@Balaur I'm sorry to hear you are experiencing weird issues too. I too have a lot of car experience, building motors and transmissions and even tuning ECUs. But honestly those types of projects are not why I bought a CX-5 It does make things more frustrating when problems occur. Without the specialized dealer tools it is very difficult to diagnose a car like this on your own. I also don't have much experience with high pressure direct injection motors so the troubleshooting is a bit more foreign.

In both cases, the dealer has concluded that a low fuel rail pressure condition must mean the HPFP is bad. But clearly there could be other causes. If that pump is not being fed properly, pressure will drop even with a proper HPFP. I've never been given a "satisfying" explanation for their troubleshooting. It just boils down to "the manual says the first step is to replace the HPFP."

I did submit a complaint with the NHTSA. In both cases what happened to my vehicle could have caused an accident or serious injury. I used to race cars so I'm fairly comfortable with crazy stuff happening while driving. But at least in that environment you are expecting it. When you are driving down the highway talking to your wife and the car just shuts off instantly, well that's a lot more shocking.

Last thing I'll say is that every dealer (3) I have talked to has indicated this is not a common problem. Yet, as my friend pointed out, if it's truly a rare failure then it's kind of curious that they had this fuel pump in stock.

Best of luck getting your situation resolved!

 

Mazda does not provide any specs for testing the spill valve signal in the FSM. And let me make it clear that what I wrote about the valve possibly 'sticking' is pure speculation on my part. IDK if the valve sticking is even possible, and I've never seen it mentioned in any reliable source of information that I'm aware of.

But what you can (and should) do is get multiple fuel rail pressure readings, using a live data ODB reader. In addition to getting the FRP at idle, it might be very helpful to record the values during one or two drive cycles. That pressure data is the only thing that's available to attempt to diagnose your vehicle's issue, and I recommend trying to make use of it, before you hang on a replacement HPFP.

 

My feelings exactly!
I bough this car from new to be our family car and to get my kids into holidays, adventures and nice places with icecream.
Any hobby becomes a chore if forced upon you.
At least with Rovers/MGs I can still find most parts new and very affordably.
Mazda parts, at least in Europe are difficult to find and expensive.

BTW, I'm quite dissapointed in this model.
I had to replace the windshield three times (it cracks if you look at it wrong or even without any provocations), the blind spot detection decided to not work during a few days, rear camera frozen once and required an entertainment system restart.
Plus, now that the plastics and glues are all dry, there are small noises everywhere.

Thank you for you feedback. I have monitored the fuel pressure many times and with several OBD readers or OBD dongle + PC/Android programs.
It appears that the fuel pressure is around 430 KPa, with a range between 460KPa and 360KPa, when driving or iddle. In very rare ocasions, it goes up to ~600KPa when reving down.
BTW, this 430KPa value is the setting of the pressure regulator for the electrical pump.
That looks to me that:
- fuel is sent correctly from the tank
- the fuel pressure sensor is working
- the high pressure pump is not doing anything

The spill valve is buzzing for ~20s after power down and stops when a relay clicks out. As an electronics engineer, I surely don't like that sound, but that may be normal for our cars.
Scratch this. There is something else buzzing in that general area.
BTW, unplugging the spill valve connector doesn't make any difference. There is voltage on the connector when the engine is on.

Best regards

 

Again, I would inspect the fuel tank vent hose and filler pipe, that does not mean "checking" the gas cap.
If a vehicle vent hose or filler pipe becomes clogged, pinched or perforated it will slow or stop fuel flow.
Clogged or pinched is condition that can "self-correct" at times.

 

Have you tried testing the signal with a scope with the vehicle running? As I wrote previously, there's nothing (that I'm aware of) from Mazda to be able to compare to, but IMO it would be good to at least confirm that the PCM is sending a decent-looking PWM signal, and the circuitry in the valve isn't somehow affecting it. Based on your background, it sounds like you're quite capable of doing something like that very easily.

In any case, it certainly sounds like your situation is simply a HPFP having initial minor symptoms and eventually failing completely. And that IMO is different than what happened with HiBoost's vehicle, where it stalled, couldn't be restarted, but just a short time later started and ran perfectly ok. And another piece of information is the different DTCs from those 2 situations. The P0087 code from your vehicle is the one I've seen consistently mentioned in other reports where an HPFP has failed. JMO FWIW.

 

Thank you for your help and comments.
Yes, the ECU is driving that valve adequately; there is a PWM signal with a reasonable duty factor and I see the voltage spikes caused by the valve coil. Also, the valve coil resistance is in the range.

 

Nice diagnostic work, which is something I very rarely see reported these days. So it sounds like the issue is on the mechanical side, which is not at all unexpected. Perhaps if you disassemble the factory pump, something will stand out as being the likely culprit. You're probably going to reuse the HP pipe, correct? Other DIYers have reported reusing that pipe with no issues.

 

Yes, I plan to reuse it.
I've read the steps in the workshop manual and I'm usually very careful when dealing with cars; I hope that everything will go OK. (I may still chicken out of the whole thing and let the dealers, well, deal with it)

 

Based on the fine effort you've already demonstrated on this issue, and with the FSM to guide you, I have no doubt whatsoever that you'll get the pump replaced with no problem. And if by chance you discover something about the factory pump which possibly might be related to the failure, by all means please post that information here. Although there's nothing that we can do about the cause of these pumps failing, nevertheless inquiring DIY minds would still be interested in knowing what you find.

 

VICTORY!

Many thanks to you all for your support and feedback.
m5xguy, I've deeply appreciated your kind words and encouragement. Thank you sincerely.

So, this is the story:
The eBay replacement pump arrived earlier than expected. It's a genuine Mazda part, used and from an accident car (as per the seller claims). The auction was for the pump + the rear housing.
I've separated the pump and inspected for damage and everything looked good. The pump was quite compliant, it was perfectly possible to push the piston inside and the valve was clicking clearly.

I've removed my pump from the car.
I've followed the service manual, but I didn't remove the bracket with the purge solenoid valve and catch tank.
The space was limited, but with a creative combination of 1/4 and 3/8 extenders, it was feasible.
When the pump was finally clear, I've found that it seems to be seized in the fully depressed position.

Installing the eBay pump on the car was straightforward. I've been lucky that the cam was on the lower position and didn't had to turn the engine.
I've taken extra care of the fuel rail and I hope that it will be fine.
BTW, to tighten the fuel rail nut to the correct torque, I've used an open-ended wrench with a luggage scale and calculated the required force (well, weight) to pull the spanner with, according to its length.

Finished putting everything back as needed and the car started without any issues. Pressure is fine (>15MPa) and all the fault codes cleared up.

I will try to understand what is wrong with the old pump and get back to you. The initial report is that the pump seems to be seized in the depressed position (well it could not be seized in the out position, right? as the camshaft lobe would have pushed it inside)
It looks challenging to disassemble it easily, even more so non-destructively. The cover and the purge valve seem to be epoxied (soldered?) on the pump body.
BTW, the spill valve seems to be free to move inside, it clicks nice when actuated.

Thank you and best regards,