Quote:
Originally Posted by ngoma
So with the non-intact Pressure Control Valve in the IP like that-- means no dynamic timing advance, correct? Excess smoke, loss of power at mid to high RPMs?
|
Yes those were the symptoms. Presumably resulting from not enough timing advance at high engine speed because with the valve fallen apart like that and in effect running full bypass all the time, it couldn't crank up the pressure to move the timing piston and give the required dynamic advance.
Surprisingly it did not set a check engine light on my engine, which sometimes this issue will, if the injector needle lift sensor reports that the timing is retarded and the computer realizes it is unable to add enough timing advance to bring it to the desired value. I suspect that was because on this engine I happened to set the static injection pump timing set at the high end of the acceptable range, so that even without the benefit of full ability to control dynamic advance, the final timing was still close enough not to trigger a fault. But it was obviously off enough to affect the way the engine ran, fault or no fault, which is funny. The primitive onboard diagnostic functions of these early engines presumably allow for more error / wider tolerances than a modern system would, especially for something like injection timing that has a very large impact on tailpipe emissions.
As you pointed out before also, on a mechanically controlled IP like on D24/T, simply tapping the valve back together like I did on this pump would not necessarily achieve a satisfactory result for restoring the dynamic timing curve. That would have to be done theoretically on a calibration bench for our engines I suppose. Whereas with these TDI pumps less precision is demanded from the mechanical systems, since the computer has control over a solenoid valve in the IP that can fine-tune the timing advance with feedback control. Thus as long as the mechanical systems can get the timing to within any kind of reasonable ballpark range, the computer will be able to take it from there and hit the desired value. So what's "good enough" for this IP, in terms of shade-tree methods, would not necessarily be good enough for a D24/T.
I see my pictures aren't showing in the earlier post, I'll fix those.