Injector leakage in low temperatures! 3

[amirfarzad]

I work on full assembled fule rail validation test and want to know is it normal for a gasolin injector to have some internal leakages in low temperature (for example -30 deegrees centigrade)? nominal pressure would be 3.5 bars but I performed 5 bars.

 

[MikeHalloran]

At that temperature, any rubber seals could be expected to leak, and to be stiff enough that the extra pressure wouldn't help.





Mike Halloran
Pembroke Pines, FL, USA

 

[TStaples]

amirfarzad,

It is not normal for a properly engineered fuel system component to leak at -30°C. Most automotive fuel system components use fluorocarbons optimized for lower temperature use in gasoline and gasoline/alcohol blends. These compounds typically have a Tg around -30°C. 3.5 bar or 5 bar doesn't really matter. Both are low pressures, giving the o-ring very little deformation to fill the groove. In order to be successful, two critical areas have to be addressed; squeeze, and pre-test conditioning.

The design engineer can increase the nominal o-ring squeeze, vs. what is typically recommended in o-ring handbooks, in order to improve low temp. performance. An alternative is to tighten the allowable tolerances on the o-ring gland to raise the lowest permissible squeeze percentage limit that can occur in the gland. This extra squeeze helps to overcome the high stiffness of the elastomer once the temperature is hovering near, or below, it's Tg.

The other gotcha is that the components must be exposed to gasoline for some period of time before testing. The gasoline exposure causes the o-ring(s) to swell, improving their lower temp. performance. 12 hours at 85°C at pressure (i.e. presurize the rail overnight at elevated temp.) is usually MORE than sufficient to saturate the o-rings to simulate typical use.

I'm ignoring the basics, such as proper o-ring gland design, and ensuring that the o-rings aren't being damaged during assembly. Damage is another key issue, as slight damage often won't manifest itself at room temperature, when the elastomer is nice and pliable. It will definitely bite you in the keester when the temperature drops, however.


-Tony Staples

http://www.tscombustion.com

 

[TStaples]

Mike,

The standard automotive low temp. limit is -40°C. Nearly everything on the car must operate to some level of performance at that temperature. People who live in places like Canada, and Finland, and Maine still need to drive in cold weather.

I 100% agree with you regarding the 3.5 bar vs. 5 bar. Both pressures are too low to provide much energization of the o-rings when pressurized at such low temperatures. That's why higher percent squeezes are often employed, where possible.

-Tony Staples

http://www.tscombustion.com

 

[reidh]

TStaples,

Our design criteria is also -40 F (also -40 C) on the low end. Maybe that point was picked to eliminate confusion when dealing inter-continentally

At -40 F we call it "limp home mode", in which everything will operate, but you won't be setting any speed or fuel economy records during your trip.

Reidh

 

[TStaples]

reidh,

The good news is that for injectors, at least, is that they are equipped with internal heaters! The power dissipated into the coils does a nice job of warming them up slightly when the vehicle is first started at -40°C. Engine heat will warm them up further after a few minutes of operation, improving the leak integrity if they are borderline.

I find the limp home mode terminology somewhat misleading, in that you generally only need to deal with -40°C temperatures during a cold start, after an overnight soak (at least for engine, and engine compartment components, with which I have the most experience.) It does however, sound better than limp start mode!



-Tony Staples

http://www.tscombustion.com

 

[reidh]

Tony,

You are correct. I should have better stated my thoughts. We use the term "limp home mode" for operation at our system's limits, whether it be heat, cold, wind, low oil, etc.

Furthermore, limp home is a minimum target for each subsystem. Some subsystems will obviously be able to perform well outside the operating limits, but the subsystem which is the weakest link at each operating limit (high temp, low temp, etc.) has to be able to perform at some percentage of full capacity in order to be able to "limp home".



Reidh

 

[amirfarzad]

Thanks alot for you all attending this discussion.
I would like also add, if there is any leakage in the first seconds of operating 3.5 bars, through warm up duration, is there a risk of firing? Never the less I will try to check actuated injectors also to see what happen!

