CR engine HPFP analysis

[JSWTDIPilot]

What vapor pressure are you talking about in the pump??? Fuel flows into the low pressure side of the hpfp at around 73 psi supplied by the boost pump. The parts that are failing are in the low pressure side of the hpfp,not in the high pressure side.

And as far as the boost pump not suppling enough fuel. While I have not checked the fuel output of a boost pump yet. I can tell you it pumps A LOT of fuel very FAST!

dweise

I'm talking about the pressure inside the HPFP case that lubricates the roller and cam. All it would take is a millisecond or two of disruption from the low pressure fuel pump to starve that area of fuel. (A simple computer fault turning off the low press fuel pump.) A few turns of the cam, a second of vacuum inside the case from the HP pump continuously pumping fuel out and the diesel would vaporize and viola! you have micro bubbles and cavitation on the roller/cam assembly.

Don't know how plausible that scenerio could be, but we are exploring all possiblities. Until anything can be proven right/wrong, IMHO, this could be another reason.

 

[JSWTDIPilot]

ALL of the return fuel does not flow through this filter. 95% of the fuel that flows into the hpfp flows into and right back out of the low pressure side for lubrication. The othe 5% of the fuel enters the high pressure side.

dweisel

Look at the cutaway again in the VW study program booklet. There are "fine filters" located on both the HP outlet before the HP plunger and one fine filter located prior to the fuel leaving the low pressure fuel return from the HPFP. All the fuel goes through a "fine filter" when leaving the HPFP no matter low side or high side. It wouldn't take much to plug one of those filters after the pump started coming apart or even prior.

(btw... this theory has nothing to do with my cavitation theory)

 

[JSWTDIPilot]

With all due respect............I think if these Bosch hpfps are so sensitive that a millisecond or two of the boost pump not pumping would cause or precipitate a hpfp failure is beyond my comprehention. I think we are ALL in trouble if this were even remotely possible. Sounds like complex physics to me, or maybe a little sci-fi.

dweisel

I suppose if the HPFP weren't so sensative to something wrong happening, we wouldn't be having any HPFP discussions like this?

Sci-fi, only if we're talking about aliens growing in the pump I suppose. Those gremelins can chew through metal with acid spit in milliseconds I suppose, with all do respect.

Complex physics, yes... but then again, I only offered my opinion to a possible reason for failure. I didn't go into any physics calculations. If you can prove this theory dead, please do with calculations or evidence.

 

[silversx80]

Could it actually be that... it... is... bad... fuel? Not the most recent tank, but say 15 tanks ago, to where a sample of the most recent tank shows to be... inconclusive (since damage wouldn't happen immediately, or from the most recent tank, but a chronic issue that progressively gets worse until part failure)? Could it be that it isn't the quality of the fuel delivered to the fueling stations, but rather the quality of the storage at said station?

Could it actually be that VW is selling a quality produce that has been fouled up by a low-volume diesel station with tanks in disrepair?

I've read enough. I'll be unsubscribing from theses threads soon enough.

 

[silversx80]

Seems to me that if the fuel station tanks were in such bad shape other tdi's and diesel vehicles would be failing also. How come there is just a rash of 2009/2010's that are failing? When our 09 Sedan failed we also fueled our other 09 tdi and an 03 tdi at the same station. BAD FUEL would have affected all three engines and not just one. Yes or no?

Same station every time, with every vehicle?

 

[silversx80]

98% of the time they are all fueled at the same local station. I guess it must have been the other 2% that was contaminated. Just my luck! My wife and I have been driving diesel cars for 23 years and NEVER had a fuel problem or injector pump failure. We buy a brand new 09 Sedan and WHAM. Bad fuel! I guess after 23 years I was bound to eventually get some contaminated fuel.
dweisel

Well, certainly a possibility. Since there was a 2% chance, there was a chance. Take the the well known variability in manufacturing of both fuel (possible lower lubricity) and fuel pump (more sensitivity, possibly), plus the already known Bosch outlook on 460 micron vs. 520 micron wear scar tests, and the 2% chance that the fuel could have been from a contaminated station... well, that is crappy luck and really sucks, but c'est la vie.

