Coolant leak: EGR valve or cylinder head?

Hello, my old TDI is experiencing significant coolant loss (2.0 TDI, BRD, 16V engine, 2007, 305,000 km). It's clear what's causing it: it's "eating" the coolant and sometimes emitting dense plumes of steam. When it starts, it runs very poorly, but after a few kilometers, it improves somewhat, although the problem persists. This has only been happening for a few days, but it's now practically undrivable. You have to refill it after 10 kilometers.
Now the question is whether the problem lies in the AGR cooler (a steel pipe cooler, not aluminum), the cylinder head gasket, or the well-known tendency for cracks to develop in the 16V cylinder heads?
I've been researching for hours, but I can't find anything comparable that actually helps. It's clear that water is reaching the combustion chamber (bad leak), which is why I initially ruled out the AGR cooler. However, I now read in an article that enough water could still reach the intake manifold through the AGR cooler to cause such plumes of steam. A CO test probably won't be helpful either, because all three of the mentioned components (AGR cooler, etc.) could be the cause. Points can be released into the water as exhaust.
Therefore, I hope that someone with similar experience can offer advice on what is typical for this engine, considering the symptoms (poor starting, rapid progression of the problem, etc.).
If I could definitively identify the AGRK, I would save myself the trouble of disassembling the head.
Does anyone have instructions or descriptions for disassembling the AGR valve (or Z-head) on a BRD engine? It seems like everything is quite specific and, at the very least, very difficult to access. I haven't found anything useful so far either.

Now, one could look into a crystal ball, or through an endoscope... if one had one .

I would first pressure test the cooling system while it's at operating temperature.

If the pressure drops rapidly, disconnect the AGR cooler and bypass the lines.

Repeat the process. If the pressure remains constant... error found. You can also test the AGR cooler separately.

If not, then Der ZK is the prime suspect.

I'm less inclined to believe it's a ZKD (cylinder head gasket) issue, because usually it's only slightly leaky, allowing combustion gases to be pushed into the cooling system, which builds up enough pressure to force the coolant out through the expansion tank, but with very little coolant actually entering the combustion chambers.

I wouldn't use the car anymore, the keyword here is "engine failure."

Thank you in advance for the suggestions.
"Bleeding the system probably wouldn't have helped here, because the next morning, when the expansion tank was opened, a clear equalization of pressure could be heard. It seems that a vacuum had built up during the engine's cooling process, once the suspected hairline crack had closed." I can't explain it any other way.
There's actually a quite simple way to insert a corresponding disc without a hole between the flange of the EGR pipe and the valve, instead of using a gasket. I've now done the same thing after reassembly (see picture). If it were the AGR cooler, the problem would likely be gone by now, but... (see below).
Unfortunately, before realizing this, I had already disassembled the intake manifold to take a look inside (be careful, there's a nasty trap: the last screw is hidden inside, underneath the oil filter, see picture).
Unfortunately, it doesn't seem to be trying to start at all. Not even a cough. The only thing that could prevent it from starting on the suction side would be the throttle valve, and that is definitely open. I'm also not sure what to do right now.
I couldn't clearly see any water in the intake manifold, but the AGR pipe was slightly damp. It remains unclear whether this leak is directly from the AGR cooler or if it's coming from the exhaust manifold. And strangely, I also saw red droplets in the turbocharger pipe, right near the throttle body (see picture). I don't understand how they got there. Perhaps they were pushed in there when the engine was turned off? "Surely, the turbocharger shouldn't allow exhaust gases to flow back into the intake manifold?"
Note: It was probably urgently necessary to disassemble the intake manifold after 18 years. All the inlets were heavily clogged. It's a miracle that it was still running properly... So far, I have only had to verschandeln the EGR valve twice.
Okay, now I need to figure out how to get it working again. Unfortunately, my new laptop isn't compatible with VCDS or its drivers yet. Reading comprehension is out of the question for now.

Help! Unfortunately, it still won't start, but it cranks normally with a full battery. After finally getting VCDS running on the new laptop, at least with an older version: no errors in the engine control unit (ECU), but the readiness test for the AGR (selective catalytic reduction) failed (which I don't think is relevant for starting). I wanted to check the phase angle between the crankshaft (KW) and camshaft (NW), at least during the starting process, but I can't find any measuring blocks (OBD-II parameters) for that (for engine BRD); it exists for other engine codes. The start synchronization is at 3 when stationary and 255 when cranking (although there's no description for 255 in the context menu). Is that okay? Plausible engine speeds of 220 and 110 RPM are displayed for the crankshaft and camshaft, respectively. What measurements or readings can I check during the startup process for further clues?

Here's the history again, maybe someone has a good idea.
- Significant coolant in the exhaust, possibly due to the EGR cooler or a blown head gasket.
- Intake manifold and both flaps removed and cleaned (the engine was not running while this work was being done).
- Since then, it hasn't started.
After that, I checked the following:
- Both connectors for the AGR valve and throttle body were securely attached (otherwise, there would be an error message, which I've experienced before).
- Nevertheless, I removed and opened both flaps again. I had recently cleaned the AGR valve (from the inside) and the internal connector is fine. The other one was completely full (!) with oil, more than I've ever seen, and I thought that problem only occurred with the AGR valve. Well, both of them can be ruled out as the cause. Moreover, the AGR valve was closed and the throttle valve was open when starting, which is how it should be mechanically. Even if the actuator is faulty, the engine should still start if the flap positions are correct. A diesel engine should actually be able to start, even without a manifold and valves, although it might not run well.

