2004 Citroen Berlingo 1.4i TU3JP Engine noise: Is this detonation knock or rod knock?

[Pilot65]

2004 Citroen Berlingo 1.4i TU3JP Engine noise: Is this detonation knock or rod knock?
Lexia says normal temps (engine/coolant and IAT), no DTCs, normal MAP values.
The noise was noted about 6 weeks ago, and seems to be getting worse every week.
Changed from 95/E10 to 98/E5 but it made no difference.

Listen to the noise here:

[Paul-R]

Link says

Video unavailable
This video has been removed by the uploader

[PaulC5]

Sounds as if there is something loose that is knocking. Have you checked the exhaust, engine mounts, crank shaft pulley, etc. Engine knocking (pinking) I would have thought is heard when accelerating whilst driving and not whilst stationary so trying to show it as you have done might not be the best way. Has the timing belt been changed when required ?

[xantia_v6]

Video is fine here. It sounds slightly more like a combustion knock to me, but it is hard to tell. If you put it in gear with the brakes on and slip the clutch to add some load, can you get it to knock continuously? the noise might be more distinctive under those conditions.

A diagnostic session, noting the timing values and knock sensor output may tell you something.

[Pilot65]

Video is fine here. It sounds slightly more like a combustion knock to me, but it is hard to tell. If you put it in gear with the brakes on and slip the clutch to add some load, can you get it to knock continuously? the noise might be more distinctive under those conditions.

A diagnostic session, noting the timing values and knock sensor output may tell you something.

Well, the noise do not increase with load or rpm. When driving uphill there is no usually no noise. That noise comes only when the throttle is opening, and at light load or no load, like when driving at steady speen on a flat road and just barely touching the gas pedal.
As seen in the video the noise is really only there while the throttle is opening, not when it is closing or held steady. The noise is locked to, and synchrounous to rpm, and from what I can tell it seems to come from the middle inside of the engine. In the video there are a few moments where we hear what I believe is real ignition knock as I play with the throttle, and this is a totally different beast, its loud and hard and and the engine feels and sounds as if hit by a sledge!
Today I:
-adjusted the valves, were a little bit out of spec but it made no difference. Checked the plugs and all looks fine.
-while the plugs were out I tried to probe for any piston slop (=worn rod bearings)... moved the crank so the piston starts moving down (any play will now be at the top cap) then using a wrench extender pushed at the piston with 5Kg or so, and measured at the end of it with a 1/100mm dial gauge.... Difficult to do and sketchy setup so the results were inconclusive, but sertainly less than 0.1mm.

One thing I noticed in Lexia live data is the ignition advance... its all over the place. At 2-3000 steady rpm and no load its pretty stable at about +25deg or so. But at idle no load its all over like +15 to -8, for no apparent reason...

[ozvtr]

One thing I noticed in Lexia live data is the ignition advance... its all over the place. At 2-3000 steady rpm and no load its pretty stable at about +25deg or so. But at idle no load its all over like +15 to -8, for no apparent reason...

It does sound like a timing or fueling issue. It sounds like it's "bogging down" on acceleration. Is the acceleration poor? Got hesitation?
In the old days I would have said it was the vacuum advance in the distributor, or the accelerator pump in the carb.
Vacuum leaks can cause the engine to bog down on acceleration, but it tends to affect idle more (poor idle or hunting).
Timing is controlled by the crank angle sensor. I would expect problems across the board if it was the crank angle sensor. Or the engine just wouldn't run.
Manifold Absolute Pressure (MAP) sensor might be a bit lazy and not reporting the change in pressure to the ECU. Could try unplugging it and see what happens?
Have you checked the fuel rail pressure? Should maintain a constant 3.0 to 3.5 bar.
Doubt it's the fuel injectors. One maybe, but not all four. More likely the fuel pump rather than the injectors.
Not much to the TU3JP engine, pretty simple little beast. That's why I like it.

[Pilot65]

One thing I noticed in Lexia live data is the ignition advance... its all over the place. At 2-3000 steady rpm and no load its pretty stable at about +25deg or so. But at idle no load its all over like +15 to -8, for no apparent reason...

