Alan

A quick question because I'm confused by some things you have said.. is your Provent located before the GZ pump or after...?

Alan

blau928

It is located after the GZ pump, as the pump needs oil to work. The pump pushes the crankcase gases through the PV200..

Hilton

Interesting discussion again in this thread.. yay!

Richard - where are you planning on putting the PV400? Its 6" diameter and about 10" deep - a lot bigger than the 200 model.

Given the experiences of others (Alan, Ducman, Colin) it certainly seems the common Provent isn't up to the job.. cyclonic does seem viable (I hadn't seen the one linked above - but I've built a larger scale one for dust extraction from my blast cabinet and figured getting a small one welded up wouldn't be hard).

Since my original random thoughts earlier in this thread, I've gone through a couple of design revisions, and settled on two possible alternatives. Both of which have some drawbacks, but which are fairly similar so I can use empirical measurements (vacuum measurement of manifold and crankcase and data-logging) to determine final outcome.

Below is my preferred - the design takes metered fresh air from the MAF boot before the throttle plate, using a pair of vacuum-cracking check valves (likely 12" Hg), and feeds back into the intake - so its a closed system.

Tube sizing is, as noted, important to modulating crankcase vacuum, recirculation flow etc. - I'm guessing an oil drain ID of 10.5mm or so, and an air return to intake which is 23mm or so, with a pump inlet of 1" from the oil filler neck (metal one). I'm doubtful the factory ports on the sides of the MAF boot are large enough, but thats pure gut feel rather than actually anything calculated at this stage.. but if I need to I'll have bigger ports welded to the throttle body before the butterfly.

Given the pump specs on GZ's site; their graph starts at 1600 pump rpm, but I'm guesstimating 4-6cfm with a 1:2 pulley ratio at idle speed (i.e. ~385rpm pump speed), I think this design will work. Engine requirements at idle are around 1650l/min (assuming VE 85%) at idle rpm's, so a pump pushing 120-170l/min of air into the intake won't swamp the idle-control system. At high engine speeds, based on pumping requirements at 6700rpm (redline) of around 375-425l/min blowby gases (depending on assumed VE 75-85%), which per the GZ graph needs pump rpm's at around 2500 - so a 1:2 pulley ratio giving around 520l/min pumping capacity (18.5cfm) should be plenty.

What I'm not sure is how much fresh air should be pulled in through the cam covers - the goal there is to keep flow of air from the heads to the crankcase so that any time a plug of oil blocks a drain, it is broken up by air travelling down the drains. From rough calcs, a 1:2 pulley (6" pump pulley) ratio will give 100-125 l/min of fresh air sucked through the vacuum check valves at redline, once crankcase vacuum has reached equilibrium, and depending on actual VE of the engine at high rpms.

Alan - what size pulley do you have fitted to the GZ pump? And do you have any idea how large a pulley will fit there? (PS - I saw your radiator shroud mod - nice place for the separator!)

Hilton

I decided I'm not willing to trust the air/oil separation enough to risk putting air back in front of the MAF - if any oil fouls it, it will run progressively leaner.

Also - if you're regulating crankcase vacuum with the hose sizes, you're presumably tuning the system for the most efficient point on the GZ pumps performance curve? How much effect will a drop in the GZ efficiency have on crankcase vacuum at lower/higher pump rpms?

Some links for those reading along:

Provent models catalogue: https://www.mann-hummel.com/industri...BVFLGOcGNM.pdf

GZ Pumps spec chart: http://www.gzmotorsports.com/images/...flow-large.jpg


Cool to see you and Alan both have this stuff installed. I'm hoping to get closer to final design before I order all the bits. My biggest question is the separator still..

Alan

I said:

If the pump seizes and the crank is otherwise sealed - then I think the blowby will blow out seals somewhere.. where else is there to go? I don't understand what other paths exist?

And

So how does the blowby bypass a seized pump? or are you just assuming it routes backwards through whatever restricted orifice you use to control vacuum level.

Note you also mentioned you don't think your crank is very well sealed - if that is the case I think you'd be seeing notable oil weepage at those spots since the stock crank does marginally pressurize at WOT.

When I develop something I like to go through what I think would happen for at least all the single point of failure issues, and anything else I can imagine with potentially catastrophic impact.

I did not previously mean to imply that the plumbing and configuratiuon aren't important to the system operation - of course they are - especially at the max cases (RPM/flow). However solely for the question of whether you will see more oil than the provent can handle - I don't think it matters - I think you will see this in any configuration that otherwise works. That was my point... It seemed irrelevant to me to discuss detailed implementation if in fact it wouldn't affect the question of overoiling the Provent. Seems we already got too deep for probably 95% of folks anyway.

My final reconfiguration isn't done - so I'm not ready to talk about it - but you already know from what I said that it doesn't rely solely on a Provent 200 for separation.

