The valve seat has nothing to do with it. On a port-injected engine, the fuel is introduced into a moving air mass at reasonably low pressure but a high degree of atomization, this promotes a very even and homogenous blend of air/fuel. With the direct injection charge, the fuel is sprayed, at high pressure, at the top of the piston, where it deflects and ends up on the cylinder walls. If those walls are cold, this results in droplet formation and subsequently liquid fuel on the walls, which of course washes-down the oil, and makes its way into the sump.
Unless you sampled the oil and sent it to Polaris/OAI, you'd not be aware of how much fuel was making its way into the oil. That said, certain designs are far more prone to issues than others. The Honda 1.5L seems particularly bad, the BMW B58 seems particularly good.
Thank you
OK for the links, some more late night reading for all of us who are interested in this subject.
More of my thought and points:
On batch or even SMPI the injection period is not always into a 'forward' moving air mass (depending on load) and partial obstruction of an opening valve will compress then expand the charge which may cause a % of droplets to condense or fall out of suspension - but I agree that running
λ = 1.0 at high throttling with the injection period during the valve opening event does promote good fuel atomization with mixture swirl or tumbling depending on valve actuation. This mixture will be exposed to the full area of the cylinder wall be it cold start or fully warmed. Any condensate will be wiped on the up stroke. We have all read reported dilution issue with SMPI for those who idle the vehicle at warmup - especially with starts and coolant below freezing.
It appears side-positioned D.I is at its best at economy cruise, say
λ = 1.1 were the injection pulse width is >/= 10ms
and the fuel enters when the piston is at its upper 1/3 of compression stroke travel and the piston top contour - if so employed - will deflect fuel away from the opposite wall. In this scenario there is simultaneous charge phase cooling to prevent pre-ignition and a mostly concentrated charge away from distal surfaces to prevent detonation. An issue with D.I is when the fuel must be injected during the intake stroke and if side positioned and vectored it will impinge and likely wash the opposite wall - especially with long stroke narrow bore engines. Now maybe 1/3 of the ring circumference will be running up along a "dry" cylinder bore.
D.I. cannot inject to great effect under high load/high rpm during exhaust and compression "off--cycle" as SMPI or Batch can,
So we see high velocity/volume injection attempted in milliseconds during un-throttled high load scenarios.
Again this is all specific to any particular design, geometry and timing.
I see some BMW with the Bosch system use a center injector near the spark plug. I wonder what other engines use this approach.
This would be optimal along with SMPI if it doesn't impact valve sizing to a great degree.
My FORD uses side positioned injectors, and is likely an older, sub optimal arrangement.
But I only say that due to my current dilemma.
I see one SAE paper above was presented by engineers from
Jiangling Motors Co., LTD., China
The hotbed of research and development. Looks like a discussion of an early implementation. I don't disqualify it, I will read it.
On oil sampling, I agree without a proper lab test I would not have presentable data on my Jetta D.I. Turbo
fuel dilution tendencies.
But, after initial wear-in, I experienced no increase on the dip stick, no ugly appearing oil, and no engine racket
when warmed. Maybe Fat and Happy, but you see I only fret the discordant effects I see and feel.
Or, in other words, I don't go looking for trouble, it finds me all on its own
Good discussion.
- Ken