Since I believe it is most likely the fuel that is causing the issue (we don't hear much about these failures in Europe, which we know has a better and more consistent fuel) you'd better believe I'm using additives after reading about these issues. Once I go through the coop stuff here, I'll also be getting some B100 to supplement each tank.

 

[JSWTDIPilot]

I'm familar with that document and have read and looked at it many times. If you had looked at it carefully and knew how the fuel flowed through a hpfp then you would have known that the screens/filters you are talking about are before the fuel metering valve and before the overflow valve and NOT in the main return fuel port. Yes,fuel does return through the overlow valve when the fuel pressure is over 62 psi. Look again at the drawing and you might understand it this time. Since I have a hpfp to actually take apart and examine, its way better than your paper verison that you are basing your knowledge on. Maybe you need to revise your theories.

dweisel

Before/IN, same difference in this case! Let's not split hairs, please.

Also the overflow valve that you are refering to, it also serves as a pressure metering valve that meters the fuel OUT the low pressure fuel return.

"The fuel delivered by the Auxiliary Fuel Pump V393 acts in opposition to the piston and the piston spring of the overflow valve. With a fuel pressure over 62 psi (4.3 bar), the overflow valve opens and clears the way to the fuel return. The excess fuel flows through the fuel return into the fuel tank."

Pg. 38. of the training book serves my point. The fuel will flow through the fine filter prior to leaving the HPFP unit through the low pressure side above 62psi which is where the pump internals sit the marjority of the time. Yes, some fuel does leave through a metering orifice prior to 62psi unfiltered but its only a small amount while the pressure inside builds as the low pressure fuel pump starts up. The low pressure fuel pump is always supplying 73psi.

My MAIN point that I'm stating, is that the majority of fuel goes through a pair of fine filters before leaving the HPFP. Whether its through the return or prior to the HP metering valve. IF those filters get clogged, WHAT WOULD HAPPEN? Question to a theory, thats all. It seems to me those filters would clog quickly in a bad fuel event?

In the event of a HPFP failure, these filters clogging would restrict all flow to the CR and back to the fuel tank. We do know that metal chips get back through the entire system, so do these fine filters break away because of the clogging and then flow through the entire fuel system?

 

[VLS_GUY]

Dweisel,

Any plans to do piston spring compression force testing through out the displacement range of the spring?
How about doing a metallurgical exam of the metal particles to see if they were created at an elevated temperature? I also would like to know the metals used on the roller and cam to determine if they were appropriate for the application.
Any tribologists on TDI Club willing to do a failure analysis such as those connected to a university?

 

[VLS_GUY]

I am going to ask the mechanical types if we can do spring compression testing at work; perhaps using the NAVFAC lab. I going to try and get them interested in the problem as a learning experience for their personnel. The big problem is that this can take time as the Navy's problems always take priority.
An alternative would be to have some one in a research university with personal access to a lab arrange something for more timely results. I can think of some one with those exact qualifications that moderates on this board...

 

[DanG144]

I have a failed fuel system (HPFP, injectors, rails, pressure regulator, aux boost pump) that I would be willing to ship to be evaluated.
I would like them back for show and tell at GTGs.
I would pay shipping one way, if someone would pay to return them.

This stuff is from a failure already documented on this web site, not a new failure.

Dan

 

[VLS_GUY]

Ideally I would like to see a comprehensive analysis of all the components done; not just on the springs. I still intend to talk to our mechanical engineers about this as they are car nuts and will interest them. I am worried merely testing the springs will not tell us much with out looking at the entire fuel system.
An another idea is what about using this as a project for training diesel/automotive technology students in troubleshooting? Nothing makes a problem more interesting for a professor and students than a potentially new problem.
What universities or colleges have rigorous programs where they could investigate this problem as a way to teach students? What about SIU at Carbondale? Any more ideas in this line of thinking?

 

[Elfnmagik]

This is good stuff.

VWoA, I hope you watching!

 

[aja8888]

This is good stuff.

VWoA, I hope you watching!