It does sound like a timing or fueling issue. It sounds like it's "bogging down" on acceleration. Is the acceleration poor? Got hesitation?
In the old days I would have said it was the vacuum advance in the distributor, or the accelerator pump in the carb.
Vacuum leaks can cause the engine to bog down on acceleration, but it tends to affect idle more (poor idle or hunting).
Timing is controlled by the crank angle sensor. I would expect problems across the board if it was the crank angle sensor. Or the engine just wouldn't run.
Manifold Absolute Pressure (MAP) sensor might be a bit lazy and not reporting the change in pressure to the ECU. Could try unplugging it and see what happens?
Have you checked the fuel rail pressure? Should maintain a constant 3.0 to 3.5 bar.
Doubt it's the fuel injectors. One maybe, but not all four. More likely the fuel pump rather than the injectors.
Not much to the TU3JP engine, pretty simple little beast. That's why I like it.

Yes I agree. No, the acceleration speed and power are all normal. I once had a engine with distributor ignition, and one of its vacuum hoses came loose while at 60Kmh... worst engine noise I've ever heard, I was sure a broken conrod was slapping around on the crank. So engine detonation knock can be really bad, no damage in this case as I got the ignition off after 1-2secs.

For the TU I lean heavily toward a fuel/ignition/detonation scenario... The noise is what I'd call a ignition knock, not 'fully developed' detonation of the worst type, but just barely there and maybe too weak to be detected by the knock sensor. But why is it there now when it wasnt before... Something must have changed. CPS fail would totally upset the engine so it wouldnt even start. MAP? Maybe, it reads the right pressures but if it were way too slow in response, or has developed hysteresis,

Have tried spraying brake cleaner around injectors and manifold flange, the engine did not respond to it so no apparent air leaks.
I think we can exclude low fuel pressure as it accelerates and climbes up-hill as good as its able to with all of its 75Hp .
Will try a BSI/ECU reset and throttle body relearn. And have a look at the front for loose pulleys and camshaft timing. And examine the knock sensor, maybe remove it and knocking on it while engine is running...

[ozvtr]

For the TU I lean heavily toward a fuel/ignition/detonation scenario... The noise is what I'd call a ignition knock,

It does sound like the timing is off...but I don't know how that could be. The timing is controlled electronically by the engine ECU.
Could it be the knock sensor screwing things up? I can't see how. The engine ECU is looking for a particular...um...sound, from the knock sensor. I can't see how the knock sensor could be replicating that sound and fooling the ECU. Disconnecting the knock sensor might bring the engine light on, but it won't do any harm.
Would be nice if you could put a timing light on it. LEXIA gives timing details in live data. Don't know if OBDII gives timing.
Hmmm. Strange that the acceleration is ok. I would have bet that the engine would bog down, judging by the sound.

[Pilot65]

For the TU I lean heavily toward a fuel/ignition/detonation scenario... The noise is what I'd call a ignition knock,

It does sound like the timing is off...but I don't know how that could be. The timing is controlled electronically by the engine ECU.
Could it be the knock sensor screwing things up? I can't see how. The engine ECU is looking for a particular...um...sound, from the knock sensor. I can't see how the knock sensor could be replicating that sound and fooling the ECU. Disconnecting the knock sensor might bring the engine light on, but it won't do any harm.
Would be nice if you could put a timing light on it. LEXIA gives timing details in live data. Don't know if OBDII gives timing.
Hmmm. Strange that the acceleration is ok. I would have bet that the engine would bog down, judging by the sound.

Yes but the timing is entirely sensor driven, no kind of adjustments anywhere. Last night I took the knock sensor to the workbench and connected a DVM on AC range and knocked on it with a 13mm spanner; even with a simple DVM I saw voltage spikes of nearly 0.1V, so sensor is good. Did the same in the car with engine running but noticed no ignition retard. This was as expected as this 'test' produces no proper high frequency signal at the correct time, The ECU filtering will most likely require a HF signal in the Khz range, and in the proper crank sector, I guestimate 0-60deg before TDC.
Figured I'd check the sensor wires for continuity between the sensor- and ECU connectors... and wouldn't you know, one wire shows a open circuit; Wire 134 from crank sensor [1120] going to ECU harness connector 32V MR pin B3 is ok, but wire 135 going to B4 has no continuity... Unbelievable! I'll have this fixed but I doubt it will make much difference.

