When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
I've spent most of the day researching the purpose and function of the AOS. Put shortly it reduces the over-pressure inside the engine and recirculates that air into the intake(recirculation is due to environment/air quality). It's not a stright pipe like on most cars but it passes the Air Oil Separator which pulls out all the 'gunk' from the air so that clean air is feed to the intake. This is good, becuse oil-mist lowers the octan which potentialy can cause problems like knocking.
Well, so-far-so-good. But, what does the AOS do with all the 'gunk' it takes out from the engine air? It dumps it down into the sump, mixing with the engine oil. not. good. at. all.
Here you can see what 3000mils of 'gunk' look like: https://www.youtube.com/watch?v=tWLYRvBhJgQ
Ok, I'm thinking a catch can would be much better. I prefer to drain some gunk from time to time rather than having it mixed with the engine oil. After a lot of googling I found someone who had done some research to how well they work: https://gallery.mailchimp.com/807089...Comparison.pdf
Seams fairly easy to re-route some pipes and install a can? Anyone done it?
Hence the need for 4000 mi. oil change intervals...
After a small de-tour to the AOS thread I'm back here. My proposed system looks like the attached picture.
What differs it from your normal catch can setup is the dual vacuum source: air box before filter and normal place behind the MAF. The car has a ECU map which expects some by pass air. By having a long and thin tube going there we can by pass some air to satisfy the ECU. I will be using I.D 3mm PVC tube. Tuning: I will look at the fuel trim on idle and shorten/lengthen the by pass tube to have close to zero fuel trim. The new catch can will dump it's gunk into a container which needs to be emptied from time to time.
The normal AOS sump return hole will be plugged. The normal AOS coolant line I'm not sure what to do with, maybe add a small radiator on top of the engine?
Edit:
There will be a check valve at the air box connection to prevent the by-pass line from sucking air from there.
Last edited by Nick_L; 11-15-2019 at 11:06 AM.
Reason: See post
I think if you connect the the catch can to the air filter box you will end up with a massive crankcase vacuum leak and the engine will not run right.
When you way you've dome research, does this mean you just Googled for hours or actually did some testing?
Before I do some testing I need something to test, right? I read up how it works and getting a good start for testing.
Connecting it this way will create a little less vacuum than the standard system, mostly at idle. At full opening(of the throttle body) the difference will be minimal. At idle it will be the biggest difference since the vacuum is the biggest. 'Silk' in the AOS thread have his AOS plumed to the air box (and without the by-pass tube) and seams to work very well for him. Mine will be closer to OEM vacuum than his setup.
While your system looks like it will work in theory, how do you plan to balance the vacuum pressure? Are you just going to "trial and error" it with different diameter hoses until you get it right?
What you want is a slight vacuum in the (vacuum in this sense is relative the atmospheric pressure...that is, less than atmospheric but not complete vacuum) system so that you will avoid over pressure inside the engine. You need to avoid too much vacuum and higher than atmospheric pressure. It's not more complicated than that.
While writing this I realised that my check valve that I 'Edited'-in in the post above will create too much vacuum if the by-pass tube is not correctly dimensioned. With the check valve in place the tube is the thing which mostly controls idle vacuum in the system. Otherwise it will be the vacuum in the air box. On second thought I might skip the check valve since it might make the system less robust. The small possible gain in vacuum might not be worth it.
When connecting to the air box you will definitely have lower than atmospheric pressure, but not by much.
I need also to add: The pressure in the system is not the same due to flow. When connecting to the air box the beginning of the system will have slight vacuum but you might have slight over-pressure in the engine. This is because of pressure drop/increase in the hoses, in the catch can etc. Larger diameter hoses, shorter hoses or less flow will make the pressure difference less. So, when using a weak vacuum like in the air box I will use large diameter, and short, hoses for the venting.
So the short answer to my question regarding "trial and error" is - yes. I think a way to test this (at least initially - assuming your AOS is still good) is to take a baseline at your oil fill with a manometer, and then build your catch-can device to that baseline. At least you'll have #'s to look at as opposed to guessing (if memory serves 5-7" is the manometer baseline for a properly functioning AOS?).
