Ford's answers to the NHTSA 6.7 Investigation

BenK wrote:

Lots of watts generated by losses..anyone now the power consumption
of this pump?

That alone begs...what is the MTBF of record for this thing?

No idea the power involved, but it sure ain't zero to get that much fuel pressurized plus cooling losses.

That is another clue... failures clustered around summer, when fuel in fuel tank is hottest.

I am willing to guess that there would be a very hot run with a near empty tank (minimum thermal mass) that factors into the failure.


MTBF design basis must be around 250,000 miles, or over 400,000km because Ford bragged that they tested that.

Note a slight lag between the "event" and the subsequent catastrophic failure --- suggesting the damage takes time to work before it bombs.

Actual MTBFs have been as low as 30,000 miles / 50,000km range on the units that experienced failure.

But on units without trouble... they don't have any problems.


Something says there is a causal agent --- it is not universally bad, but something unique to the failed samples.

NewsW wrote:

stsmark wrote:
Another item for the cavitation discussion is the fact the Volume Control Valve is on the inlet side of the CP4.

High pressure side bleed off is at the common rail.

They basically want lots of fuel to flow through the pump --- for cooling.

Lots of watts generated by losses..anyone now the power consumption
of this pump?

That alone begs...what is the MTBF of record for this thing?

stsmark wrote:
Another item for the cavitation discussion is the fact the Volume Control Valve is on the inlet side of the CP4.

High pressure side bleed off is at the common rail.

They basically want lots of fuel to flow through the pump --- for cooling.

Another item for the cavitation discussion is the fact the Volume Control Valve is on the inlet side of the CP4.

Excellent point on lubricity, viscosity and PSI at the cam/follower

Agree that using diesel as a lube is a huge mistake, especially in light of the PSI's

Exotic coatings is well known in the gear industry. From case hardening, to nitriding,
to exotic coating and the biggie...designing the lube film to act as the barrier
to metal on metal contact

In theory, our diff gear boxes are designed to have the gear faces never go metal
to metal. Why the viscosity is so high both for the film strength and the heat
generated that reduces viscosity.

They do go metal to metal often from shock loads and plain over loading where
the heat reduces the viscosity and the forces squeezes the lube out

When gear tooth faces are over stressed, it work hardens the surface, then the
surface micro fractures to either crumble (pitting) or flake off.

Why there are metal particle, however small, always found during a diff fluid
change. Why there are magnets (in most) to capture the larger particles and most
of the smaller particles

There materials explaining this is numerous and we should take a look to see
if there is anything in reference to a roller.

On that, anyone know if there is a roller that has needles are it's bearing?
Or hopefully not, a journal bearing for that roller?

Staying on lubricity, how tightly tolerance'd are the lube attributes withing
the diesel fuel specifications? Think there is a new Ah Ha there Niner...good
point you outed

Then hypothesize that there might be chemical interaction within the diesel
itself (the additives that the OEMs employ) to the additives that the owners
toss in there.

Add some heat and passivisation to this brew and what do you have?...betcha
lost lubricity to some level

Oh, then organic amines...how the heck does that play in the now witches brew?

My experience with exotic coatings says that the coatings solve frictional issues
but DOES NOT solve base material issues. Meaning PSI elastic working of
the cam surface.

That is where I think the Bosch engineers 'think' they have solved the cam
follower PSI issue on the cam...an exotic coating.

Yes solves the frictional issue, but in no way do I see even addressing the
overall issue of strength of materials. That goes back to reducing the PSI
and we have gone over that several times (form factor must be increased...or
change the architecture)

To me, Bosch is missing that talent with my guess, mistaken ID of the true
issue(s). Silly, as they most likely do have that talent, but not on this team

Then that 'thrust' (I used shock) moment is transmitted to the shaft and the
bearing, which has an aluminum housing, which most likely is the material that
bears these forces.

What kind of shaft bearing? What kind of seal? Then what kind of lubrication
(type, viscosity, etc) does that bearings specification require? To then compare
it to the diesel lubricity specification (and how tightly tolerance'd) along
with whatever additives are in that brew

Think that if there is ANY movement at the bearing, there is a cascading
effect up stream to the cam follower, therefore the piston.

On that, with this kind of 'thrust'/'shock' loading...if the sizing/type/etc
of the shaft bearing doesn't have enough margin, it will wear to become loosey
goosey enough to create havoc on the follower/piston/cylinder

That would in my guess, have the follower leave or lose contact with the cam during
the intake stroke. To then bang back on during the compression stroke. To create
an even larger PSI thrust/shock loading between the cam and follower's bearing

This scenario needs access to worn but not broken, yet, parts for examination
and measurement


{edit}....a new variation...if there should be a cavitation event
in the cylinder after that one way valve slams shut...would that then
lift the piston cam follower off the cam?