 

[TStaples]

amirfarzad,

I'm assuming you mean the 'f' word that rhymes with hire. Since I work in automotive fuel systems, that word is not in my vocabulary, at least according to our corporate attorneys.

Anytime there is a fuel leak, you do run the risk of a thermal event, which could result in a crispy customer. You do not want to even think about allowing leakage, regardless of how low the ambient temperature is.



-Tony Staples

http://www.tscombustion.com

 

[dgallup]

Add a little methanol to your fuel and those o-rings will swell up and seal.

Tony is correct about the need to precondition the o-rings prior to low temp testing. No one but a test engineer is going to take brand new parts and expose them to -40C.

Tony - don't you have any work to do at that new job? I could send you some.

 

[TStaples]

Dave,

I have plenty, but I can always find a little time to help.

The methanol is another good trick, but you can't always count on it being present in the fuel. Here in Cali, the standard is 6% moonshine minimum in the fuel, which would do a great job of sealing the o-rings at low temp. However, it has yet to get below freezing in the 4 years I've been here.





-Tony Staples

http://www.tscombustion.com

 

[amirfarzad]

Tony,

Thanks for your help. I already prepared a mixture of %85 petrol and 15% methanol. would it be enough to leave the O-Rings for some 10 hours in that mixture before repeating the tests.

Regards

 

[TStaples]

amifarzad,

If you are referring to the external o-rings, then I see no issue with exposing them seperately. You may not have a method to safely expose them at elevated temperature. If that's the case, then I would measure the I.D. and cross section of the o-rings, start the exposure test at room temp, and remeasure the i.d. and cross section at 12 hour intervals. Just drop the o-rings into a capped beaker with a small valume of the gasohol mix. Once you have seen an approximate 10-15% swell by volume then, the exposure time is long enough. Based on my experience, you will likely need 24 hours at room temp.

However, the internal injector o-rings also need to see exposure. I recommend exposing them by charging up a capped fuel rail with gasoline, and exposing the assembly to elevated temp. It helps facilitate exposure to the installed o-rings, which are already under compression. Just be sure you have a safe place to do this (a suitable explosion proof facility, or installation on a vehicle works even better. A couple of engine heat soaks and you are all set.)

-Tony Staples

http://www.tscombustion.com

 

[turbocohen]

Older injectors used intrernal o-rings but modern injectors are hermetically welded to minimize evaporative losses. The upper oring to the fuel rail cup should be all that you have to contend with. Methanol is a bit tricky since it can blow the oxygen molecule out of the o-ring material after it is exposed then allowed to dry sometimes leading to shrinkage. USE a meth tolerant o ring! Rewetting with fuel swells it but after assembling a few oem alky injector rails from Kapuskasing.

http://www.gmcanada.com/inm/gmcanada/english/about/FeaturedStories/Story037.html

 

[TStaples]

turbocohen,

You are correct, for high volume mainstream liquid injectors. I have no idea how modern amirfarzad's injectors are. There are still a lot of older dsigns being produced with o-rings, and still in use in many global markets outside the US. Good point, though!

I had the priviledge of working with fellow eng-tips contributor dgallup to design one of the first (if not the first; you can never be sure because every injector OEM's development timelines are proprietary) automotive gaoline injectors with no internal o-rings, back in 1993-98. You would think that would have crossed my mind, but alas . . .. However, I have worked with many more injectors that still have o-rings, primarily in alternative fuels, and non-US markets.

Do you work, or spend time at Kap? I have been there several times, most recently in 2001-2002 supporting a GM CNG bi-fuel truck program. Darn COOL job! Brrrrrrrrrrr.

-Tony Staples

http://www.tscombustion.com

 

[dgallup]

Actually, quite a few current OEM injectors still have internal o-rings. The Delphi Multec 2, which is still in production has two. That's better than some of their earlier designs that had 4. I'm about to cut up a Multec 3.5, I'll let you know what's inside. Always fun to see what the other guys are doing. BTW, the dynamic range on the Multec 2 is awful.