It makes the following improvement iver the stock AOS:
+No risk of hydrolock in case the AOS break +Cleaner air gets re-circulated (less oil in it) +No bad oil re-circulation in the engine +Easier to service
On the negative -More labour, one more filter(every 40000km) to change and oil(5000km) to empty on regular times -Not original (but for me that is not important)
And then we have the discussion of engine problems with the M96 engine. Having bad oil to lubricate the internals doesn’t help. And using it as an oil-burner(read sot) also doesn't help.
For me this is the perfect solution as the oil in the engine is kept as clean as possible.
I do like numbers, I'm an engineer after all, but it's not necessary in this case. I know from the start I will have less vacuum than stock. That's beacuse I will not run the return to the intake, but to the air box. I will do as good as I can with this. 'Silk' in the AOS thread have been using the air box as vacuum source for over 45t miles and it's proven that it works( how good it works is another question).
Will it matter? I think not very much at all. It's a give and take. I'll gain in much better oil quality and cleaner combustion which I think is better than a high vacuum inside the engine.
As long as I don't have pressure it should be fine. And I will use the Porsche T-fittings to the block, which have PCV valves preventing over pressure. If I see oil on them I know vacuum is not enough.
And yes, this is a plan, let's see how it turns out. I'm confident enough to test and develop it.
What better oil quality? I work in the engineering world to and numbers matter to me as well but I just don't think there is a real problem here that needs fixing. The stock AOS is a serviceable part and if people change them at a somewhat regular interval, all will be well. It is a self contained system that doesn't require any extra steps or procedures to keep it from working.
I guess I have to show some numbers to get some 'cred' here. So, this morning I went to the work shop and looked in my box-of-many-components. I fished out a differential pressure sensor which I wired up (with some high precision electronics) to an Arduino. Then, on to calibration. Accuracy is within roughly 1/10" h2o, but from now I will use SI units.
A brand new AOS arrived in the mail (I ordered it before really understanding the issues with the normal AOS system). The test shows that it limits vacuum to 2.7kPa with some over/under shoot, see graph.
I believe it's only idiots that don't change their mind. So, with this new information I will not build the system like I said before, it's going to be plumbed in a different way.
Now the 3D printed oil filter socket is tested and working. The first one used a 1/2" driver which was not strong enough. Enjoy. https://www.thingiverse.com/thing:3968493
Good luck!
I started down this rabbit hole when my AOS failed and decided to just replace it with a genuine part, keep my oil changed at 3000K or 6 months and manometer test once a year. When the time comes to replace my AOS like many other parts, I will do just that.
Can I ask why you decided to go original. Was it a lack of a good solution or just that you feel that the original is good enough?
I remember that you adviced me to upgrade the IMSB to a 'IMS solution' in the beginning of the thread, and that's far from a genuine original part. I guess we are the same: selective of what can be, or not be, original.
Continued with flushing out the oil in the afternoon, it's been in there for almost 20000km. As a proper 996 owner I cut the filter open and search with a magnifier for pieces from the IMSB. Lucky I just found microscopical metal parts (tried to take a picture, but with macro and tripod the pieces are just one-two pixels big). They are more like dust particles than pieces. A few plastic pieces of the same size was in there also. Thinking that the car have gone 10% of it's life since last oil change this is absolutely no worries. Good news, IMS stays in place untouched.
It also means I can go ahead with another experiment of mine (O-yes, I have a lot going on). I bought an extra knock sensor which I will mount in the hole in the bottom of the crank case, close to the IMSB. Connecting it up to a spectrum analyser it will be possible to create a base-line FFT of a healthy engine. Not sure if it will show useful information or not, but I'm hoping it will be possible to detect a failing IMSB by the vibrations it makes when going bad. I worked for SKF(producing large roller bearings) when I was a student and heard about the technology. Can be read here: https://www.skf.com/binary/tcm:12-11...20Analysis.pdf
Is it obvious that we got snow here the last days?
11-14-2019, 04:06 PM