To then slam back into the cam during the compression stroke?

Anyone know if the piston/follower are positively captured together?

There is a spring in there, but is that spring sized in anticipation
of a cavitation event?

Can it be designed to manage cavitation? How? Can't get my noodle
around that at this moment...







On the higher pressure vs emissions related combustion vs going back to the old
pump. Noodling if they can make up for the lost efficiency of higher PSI with
some exhaust filtration/catalyst/etc.

Another very good one by Niner (De Laval) and had forgotten about that. Assume
the same/similar effect at a lower pressure, albeit less efficient atomization








I'm convinced that the CP3 pump is the reason cavitation is NOT an issue. It's
pressure is also conducive for cavitation, but there has not been any indication
that cavitation was an issue

With the gear pump feeding the piston pump, there is a constant, positive flow
to the piston/cylinder. The pressure drop would have been much, much less

The only potential cavitation is during the piston compression stroke where it
out ran the gear pump's GPH, as that is how it 'should' work between these
two types of pumps









On timing...there are issues with cogged belt drive. Stretch and the change
of timing during that stretch and then catch up. Ditto during pedal release

Just interesting indication that they either confirmed (must be) or 'thought'
there were injector shock waves.

Again, accumulator of some sort would have been a better solution than timing









There are many other extremely high PSI/BAR systems using fluids to cut metal,
concrete, etc. But they are constant flow vs ICE pulsations vs RPM

Wonder what types of pumps they employ?







Am now noodling whether H2O is truly a culprit. Think plain strength of materials
at play, or the lack of understanding what they really have going on in there.

{edit}.....maybe still in the fray but not as significant directly
Lubricity affected by H2O is one, oxiding of the pump innards still
there.

A way to test the cavitation hypothesis


At this point, it is time to turn to experimental science rather than paper and simulations.



Simulations used during the design up until the 2010 model year pumps were design / manufacturing process frozen have probably failed the engineers as they were suppose to test and simulate for cavitation using a variety of fuels and contaminants during the design / engineering / validation phase.


A way to test the cavitation hypothesis will be to set up a number of benches with reference fuel, with samples arrayed as follows:

A) Reference fuel / reference operating conditions

B) Reference fuel w/ contamination known to be sufficient to cause cavitation / failure

C) (B) with a pump ahead of CP 4 to raise the pressure so that it should be handily below the cavitation threshold.

D) (B) with pump operated at higher rpm / volume believed to be sufficient to cause visible cavitation damage after (x) hours.


That kind of a bench would very quickly identify whether cavitation is a rat hole, so to speak --- after sufficient hours are run on the bench, the pumps can be taken apart and the surfaces inspected to see what damage is done.



Further thoughts:


Cavitation is detectable acoustically.

Need to fit accoustic, plus temperature and other sensors to test rig.

Normally you can hear cavitation --- so that might be a good way to use ONE pump, vary the rpms etc. and listen for the telltale bubble collapse sound of cavitation --- to precisely detect the point long before any damage is done.

Side note:

Can insufficient fuel pressure from tank / feed / volume also cause cavitation?

It would certainly raise the load on the pump...

My theory for failure.

First, diesel oil, while still an oil, was never designed for lubricity properties. Not when pumps were lower pressure of 260 bar, 5000ppm sulfur pre hydro treatment to bring it to ULSD 7ppm fuel. Pump designers did take advantage of the lubricity propertied of the fuel while using the lower pressure pumps, in more lax pollution standard times. Diesel oil is designed as a fuel first, lubricating agent was an after thought or a by product of the simple distillation process back then. CARB and the EPA changed that.

My take on the Bosch CP4 failure is that the design is faulty for the pressures it is required to make. It is my belief that that big fat lump stick known as a cam, at maximum rpms (because some folks on TDI club insist on "drive it like you stole it, for proper break in, to seat the rings properly") is creating thrust forces on the piston and the aluminum wall, with that small roller, that are astronomical for a hydraulic pump with a fluid that has a viscosity of 1.5 to 4.0 CSt. These thrust forces on that aluminum bore create wear by products, metallic fines that contaminate the fuel that gets recycled back to the fuel filter and the fuel tank.