- The cable to the camshaft sensor, which runs underneath, I might have accidentally brushed against at some point, but definitely not squeezed or anything like that. It looks completely normal, and when I had a faulty sensor a few years ago, it also gave a clear error message.

To me, it initially looks like the injector might not be working, or the valve timing is incorrect, which is why it's not firing at all, not even a spark. I have another possible explanation: when cleaning out the channels, some small pieces probably remained behind that the vacuum cleaner couldn't remove; it's just really sticky. This could temporarily prevent individual valves from closing, and the compression needed for ignition is missing. However, I find it hard to believe that at least one valve is affected in every cylinder, and that this flimsy paper-like material is still blocking them even after a lot of cranking. But who knows? Has anyone had any experience with this?

I really can't imagine that the belt just slipped off by chance.
My next step would be to take some measurements (see above), but I'm not sure which other measurements might be helpful.

Synchronization angle:
Okay, I understand. Please provide the German text you want me to translate into English. I will only provide the translation.

The toothed belt principle is the same as that used in:
Okay, I'm ready. Please provide the German text you want me to translate.

Thanks in advance to Rainer for the links. These descriptions are really very good and detailed.
Unfortunately, as I mentioned before, I don't have a "synchronization angle" setting. However, in MWB4, there is a "rotation angle" (see image). Because of the heading "Control of Pump-Nozzle...", I didn't pay much attention to it. The value might indicate that (?). It remains stable at 1.87 during the initial test, unlike the other measurements which show fluctuations.
But with a height of 1.87 meters, shouldn't he at least be able to start the engine?
The next step would have been a mechanical inspection of the valve timing. However, due to the inaccessibility of the engine, this would have been more of a hassle, so I preferred the software-based solution. One more reason to have VCDS .
With a value of 1.87, it seems that issue has resolved itself.
I can't imagine that the "water surge" could have any effect on the injection system or the valves. And electrical problems with the fuel injection system should normally trigger an error message...
The engine has never been extensively disassembled or modified, except for peripheral components (valves, tandem pump, oil cooler; all of which were done a long time ago). Therefore, everything is 100% original and untouched. And besides that, I haven't even touched on the core of the problem.
Any other ideas?

Sure, maybe try activating the starter and either repeating the last Screenshot or simply logging the data?

Nothing meaningful can be displayed while the system is stationary.

Sorry, yes, I understand. I was just fixated on one specific angle value, which is fixed anyway (is it within spec?). I've attached a picture with the starter running. I don't see anything unusual. The question is, what other measurements would help me further diagnose the issue? I've already checked everything that I think might be related. It doesn't necessarily mean anything.

Hi.

According to the Screenshot, the engine is drawing in air, the NW-KW (North-West-Key Word) is correct, and the PD (Pressure Differential) elements are being controlled appropriately.

Were the PD elements removed, and are they now somewhat properly adjusted?

Sure, here's the translation:

"LG" translates to "Best regards" or "Sincerely" in English.

Hi!

Very strange. Does the tandem pump promote [something]?

Best regards,
Brezelmann01

Good morning. Besides the intake manifold and the two flaps attached to it, nothing else was touched, absolutely nothing. If anything else had been changed, I would have mentioned it. I've even wondered if I might have made a mistake, but I don't currently see any way that could explain these effects. I really don't see any indication of what might be causing this. Therefore, I had asked which measurement values might still provide some useful information.
I haven't checked the tandem pump yet. I might try it later if I have time today. However, it's only 2 to 3 years old and it's a branded product.

Hi,
What fuel injection amounts are displayed during the start-up attempt?
hg
Herbert.

Herbert wrote:

Hi, What fuel injection amounts are displayed during the start-up attempt? hg Herbert

Hi Herbert,

In his Screenshot, a target value for injection duration of 9.x °KW is displayed, which suggests that the injection quantity would be adjusted accordingly (although this is not visible in the Screenshot, as it's from a different control module).

My guess is that no diesel fuel is actually reaching the PD element. Perhaps the hoses on the diesel filter are reversed, causing the system to attempt to draw fuel through the return line.

You can measure the pressure of the tandem pump, but in my opinion, it would be sufficient to simply check (with a cloth in front) whether any diesel is actually coming out.

Best regards, Rainer.

Hi,
I would like to clarify, based on direct measurements, whether or not any injection should be performed at all. The target value for the support duration might indicate that, but I'm not sure about it.
Otherwise, it looks suspicious as if the DK (Domain Key) is missing.
hg
Herbert.

Okay, so I have a day off today and I disconnected the hose on the tandem pump that goes to the return line (the lower one). As expected, it sprays out of there when I start it. There's definitely nothing mixed up; no one has touched it in years.
What does Herbert mean by "missing DK"? Surely not "Drosselklappe" (throttle valve)? "It's definitely open, and even then, it should still run in case of a malfunction."
I've selected a few more engine control module (ECM) values related to fuel injection (during starting). Unfortunately, I'm not always clear on what they mean. There's also a fuel injection quantity listed (see pictures). I would rule out a control issue, but I'm not sure.
Here we are again, with a mechanical problem: The initial issue was a significant coolant leak through the exhaust. However, the connection between that and the current starting problem is completely unclear to me. It's unlikely that the cylinder head gasket is suddenly so severely damaged on all four cylinders that it wouldn't even allow one of them to ignite. Furthermore, there's no visible external leak in that area. I also can't imagine that there would be any consequential damage to the injectors overnight due to this. That just leaves the task of cleaning out the intake ports (see above).
If I don't come up with anything new by the time the holidays are over, I'll remove the Z-head cover and check if all the valves are opening fully. Assuming suitable weather conditions, no garage is available.