It would be interesting if an independent team of engineers did a forensic research paper on the HPFP failure causes and published it. I suppose Bosch and VW would have to approve it ahead of publishing.

 

[VLS_GUY]

Having VW's cooperation would be ideal since you would have access to the technical data package (TDP), the designers, and their manufacturing procedures. This does not mean that an investigation can not be done however. After all VW sold the fuel systems along with the cars. The owners of the failed parts can do what they want with them. Lack of cooperation from VW and their supplier (Bosch) makes the process much more difficult and involved; not impossible. Publication of the results would require plenty of editing from multiple parties involved to avoid legal prpblems; this is the major area of impact.
If VW wants to start a turn around on their North American reputation (and sales volume) for dealing with customer problems now is the time to show leadership and conduct an open investigation into this issue.

 

[tcp_ip_dude]

Thanks much "Dwiesel". Nice to be seeing some 'real' data.

 

[tcp_ip_dude]

I'm not sure its real data. Just a bunch of broken parts.

dweisel

It's at least data coming from something that can be touched, measured, photographed and not data coming from the imagination.

Then perhaps it's one step closer to 'real data'.

Thanks again for your efforts.

MLA

 

[Ski in NC]

I do failure analyses on marine propulsion systems as part of my business. Often an analysis has been sub'd out to a metallurgy lab with an electron microscope. Rarely does that reveal anything worthwhile. Failed bearings are so chewed up that any hints of a cause are lost. Dimensions can't be verified, hardness has changed, surface finish lost, etc. Sometimes the report ends up stating "for some reason the bearings ate themselves", although I have to use somewhat different wording if I want to get paid.

Picking through a failed HPFP will likewise probably not reveal a cause.

But if anyone has a failed pump they want to send me, I'd love to go through it!!

 

[VLS_GUY]

Ski,

Do you think anything can be revealed by the piston spring compression test? The metallurgical testing with a SEM on the metal particles will tell you very little. The idea is to see if the anti-friction coating has failed under loading from the piston spring. The filter particles in past testing I have seen have shown coating treatments breaking off etc. and forming particles captured by screens (delamination) etc. Have you seen this as well?
Can you do spring compression testing and SEM particle analysis as part of your business?

 

[50R BTR]

I would think having a HPFP prior to failure, with some hours and a subjective clean fuel history, would tell you more than would a failed unit. With the cost of a new HPFP around $1k, it would take some very dedicated resources. I agree that post mortems rarely show exclusive failure modes, especially with bearing or surface loading failures. Does anyone know if there are rebuild parts available? Just curious.

 

[VLS_GUY]

The ultimate situation is to do inspections on a few HPFP's in a time series. Unfortunately we seldom get this sort of data in failure investigations.
As far as rebuild parts are concerned, as your VW dealer or better yet talk to DFIS.

 

[Ski in NC]

Ski,

Do you think anything can be revealed by the piston spring compression test? The metallurgical testing with a SEM on the metal particles will tell you very little. The idea is to see if the anti-friction coating has failed under loading from the piston spring. The filter particles in past testing I have seen have shown coating treatments breaking off etc. and forming particles captured by screens (delamination) etc. Have you seen this as well?
Can you do spring compression testing and SEM particle analysis as part of your business?

Any lab worked needed is subcontracted. My shop does not have lab equipment at a quality level needed for forensics.

I'm not familiar with an anti-friction coating on a spring. In general, springs don't have friction surfaces and are not coated for such. No expert on this pump, though. If there is a coating on the spring, and it is flaking off, that should be readily determined visually.

Edit: I must not have read your post very closely. Realized you were discussing the rolling surfaces, not the spring itself.

 

[BikeFlyer]

Looks like the cam shaft that the HPFP roller runs on is a changeable part. True? Or does it require an engine tear-down?

 

[VLS_GUY]

I am interested in finding out about the failure of any surface coatings on the cam and roller not the spring. If it is delaminating under compressive failure with out heat the particles may appear different from those that would happen under heat stress. The heat may come from friction once the coating is gone.
Does anyone know if DFIS has the capability to rebuild the HPFP yet? Their rebuild test procedures would be the best guide as to what has failed in the pump addition to the obviously worn items in the pictures.