Side note; a good while back i got tired of servo squeal so I finally tightened the serpentine belt. I was some time AFTER after that this noise came (or was noticed). And now I see a possible connection.
So now I'm thinking; if the crank is pulled up hard by the belt it (the front end) could now be riding mostly in the upper bearing shell, with 'some' clearence to the lower bearing shell. And in 3 of the 4 engine strokes the vertical crank force is downwards, only on the intake stroke is the force upwards as the crank pulls the piston down, but at this time the 3 other pistons are at Compression - Power - Exhaust, so their force is downwards.... At idle or no load this may not be enough to overcome the belt tension force (which is probably like 100Kgf), but as soon as the throttle is opened and moving up we get a much bigger 'bang', and that may be enough Power stroke force to slam the crank in to the lower bearing shell. Just my theory, and today I'll test this because its time for an oil change:

A) Take the force off the serpentine belt, or remove it.
B) Drain 1/2 of the oil and replace with diesel, this mix should result in a pretty thin oil roughly like the dreaded 0W-5, engine should cope with this for a few secs at idle no load.
C) during oil changes I always flush out the last of the old goo with cheap thin mineral oil, diesel or kerosene, but today I'll use 20W-60 for that. Then fill with correct type oil.

The large difference in oil viskocity between B and C should reveal if there is large bearing play down there. B = more noise, C = less or no noise.

[ozvtr]

The noise is what I'd call a ignition knock, not 'fully developed' detonation of the worst type, but just barely there and maybe too weak to be detected by the knock sensor.

I'm sure I could here it knock at the 55 second mark of the video. So I would tend toward a lean mixture. But I still cant understand that the engine DOES NOT lack power on take off! If there was no fuel boost on acceleration the engine would bog down.
Have you tried disconnecting the MAP sensor?
The knock sensor is really only there in case someone uses poor fuel. The fuel maps are designed around 95RON fuel, or better. Well here in Aus anyway. I understand that in Europe there are all sorts of crazy versions like duel fuel and LPG. I'm not familiar with any of those versions...sorry. Yes, carbon build up will also cause knocking but I'm wandering away from the actual problem here!

[Pilot65]

The noise is what I'd call a ignition knock, not 'fully developed' detonation of the worst type, but just barely there and maybe too weak to be detected by the knock sensor.

I'm sure I could here it knock at the 55 second mark of the video. So I would tend toward a lean mixture. But I still cant understand that the engine DOES NOT lack power on take off! If there was no fuel boost on acceleration the engine would bog down.
Have you tried disconnecting the MAP sensor?
The knock sensor is really only there in case someone uses poor fuel. The fuel maps are designed around 95RON fuel, or better. Well here in Aus anyway. I understand that in Europe there are all sorts of crazy versions like duel fuel and LPG. I'm not familiar with any of those versions...sorry. Yes, carbon build up will also cause knocking but I'm wandering away from the actual problem here!

Yes tried disconnecting the MAP, and the engine runs kinda funny... but still same knock on acceleration. Also the knock senssor wiring I mentioned is actually correct, I just looked at the wrong schematic version.
About the belt & oil trials I mentioned: Not really conclusive but what I got is; notably less knocking after taking tension off the serpentine belt. Retightened it and changed to 10W-60 oil (!)... again notably less knock, alnost gone. What all this tells me is that a (normally) tight serpentine belt increases the knock, which points to worn main bearings, says nothing abot big end bearings. And thick oil affects both main and rod bearings, maybe targeting the rod bearings more. So the conclusion is that all bearings are well worn, as would be expected after 225000Km and bare minimal maintenance... head top looks like its seen nothing but mineral oil, IOW black as an oven.

Another possibility is that the upstream lambda has drifted and reads too rich, causing the ECU to lean and create knock. But I doubt this as the knock is there even on cold engine when it is operating open loop and the lambda is ignored. The knock sound is less on cold engine and increases with temp, again pointing to bearings. I may have a go at replacing the rod bearings from below, but this is impossible for the main bearings, the engine must be flipped upside down to remove the bearing ladder / bed to get at the mains.