Overall the main point was that to my knowledge of 3 people with a GZ vacuum pump system (you now make 4) none of them are running without issues through a Provent 200.

Anyone else have this configuration - with what results?

Richard if your system does work when driving please share more details. Maybe the Provent 400 will work well enough? I still think its the wrong kind of primary separator and it is indeed truly huge... have you seen one of these things? I think it would have to live outside the engine bay

Alan

Alan

I think it is possible to get clean enough - but we shall see - certainly I'll test to atmosphere (w/ capture) first before taking any risks.
If you tune only with fixed orifice relief then you have to set for the max desired vacuum you want to see and live with the consequences (lower vacuum at low pump speed and at under high blowby conditions) - Empirical.

However you can also use other means - vacuum limiting valves, Variable orifice, multiple orifice diverter selection etc

For more reference points Colin (Lizard928) and Jake (Ducman82) both also have GZ pumped systems. The general configuration of these (and mine) is evacuation from the crank vent only and recirculation flow back into the cam ports. Colin and Jake are atmospheric vent - mine intends not to be.

Richard's sounds more similar to stock - evacuating both heads and crank to the pump inlet.

Looking at your proposed schematic - getting suffficient sized in/out porting between the MAF & TB is challenging - its small and you need the booster connections too.

In a pumped system you may not be saving the MAF. Running high speed, decellerate - throttle is closed - pump is still dumping flow between the MAF & TB - it pretty much has to back up into the MAF anyway... yes its part time not contiuous - but it actually happens frequently. So challenging to do and likely not ideal even if you can do it.

In a stock system high speed decelleration => blowby stops, crank pressure dissipates with little flow...

Alan

FUSE69

Has anyone that is running a vac pump considered using a BMW M5 cyclonic oil separator... And then perhaps the outlet of that into a provent?
Could be an interesting idea??
M5 actually uses 2, one plumbed to each head.

Oil separator LH 11151406788
Oil separator RH 11151406789

Alan

I did look at those but one is clearly too small... even 2 may be marginal.

I do think a more open centrifugal separator in the main pump flow is a good idea - have it handle the majority of the oil stripping and high flow back into the recirculation loop (low restriction) and then use a filtered medium separator on the net output flow.

I do think the efficiency of this type is likely very dependant on the configuration & flow rates - so lots of experimentation would be needed to evaluate that.

Alan

blau928

Hi Hilton,

I can comment on based on what I know and or have done so far based on the details you provided.. I am glad to see that you at least get the pulley ratios and pump speed as well as line size will affect the vacuum in the case, as well as flow in the system.

My head lines are currently plumbed into the smaller side of the metal filler base, so when oil comes through, depending on vacuum, it will be going to be in the crankcase, or going out the evac line to the pump for a spin around the provent and back to the sump based on the differential pressure in the evac line, and motor, the oil will be sucked back into the sump. Again, I want oil in the evac lines to keep the pump happy, but not too much as to make the Provent unnhappy.. (On the setup for the SC with the custom oil filler, it is also this way.)

Consider this... If you flow more than 200L/min, the Provent will become pressurised, and it will not filter the gases properly, and oil will end up in the outlet... The limitation on the PV200 is nothing but the filter's ability to flow air. If you overdo it, the oil in the gas mixture will not be absorbed by the filter, and will end up ruining the MAF.. Think about it..... When you pressurize the filter, what will happen...? Will it have the same effectiveness of filtering..? I doubt it.... The provent 200 has 1" dia pipes, it will flow a lot of oil, gas etc...

IMHO, this is the difference why the setup is expelling oil in the outlets for people who have them installed now. Even the VW guys that have cranked up their turbos and swear by the Provent notice this on their Golfs etc. and use 2 PV200's. I really do not understand why this is so hard to grasp for some of us, as it is simple physics..

Here is how the PV200 is currently mounted on the car, hope that helps...

blau928

Alan

If you feed your recirculated pump flow down the head drains - you don't need to need to be so concerned about maintaining that fresh air flow at high rpm/load for oil control. I think that will be just impractical to maintain since it would boost the maximum flow so much. There is no need to flush continually otherwise.

In practice the pump flow is fresh air intake, blowby, valve seal leakage, leakage air, and deliberately recirculated flow for vacuum control. Unless you are deliberately introducing lots of fresh air - the pump flow will typically be limited to much below its free air pumping volume specs to generate ~12 "Hg in the crankcase (except when blowby peaks - but in a fixed orifice system the vacuum would drop then too).

I have a 5" pulley on the pump and a <4" crank pulley. I think a 6" pulley is going to be a tight fit - not sure on that...

