't experience tangential or perpendicular to the axis shock a roller bearing is better suited. Thermal considerations are next, turbos are HOT and the frictional induced heat in the bearing poses a problem for most bearings, ceramic bearings are a good choice esp in turbo housings. Oil film bearing has been the std but requires proper clearance, flow pressure etc.. and of course sheer is always a problem esp with shared engine oil.

surface drag is present in oil film bearings as the Ra of the two surfaces are usually very small. depending on clearance gap and the differential velocity of the two surfaces and the viscosity of the oil, drag vs sheer become the balance point.

of course this is only a light dusting of a very complex and intensive subject.

 

Turbo lag is a function of pressure deltas.

I was not implying that the 208 turbo was poorly designed at all. As mentioned it's a compromise, pumping efficiency, flame kernal spread, intake and exhaust port length, etc.. all mean that an engine designer has to pick a 'narrow band' of the full rpm range that is desired for the most power. Technology and engineering as come a loooooooooong way since the 1980's and the useful rpm window has been increased.

Lets get back to the lag issue though. a non-inter-cooled turbo engine will suffer from lag over an inter-cooled one. Air density is the main reason, coupled with the engine management trying to reduce the risk of detonation in the non-inter-cooled engine by timing retardation and extra fuel for cooling it becomes apparent the engine will not be as responsive and that in turn adds to the lag issue already in play with a small engine and large turbo.

Your best option is to upgrade the turbo to a modern unit that is far more efficient and sized properly to the engine, anything else is a waste of time and money. If the concern is originality of factory OEM parts, then leave it be as modifying the turbo will only lead to more problems that need to be fixed down the road. I'm sure you're aware of the impeller RPM in a turbo, modifying the bearing the bore for such a high-speed shaft is not a simple thing, I can only imagine the setup and jigging required for a machinist to attempt that.

 

Depends really, most times on small engines the single turbo will outperform the twin setup. Since trying to balance both banks independently would be a nightmare the intake plenum is shared and as a result the two turbos end up being like a single massive turbo. The trick is finding two small turbos that will work over the engine rpm range and still provide the right amount of air volume at specific pressure without being either too small or too large. So far for the 308 engine that's a tough one as there are loads of single units that are perfect for the displacement of these engines. Helps that there are lots of Japanese 3L engines with turbo's

 

Agreed.

In my experience going from journal bearing to ball bearing (on the exact same turbo) has a -very- slight improvement in part throttle transients. Zero gain under WOT. (as far as response time).

Not sure what the compression ratio is but if it's like any other turbo engine from that era it's not enough.

 

I think you just need to have your turbo's rebuilt with 'zero clearance'. Essentially the turbos are built and broken in so that the compressor wheels have zero clearance with the housing, improving efficiency and decreasing lag. Common upgrade in Porsche turbos (mine included). (just google 'zero clearance turbos)

 

Yes indeed, the shared plenum approach would not work, there would need to be a separate plenum for each bank.

So, separate plenums, 2 turbos, you would need to make the engine longitudinal because the front bank turbo would have no space.
Result: 288 GTO.

 

Kevin at ultimate motorwerks does the work (web site link you posted above - if you can't reach him though the web site contact, let me know and I can pm his phone number). I believe most of us on the 993 turbo and 996 turbo forums who have had turbos rebuilt have had this process by him. He essentially asks you how you want to use your car and your hp goals and will build you turbos and ecu to take advantage of the turbos. I've had k16 rebuilt as k16/24 hybrids (with the zero clearance) on my stock 993 twin turbo engine and in my current modified andial motor have I believe k24/26 hybrids. He is busy and sometimes takes longer than promised for the turn around but worth the wait.

 

Garrett claims up to 15% faster spool speed using ball bearings. It also holds the shaft end play and run out to a tighter tolerance which allows tighter clearance in compressor and turbine.
Image Unavailable, Please Login

 

Whatever you end up doing, get dyno results before and after so you can document exactly the result (and post here)!

 

Alberto,
the most reliable/ cost effective way to improve transient response of your K26 for the 208 is to get first a brand new turbo. The guys who were supposed to implant a bearing cartridge to replace the original bearings should be able and capable to measure the clearance between impeller and compressor housing as well between turbine wheel and turbine housing. Then they would have to set the clearances a bit tighter , nearby to diesel specs. You will notice a clear improvment in the acceleration, assuming every bit of the fuel injection system is on spec.
This trick was common practice for assembling performance turbochargers at the time when no zero clearances or plasma coating or BB were industrially available. Consider all racing turbochargers of the 70-80's were assembled as hybrids with compressor side mutuated from commercial diesel engines ( big size to achieve high air flow ) and turbine side from gasoline pass car engines ( small size /low inertia to get quick response), however with "diesel" clearance setting. As last step the complete rotor should be re-balanced .You cannot do the trick on current K26 because I assume is well used and the turbine side parts may be distorted/cracked because of the exhausts temperature. Of course you cannot expect a 100.000 km turbo life but for your Sunday tours around Modena would be just fine .

 

A brand new turbocharger is available at eurospares.co.uk

http://www.eurospares.co.uk/searchResult_one.asp?S=117714&TM=1&TMo=672&TA=1&TB=39525


Best wishes
Markus