These metallic fines then find their way back to the HPFP inlet, never trapped by a 10 micron fuel filter, where they interface with the cam and roller surface. No one to date has been able to give me a load force calculation of the pressure at a tangent on that roller bearing that rides on the cam at 2000 bar liquid pressure and 5000-5500 rpm, which is what the VW is capable of in 2 liter format. No one has been able to tell me the lateral thrust force calculation of that piston on the thrust side of bore either, or suitability or hardness or wear properties required of the material in the bore to handle such load, or lubricity values. Keep in mind, we are talking fuel for lubrication, not 5w-40 motor oil lubricating the cam, the rollers and the bore. Once the fines find their way back into the pump, they start their insidious snow ball effect of damage to the roller and cam surfaces, as well as then gouging the DLC, which once those bits get into the mix and contaminate the fuel, you have very abrasive, hard diamonds basically in suspension in the fuel, until they too go through the injectors, through the motor and bore, and out the exhaust. Mix DLC, with metal fines, and it turns your fuel into slurry (ok a bit of an exaggeration, but probably true at the pressures and temperatures this equipment operates at) that destroys everything that is precision, fine and mechanical that requires micro tolerances to function within design parameters.

That is my take on the failure... either the DLC coating lets go first, or the load on the cylinder wall causes the metal to get into the fuel, but one or the other is contaminating the fuel, that gets recycled for ever, until burned up, going through the injectors and out the tail pipe.

This piston design might be fine without DLC, or DLC might be fine at lower pressures, but together, they can be a fatal combination. It explains why Bosch requires EVERYTHING to be replaced, when the HPFP fails... having micro diamonds floating around in your fuel system is a recipe for disaster, hence no short cuts when the HPFP goes south. Everything gets replaced. Prior to the CP4 designed pump, what was replaced if your CP3 failed in the fuel system? Just the pump?

The folks at Bosch need to induce failure in to these pumps in their lab tests, then work backwards to see how to redesign and prevent failures. They need fail tests, not pass tests, performed by their labs.

They should pull the fuel out of the tank of every failed HPFP Ford, VW, GM and Audi has, and then run that fuel through brand new pumps in the lab and see how fast things fail on new pumps. Maybe they should measure the fuel through spectroanalysis to look for trace elements, or look at fuel samples that almost all the fuel is distilled off, then take a peek under an electron microscope and see what solids are left from the fuel tank, that are from parts of the pump?

I believe there has to be a solution to this, and pointing fingers at the fuel is a design flaw and an assumption. Bosch has done nothing about really resolving this for 4 model years, it's been going on, the failures, since May or June of 2008, on 2009 model VW TDI's. 2012 models still have HPFP failures. This is a problem that will not go away, as much as Bosch would like it to. The pump design or implementation is not robust enough for North American applications. Bosch Engineers and PHD's made some assumptions that were defective. And we know what happens when you assume.

BenK wrote:
Oh, read this article this morning and thought an interesting side note on
diesel fuel

Think this would become more and more of a common occurance once bio is mandated
into all diesel fuel...

Used kitchen grease is thieves' latest target

It has been going on at least since the mid 2000s, with the last price spike.

The fact is, there is simply no way to control tens of thousands of home brewers, each with their own setup, and none with the semblance of a full lab, quality wise.

ASTM is just an aspiration --- that is why I believe that the B5 standard is just a sop to the politicians and EPA, because the OEMs figure out even if the biodiesel is complete garbage, 5% contamination will not be fatal.

BenK wrote:


Say go back to the older pump and it's lower BAR, which according to most say
never had or never heard of the kinds of problems with the current HPFP

Then to address the emissions, of which this current setup is designed to meet

Anyone know of other methods to clean up the exhaust besides filters, uera,
after burners, etc? Suspect the vehicle OEMs chose the cheapest, not necessarily
the 'best', right?

There is an interplay between high injection pressure (finer spray droplets) and lower particulates, lower unburnt Hydrocarbons, and with that --- smaller DPFs, SCR / DOC needed to clean it up at the tail end.

I have a Bosch paper that discuss the tradeoff, and how they are stacking the deck in favor of pressure to lower cost in the tail end.

To my knowledge... may or may not be correct, Mercedes Benz in conjunction with VAG, teamed up, at least for sedans, to create Blutec tchnology, to clean up the exhaust with urea, at least for the NOX portion of the exhaust.... Then there is the DPF portion and the separate sulfur filter portion of the exhaust system.

Someone with a Ford truck that's a 6.7 from down south, where it's still dry and warm, should crawl under their truck and get some name brands off of all those exhaust filters hanging down that's part of the exhaust system and Google them up to find out more.