Yes the shroud mod is all part of this (I want it to be as stealthy as possible too)

Alan

Hilton

My questions were more out of curiosity at Richard's setup. I'll be using some check valves - Grainger make a couple of sizes which look to work.. e.g. 3/4 sized one:

http://www.grainger.com/Grainger/CDI...e-Relief-5Z765

GZ sell another candidate:
http://www.gzmotorsports.com/VCV103-...rol-Valve.html
(which looks to be the Grainger 3/8" one - at double the cost?)

And Moroso make one:
http://www.jegs.com/i/Moroso/710/22630/10002/-1

My return from the pump will be into the intake (i.e. after the throttle plate) - only fresh air will be pulled into the cam covers from the MAF boot. Hence I mentioned that the pump output will not overcome the idle-compensation circuit.

I was figuring on adding one or two return ports to the front end up the throttle body, with protruding ends angled so that flow from the throttle up into the intake manifold will encourage scavenging of the return line. The throttle body is aluminium so can be welded. I'm pretty sure there's enough clearance for a pipe there above the front knock sensor, around where the regular breather hoses join the oil filler base.

If the pump fails, intake vacuum will continue to pull blowby gases out of the system (up the oil return line of the separator if the pump seizes).

Thanks for that info. I measured the inner circumference of the airpump pulley on the crank side to be around 3 inches (74mm) - which as the driving surface is what I assumed would be the relevant diameter. Am I measuring the wrong point? (hmm.. I should go measure the crank pulley on one of my S4's - the pulley I measured was an extra I had kicking around in the garage - from an S3 IIRC. They might be different).

From calculations, assuming GZ's quoted pump performance is accurate for the environment, a 1:2.8 pulley ratio would be around where I expect the pump to meet the blowby gas evacuation needs. Using a 1:2 pulley ratio gives a fair bit of pumping volume headroom to allow for pump performance drop-off - I'll have to take some measurements on my '87 which actually has an airpump fitted to it, to see what clearance there is.

blau928

Hi Alan,

Just a comment on your note above.. Hope this helps to understand the point I am trying to make..


At this ratio of 0.80 (5" pump pulley to 4" crank pulley), the Vac pump will be turning at 5000 RPMs when the engine is at 6000 RPM. This is not recommended by GZ unless the engine is producing 600-750 HP. (Granted I do not know if your car makes this kind of power.. I am assuming it is a stock GTS. Even if supercharged with a Murf kit, you will still be around 500HP give or take depending on tune.) In addition, it is recommended to keep the GZ pump RPM less than 5000.

With your pulley ratio, the GZ pump will be pushing 22CFM at 6000 engine RPM per the mfr. This is 623 Liters/min of airflow, and will overwhelm the PV200 capacity of 200L/min, and pressurize the filter causing oil to come out the exit tube as the PV200 cannot handle this airflow (Per Mann+Hummel spec sheet)...

The recommendation of GZ is the pump ratio to engine speed should be 0.54, or at 6000 engine RPM, the pump would be turning at 3240 RPM at about 17.5 CFM of flow..

Currently, at 4000 Engine RPM at your setup, you will be at 3200 Pump RPM speed, or what GZ recommends the ratio should be. At 3200 pump RPM, the GZ pump is pushing 496 L/min of flow (17.5 CFM). This is 150% of the flow rate of the PV200 @ 200L/min.

You mentioned prior that currently, your PV200 seems overwhelmed at about 4200 Engine RPM.. It may be that this is what is causing it..

IcemanG17

guys.....you might be making this far more complicated than it needs to be...... Yes there takes a certain amount calculations to make it work correctly.....and boosted cars making high HP or cars with larger bores....

Keeping things "smog" legal gets insanely complicated as well....

the two single most important issues that make our 928's so sensitive to crankcase pressure are a horribly designed breather system than became more and more un-engineered as the cars "developed" The early breather setup was "okay" at best....it only got worse....

Another HUGE issue with any 100mm+ bore is ring flutter.....even slightly worn rings can flutter significantly at anything past say 5000rpm....by 6000 its VERY risky..... Keep in mind our ring-pistons were designed in the 70's....not exactly high tech by today's standards....

Keeping the crank counterweights so close to the oil pan doesn't help either....

One interesting lesson I learned this weekend was it IS possible to kill a 928 engine with extreme overheating.....previously I had overheated the Estate engine to above 230F with no ill effects.....another 16V 928 engine was overheated to the point that it nearly stopped running from a busted lower radiator hose (no steam to see)....while the car runs and seems fine until 3000rpm...anything above that it spits oil at insane levels.....rings are GONE.....near 100psi per hole.... in 8 minutes on track it spit 1 GALLON into our improvised catch can

ptuomov

Why is the cylinder bore diameter a factor in ring flutter? By my math when computing the ring flutter onset rpm, everything else cancels out other than ring vertical width, ring material density, and the ring pretension.