This refers to the 03L130277B injectors fitted to the above CAY code engines. I decided a few weeks ago to disassemble a faulty injector to try to get to the root cause of why it failed, hoping that there might be some indication of something we can do to prevent it! This is a long read, but I've put in plenty of pictures
VAG's own SSPs 351 and 442 detail how these "piezo" injectors work and are worth a read. I won't repeat it all here, for brevity. The actuator element in these injectors comprise of a stack of piezoelectric material sheets, with alternating positive and negative electrode plates between them. The positive sheets are connected by a common terminal, as are the negative ones. Applying a voltage of ~150 Volts causes a minute expansion of the stack, which actuates the injector.
As the piezoelectric material is a good insulator, one might expect a very high resistance across the injector terminals (or probably a detected open circuit). On failure, my injector at least showed only a few Ohms.
This is the suspect in question...
Slicing the plastic top of the injector revealed a resistor in parallel with the two electrodes going to the piezo stack:
According to the colour bands, this resistor should (I think) have 200 k Ohms across it. Nope! Just 2 or 3. As it's in parallel with the piezo actuator, either one could be responsible.
Next, off came the piezo actuator assembly from the injector body:
That circlip was a nightmare to get off... Next, I softened the plastic piezo actuator cap with heat and gingerly removed it:
I say gently... the copper electrodes in the plastic head separated from the terminals in the above very easily. Measuring the resistance across the terminals again showed only a few Ohms. The next part got a little rough - removing the top of the capsule:
And, several hacksaw blades later, here we have it! Not what I expected at all. There was no sign of any contamination in the "capsule" whatsoever, no soot, fuel, moisture, nothing. Clean and dry. No glass fibre insulation, just the piezo stack elegantly encapsulated in silicone rubber. Again, the resistance between the terminals was low.
The only other visible oddity was this black dot towards the top of the stack (I had to peel some of the silicone away to make it more visible). It appeared to have originated from the stack and worked its way into the silicone. Again, was it like this from the factory, or is this where the short point is?
It turns out that this was actually where the fault was! I wanted to dissolve away the silicone somehow to avoid damaging the stack, but this was impossible so I had to peel it away. This resulted in the following:
I'd thought that the wires were electrode plates intersecting the stack, but I was wrong! Next I removed more silicone and the plastic casing to yield the following:
I cut away the electrodes (the stack was still showing as a closed circuit with very low resistance at this point) and then applied slight force to the stack, which dutifully split in half:
The section on the right seemed to be an open circuit, but the one on the left was the culprit! In fact, the black dot is visible just over halfway along the stack, where the wires go in. I split the stack apart there to yield the following:
Bingo. The "good" section of the stack did not have any of this nasty discolouration. The photos are terrible (sorry), but on close inspection, this dark mark actually resembles a fractal pattern, almost like a lightning fork. In total, the stack consisted of 20 of these individual wafers. Around 4 of these had the black discolouration, and were all near the top of the stack and only at the one electrode. The discolouration originates from the barely visible dark "arc" in each corner of the wafers, presumably where the electrode backbone connects to the electrode plates.
The two on the right were an open circuit as tested with my multimeter, but the darkened one on the left was shorted out (even before I deliberately cracked it!).
This is where it gets more interesting. I fractured one of the good wafers to have a look inside. It seemed darker on the inside than outside, so I put it under a microscope (only 20 x magnification, unfortunately). This is the view from above; notice the dark band in the corner. All of the wafers, good and bad, exhibited this. This is also where the darkening appears to originate on the dodgy ones.
This is a side-view showing what's inside the wafer. I was expecting it to be a solid sheet of piezoelectric crystal...
If you look VERY closely, you can see a tan-coloured uppermost and bottom layer. Between these are around 15 layers, with very thin and shiny metallic layers separating them (these are the electrode plates, I think). That means, with 20 wafers per stack and 15 layers per wafer, we actually have 300 individual layers of piezoelectric material in each injector! Now for the very fun part - dissection of one of the darkened regions:
To me, it looks dark, bubbly and rather igneous throughout. With each layer having opposing voltage electrode plates above/below, any damage to a layer could result in the plates forming an electrical connection, shorting out the stack. This is a magnified view of a darkened patch as viewed from above:
The dark material has broken through the surface of the ceramic upper insulating layer and some topology is clearly visible too.
This begs a few questions at this point: what is the dark material, and what caused it to form? Did it all appear all at once, or did it spread over time? Is it the result of mechanical damage to the stack, or is it the result of electrically inflicted damage? On closer inspection, the dark mark (sorry Harry Potter fans...) appeared to be the same material as the electrode backbone, i.e. a kind of solder.
This is what I think might have happened: The damage occurred near to the top of the stack, where the electrical current and voltage should be at their strongest. It appears as though the material has become VERY hot and that the metal electrode material has melted and permeated the stack.
A few more observations: here in the UK, VW allegedly gave a two year extended warranty on the injectors and EGR valves after applying the emissions fix. VW admit on their dedicated webpage that EGR levels are increased as part of the fix and the injection cycle is modified. More EGR means more soot, so more frequent DPF regenerations. As far as the injection cycle is concerned, this appears to mean cancelling the pre-injection on high engine loads in fourth and fifth gear (which really does make an appalling racket).
Curiously, Renault/Nissan use similar injectors, also from Siemens, on some of their Diesel engines, seemingly with low failure rates. Even more curiously, independent tests have shown high NOx from these engines and infrequent DPF regenerations (I'm happy to substantiate this if asked...).
Could it be that extended regeneration attempts, with the required rapid injector cycling, may cause the stack to overheat, thus leading to electrode melting and stack cracking? If so, it might be possible to reduce the likelihood of failure by keeping the amount of time regenerating to a minimum. Easy enough out of town, but this might not be an option for all.
That's it for now, thanks for reading
