GT3 Oil pump dissection

[Cupcar]

Below is a dismantled GT3 double scavenge oil pump.

Of interest is that the pressure pump gear, on extreme right of photo, is 51 mm long, identical to 3.3 Turbo Pump's pressure gear.

The longest scavenge pump gear is identical to the pressure pump's 51 mm length and the shorter gear is 33 mm long, giving a total scavenge gear length of 84 mm.

A Turbo 3.3 pump has an 80 mm long scavenge gear, so GT3 is 5% greater in scavenge capacity.

Of interest here is that a GT3 has two other scavenge pumps, one for each cam box, which pump directly to the oil tank. Thus the crankcase pump in the GT3 only has the oil from the crankshaft to deal with, hence it creates a negative pressure in the crankcase which, with a one way negative pressure ventilation valve, is supposed to be good for 6 to 8 HP (and rear main seal leaks due to the negative pressure)

[356RS]

Great information. Thanks.
Is that Adcom your shop Amp?

[Cupcar]

Originally Posted by 356RS

Is that Adcom your shop Amp?

Yep, I have Adcom electronics and Monitor audio speakers to relax me in situations like taking apart this pump which, if you look closely, has gears that are toast due to a bad day at the races.

[vmisquez]

Thanks for taking time time to share.

[MBEngineering]

HI Cupcar
"if you look closely, has gears that are toast due to a bad day at the races."
this GT3 pump also had a bad day at the races, well rally, took the end off the intergear also.









sorry to hijack the thread.

regards mike

[Cupcar]

Like I tell my wife, no matter how bad my day has gone, someone has one worse than me!!!

I am surprised the case of your pump didn't split down the middle or that your countershaft splines were not stripped.

[MBEngineering]

HI Cupcar
the intergear/countershaft splines did not have time to strip, thay just stopped with the pump,







regards mike

[Cupcar]

Mike - How did the nut get into the oil tank, or dare I ask...

Our situation was a #2 rod bearing failure...that took out a Ti rod...and the crank...and the side of the case...along with the pump and countershaft

[tadd]

Cupcar:
I know this might sound crass, but any idea why? You have a huge pump and Ti rods; thus I should think the cross drilled crank, enlarged oil passage to that main, and all the other goodies... yet #2 still starved?!

t

[Cupcar]

Cupcar:
I know this might sound crass, but any idea why? You have a huge pump and Ti rods; thus I should think the cross drilled crank, enlarged oil passage to that main, and all the other goodies... yet #2 still starved?!
Well, it even had cross drilled center mains, so how does such a thing happen when you just went from a small to a large oil pump?

Answer: The engine is in a sports racing car and has about a 3 foot long -16 suction line from tank to pressure pump. This line was sucking itself shut and depriving the engine of oil at high RPM when it needed it most.

Amazing thing is that the engine went through two race weekends before problem was figured out without blowing and the other main bearings show essentially no wear.

The first "symptom" was random overflows of the oil tank when the crankcase was totally emptied with no inflow from the pressure side.

The "physical finding" after the fact was the in-car video which shows the paradoxical motion of the oil pressure gauge at the end of longer straight sections of track.

Take home lessons:

1) There is a reason for those dry sump suction line support coils

2) It is hard to see the oil lights come on in bright sunlight and a glance at oil pressure is as important at the end of the straight as it is at the beginning.

Yes, it had a cross drilled crank.

[jcge]

Cupcar / MBEng (Mike)

Thanks for posting those details of failure modes

A question related to the flow comparison based on lengths of the pump gears sections...

I assume the gear OD's modules and teeth count are consistant across the models ?? (BTW, what are the gear OD's ??)

John

[Cupcar]

Originally Posted by jcge

Cupcar / MBEng (Mike)

Thanks for posting those details of failure modes

A question related to the flow comparison based on lengths of the pump gears sections...

I assume the gear OD's modules and teeth count are consistant across the models ?? (BTW, what are the gear OD's ??)

John

36 mm diameter, 10 teeth. I think this is the same from 356 through all 911 engines.

[jcge]

Thanks Cupcar
John

[kenikh]

Originally Posted by Cupcar

Answer: The engine is in a sports racing car and has about a 3 foot long -16 suction line from tank to pressure pump. This line was sucking itself shut and depriving the engine of oil at high RPM when it needed it most.

Wow...just WOW. Sorry to hear it.

[tadd]

Cupcar:
Thanks for the answer. Wiping a motor is a pisser, but I appriceate you passing on what to watch out for. I'll be using a trunk mounted tank in my car and I didn't even think about anti-collapse springs...
t

[stownsen914]

I have heard of the anti-collapse oil feed line springs. Where do you get them?

Scott

[Cupcar]

I talked with a guy at XRP hose, and he pointed out that as braided stainless gets older, the small projections of rubber that one sees poking through the stainless braid in a well made, new hose tend to come undone from the SS braid over time in a suction hose, thus allowing the inner hose to more easily collapse under vacuum.

The hose in question here was over 10 years in service

XRP are a source for support coiling BTW.

[Cupcar]

Originally Posted by stownsen914

I have heard of the anti-collapse oil feed line springs. Where do you get them?

Scott

Page 34 of the XRP hose catalog:

[tadd]

You can also get hose with the winding (spring) already molded in the hose. It's used for vacuum systems on the roughing side. If you do run an inner anti collapse spring, don't forget to use a wiggle or bend at the end of the run so it can't go anywheres.
t

[Cupcar]

We are going with the XRP HS 79 with a coil, despite that HS 79 is supposed to be excellent for vacuum without a coil.

Note that the ID for the HS 79 in -16 larger than the usual, so even with a coil, it will have a larger ID than the -16 that failed.