Exhaust Valve Sticking - Poor Emissions - Part 2

[martinr]

This case study follows on from Part 1, at:

topic23539.html

The engine is an 8-valve petrol Rover K Series engine and, from the exhaust note, it was clear - given previous experience with this particular engine - that an exhaust valve was sticking intermittently.

A 4425, 4-channel Picoscope was used to investigate, and the opportunity was taken to compare the exhaust pressure pulses, as seen at the tailpipe, with those emerging from the exhaust ports and entering the downpipe just below the exhaust-manifold connection. (This has also been covered here: topic23557.html .)

Channel A Blue trace, Fluke pressure transducer fitted to the blanked-off boss in the downpipe, next to the 02 sensor.
Channel B Red trace, a GM fuel-tank pressure sensor sensing the pressure at the exhaust tailpipe.
Channel C Green trace, a Renault MAP sensor detecting the pressure in the inlet manifold.
Channel D Yellow trace, Secondary ignition HT probe, giving No 1 firing line as the sync signal.

All pressure traces, including those of the inlet manifold, are displayed as positive pressure upwards. And all pressure sensors are DC coupled - this is important..

The phase rulers were set using data from the Haynes manual: ignition timing at idle is 15deg BTDC.

The screenshot above captures a missing-exhaust-pulse event at the downpipe (blue trace) and the corresponding event at the tailpipe (red trace), some 10ms later. From this, it is clear that the exhaust valve in No 1 cylinder is the culprit.

But looking at the green trace - that of the pressure/vacuum in the inlet manifold - there is a small anomaly that coincides with the firing line in No 1 cylinder, and this anomaly, also intermittent, coincides with valve-sticking events. So the question is: what, if anything, does it tell us, and could it contradict the diagnosis that No 1 exhaust valve is occasionally sticking?

Zooming in on this anomaly, it appears to occur between around 695 degrees of rotation and 3 degrees of rotation. And the suspicion is that this anomaly, being a positive-pressure pulsation, is caused by the escape of pressure through No 1 exhaust valve (during the compression stroke) into the exhaust manifold but that this then leaks into the inlet manifold during a valve-overlap event. Using data from the Haynes manual, No 4 cylinder is in the valve-overlap stage between 717 degrees and 5 degrees. But there is sufficient difference between the measured 695 degrees and Haynes’ 717 degrees to raise some doubt. So, a previously-obtained waveform determining TDC (as found from a Fluke pressure transducer) with respect to the crankshaft position, was used to check the true relationship between TDC and the firing line (at idle) of No 1 cylinder on this engine.

And, zooming in

It’s clear that No 1 firing line occurs not at 15 degrees BTDC but at only 2 degrees BTDC (topic for another case study?). Therefore, 13 degrees must be added to the measurements to obtain a true value. So 695 degrees becomes 708 degrees (cf Haynes’ 717 degrees), and 2 degrees becomes 15 degrees (cf Haynes’ 5 degrees). Still not quite perfect, but the middle point of the valve overlap, as given in the Haynes manual, between 717 degrees and 5 degrees, is 1 degree after TDC. And the mid-point of inlet-manifold pulsation, as measured (and adjusted for the true position of TDC, is midway between 708 degrees and 15 degrees i.e. 1.5 degrees. The mid points therefore coincide..

This now gives a warm feeling to the belief that the pulsation detected in the inlet manifold is conclusive evidence that No 1 exhaust valve is the root cause of the problem.

The solution has been mentioned here: topic23539.html and the work is currently underway at J and E Engineering in Rossendale. The refurbished (spare) cylinder head will not be fitted until next year, which offers, in the meantime, opportunity for objectively testing proprietary products such as Nitrox Hot Shot and Forte as non-invasive solutions to valve sticking. The machining work - skimming, valve-seat recutting and K-Line valve-guide inserts - was touched on in Post #8 at topic23539.html will be a big improvement on the Rover OEM cylinder head.

This case study also shows that tailpipe pressure pulses are not suitable for fault diagnosis, the closer one can get to measuring at the exhaust manifold the better; however, if the tailpipe pulses show perfect consistency, then the tailpipe can be a valuable first-look diagnostic checkpoint, avoiding the need for more involved measurements.

To be continued……

[wiki38andy]

The screenshot above captures a missing-exhaust-pulse event at the downpipe (blue trace) and the corresponding event at the tailpipe (red trace), some 10ms later. From this, it is clear that the exhaust valve in No 1 cylinder is the culprit.

Missing an exhaust pressure pulse from exhaust side, is an evidence for a misfire only. This means that the combustion did not take place. Not necessarily a sticky exhaust valve. Can be an ignition issue, an injector leak, .....anything which can prevent fuel from being burned. Disconnect a spark plug or an injector and you will get the same waveform.

[martinr]

The screenshot above captures a missing-exhaust-pulse event at the downpipe (blue trace) and the corresponding event at the tailpipe (red trace), some 10ms later. From this, it is clear that the exhaust valve in No 1 cylinder is the culprit.

Missing an exhaust pressure pulse from exhaust side, is an evidence for a misfire only. This means that the combustion did not take place. Not necessarily a sticky exhaust valve. Can be an ignition issue, an injector leak, .....anything which can prevent fuel from being burned. Disconnect a spark plug or an injector and you will get the same waveform.

Good point. But what about the proceeding positive-pressure pulse (annotated as “Anomaly in inlet manifold”) detected in the inlet manifold at TDC on No 1 cylinder compression stroke, coinciding with valve overlap on No 4 cylinder? An ignition misfire or a lean-burn misfire in No 1 wouldn’t give you that, would it.

The other point worth noting is that the secondary ignition trace (yellow) for No 1 cylinder doesn’t indicate a misfire, although at the start of this investigation I wondered what was causing hash on an otherwise fairly normal burn line. topic23536.html?&p=105613&hilit=Hash#p105613

[wiki38andy]

Roughly speaking, at idle, the valve overlap event begins at 15* before TDC and ends at 15* after TDC. Could be more or less on any side, but not important for the following analysis. Between 30*-60* before this TDC, an exhaust valve opens releasing a pressure spike inside the overlap window time, as you marked with cursors on second picture. During the compression stroke of cylinder with supposedly faulty EXH valve, the pressure leak is smooth over the whole stroke time and continuously evacuated at a low pressure level. It is not suddenly released before TDC like in EVO event, and so it is not a spike that can be detected by INT pressure sensor in overlap window. Moreover, the pressure hump you mention on INT waveform, can be noticed few times even when the EXH waveform shows no issue with any valve, on first picture.