 

[VLS_GUY]

I would not think the HPFP is worth rebuilding. I would be using the rebuild procedures to arrive at procedures to conduct the failure investigation.
I agree that Lubricity is evident on the cam/ roller interface. We also must determine if is lubricity problem for the fuel or the metal being used for either the cam or roller is the problem. Poor material design choice or bad process quality control during manufacturing could be the problem. In other words besides looking at diesel fuel quality the metallurgy of the parts need to be looked at as well.

 

[VLS_GUY]

Updates on a couple of things:

I gave some background on this problem to one of our new Mechanical Engineering PhD.s
He is still looking at the design but a couple of things he has noticed after an hour:
1. The small diameter on the roller follower increases friction loading on the follower/cam surfaces. In sort the roller diameter should be increased.
2. The cam and roller centers of rotation are out of alignment through out much of the cams rotation. This Introduces slip as a likely cause of the galling evident on the roller The roller needs to be pinned as noted by Brian a to hold it in alignment.

I contacted the metallurgical lab on base today and the section head was interested in this novel problem with new technology. He is going to give the lab manager a heads up on this to see if lab tests can be arranged on the springs and other components as soon as the lab manager get back from travel. Our PhD also feels he knows how to run their lab equipment if needed.
Not a bad start for an Electrical Engineer...

 

[oilhammer]

On the samples I have had here, it really looks like the roller/piston somehow cocked sideways in its bore. The cam has a severe wear spot right in the center.

 

[VLS_GUY]

Alan thinks the slip is responsible for the galling and the heat build up evident on the parts as stated before.
On a side note Alan wants to buy an air cooled bug as a daily driver. It seems back in Hawaii he owned a few air cooled dubs and wants another.
Any one have a bug in good shape they would sell for a reasonable cost in Southern California?

 

[dweisel]

On the samples I have had here, it really looks like the roller/piston somehow cocked sideways in its bore. The cam has a severe wear spot right in the center.

oilhammer, does the roller show any wear on its diameter or any flat spots? My thoughts are the roller would not get out of alignment with the cam unless it stops turning because of metal particles getting trapped between the roller and the piece that holds the roller in place on the piston cup. In other words the roller getting turned is the result of metal already coming off the roller/cam and not the cause of the failure.

dweisel

 

[dweisel]

Updates on a couple of things:

I gave some background on this problem to one of our new Mechanical Engineering PhD.s
He is still looking at the design but a couple of things he has noticed after an hour:
1. The small diameter on the roller follower increases friction loading on the follower/cam surfaces. In sort the roller diameter should be increased.
2. The cam and roller centers of rotation are out of alignment through out much of the cams rotation. This Introduces slip as a likely cause of the galling evident on the roller The roller needs to be pinned as noted by Brian a to hold it in alignment.

I contacted the metallurgical lab on base today and the section head was interested in this novel problem with new technology. He is going to give the lab manager a heads up on this to see if lab tests can be arranged on the springs and other components as soon as the lab manager get back from travel. Our PhD also feels he knows how to run their lab equipment if needed.
Not a bad start for an Electrical Engineer...

No engineering degree here,but wouldn't increasing the size of the roller also increase the surface area that would be subject to more fricton at the roller/cam contact surface and where the roller sits in the piston cup?

Increasing the roller diameter would also increase the lenght of the stroke of the high pressure piston and also cause the spring to be compressed more.

Can you describe how the rollers center is out of alignment with the cams rotation?

dweisel

 

[oilhammer]

oilhammer, does the roller show any wear on its diameter or any flat spots? My thoughts are the roller would not get out of alignment with the cam unless it stops turning because of metal particles getting trapped between the roller and the piece that holds the roller in place on the piston cup. In other words the roller getting turned is the result of metal already coming off the roller/cam and not the cause of the failure.

dweisel

Very good point. The roller has bits of shed metal, looks like a chrome-finish tool that has had some chrome plating chipped away. And it does NOT roll smoothly.