TBI vs Port injection

[eschaider]

You have a very good working knowledge of the various variables Luce. In general it is better not to attempt to characterize the injectors in car. The number of variables necessary to control consistently, makes it challenging to duplicate individual results test to test. Aggravating this, is the fact each injector needs to be characterized individually. By far the best (and least aggravating) way to have the injectors characterized is by a shop with proper flow bench equipment already set up to do SAE J-1832 characterizations.

The MS3Pro systems are very, very impressive — especially for their price. They will allow you to 'tune' each injector to provide proper fueling on a cylinder by cylinder basis if you want to split hairs. Once you have the J-1832 characterizations the job is not done yet. Even when injectors are done on a bench that is specifically designed to do the characterizations, you will still find data points that are 'flyers'.

Virtually (but not all) physical processes have a relatively predictable progression as you increase volumes, temperatures, voltages, pressures etc. As a result when you plot the points you should get a smooth curve. Invariably when you get the characterizations back, even from Ford, you will find flyers in the data that do not land on the predicted curve and in fact tug the curve from its natural position on a graph. Those flyers, along with how you set up your spark and fuel tables will, depending on their number and severity, affect part throttle drivability and throttle response.

To correct for this I use an Excel spreadsheet that takes the characterization data and plots it to look for the flyers. When I identify a flyer I move it, inside it's data set, to fit by using Excel's curve fitting tools to render the curve and then reposition the flyer, reflecting a real world physical behavior. When you do this, Excel will provide you with an "R" factor that essentially tells you how closely the curve fits your data or vice versa depending on your point of view. You want an "R" value of 0.99 or better. A 0.99 value means a 99% accurate fit. This extra effort takes additional time but in the end is worth it.

Below is a pic of what the sheet looks like. Each of the graphics is carefully massaged to be certain any 'flyers' are properly accounted for and then correctly rendered. This process is repetitive and duplicated for each injector. The process is also time consuming.


Ed