[v8volvo]

Thanks for moving these, this is an interesting theoretical discussion that deserves its own thread.

Originally Posted by ngoma

Would the OBD provide a way to connect a VAG-COM to produce a timing curve graph? Displayed on a laptop screen in real time?

Exactly yes, that's what the mtdi guys are doing. Like what you can do with a full electronic engine, it can do with a mechanical one in terms of data acquisition, even without the computer calling the shots. And you would be able to watch the effect of all the variables also as you played with them -- engine/fuel temperature, playing with smoke screw, playing with IP case pressure, advance piston shims, governor changes, different injector opening pressures, different head and rotor size, etc.

The computer would register all kinds of diagnostic faults also of course, due to the other missing sensors and its inability to control any outputs, but with just the inputs of injector needle lift sensor and crankshaft position sensor it's known (from others' experience doing this process for their M-TDI conversions) that the OBD will still happily read out timing and RPM information.

What those guys then do is take readings of timing curve from their M-TDI engine running on its mechanical pump and compare them to a log file of what an electronic engine would be trying to achieve at the same RPM and load. They then manipulate the static and dynamic settings of their M-TDI configuration to mimic what the computer tries to do, or at least get reasonably close.

For us, the trick would be that we wouldn't have access to a "standard" map to compare to, for these old mechanical engines that never used electronic control. So in theory, in order to establish our own benchmark, I suppose we would have to first get a known perfectly running, freshly rebuilt and calibrated IP installed on a known good condition engine, along with a set of brand new freshly rebuilt injectors calibrated exactly to the factory pop pressure, and then take a full library of timing curve readings from that engine under all the various relevant conditions of temperature, fuel type, etc.

Then once we had that benchmark reference data, we would in principle be able to make running adjustments to our old, worn IP's if we could take timing curve readings from them and play with internal pressure to get them to match up as close as we could to the benchmark. Of course to do so we'd have to get all the timing and communications equipment set up there too.

Sounds like a lot of work and expense. We are probably better off just getting an IP properly rebuilt and professionally calibrated once in a blue moon.

Maybe a simpler but sufficient method would be using one of those Snap-on diesel timing kits that use a sonic wave sensor that clamps onto the injector hardline to signal what looks like a common timing light?

Interesting idea, that would certainly be easier to set up, though would lack the ability to give real time communication when driving under load since you'd have to be standing there under the hood holding the light. Maybe could still get close enough though? Or maybe there would be someone clever enough to transform the signal into something that could be used an input for a data log?

I think (?) some of the bench calibration process for setting internal pressures occurs with the pump at a variety of load/injected quantity settings. Maintaining a stable pump RPM to take readings at anything other than idle fuel quantity of course would be pretty tough without being able to drive the car and apply a load to the engine. Maybe having it running on a chassis dynamometer would do the trick though?