There is a lot of misconception and misunderstandings about how the DME operates, and what some tuners are doing to improve performance.

The 944/951 uses three different timing & fueling tables: Idle, Part-Throttle (PT), and Wide-Open-Throttle (WOT).  Which table the DME uses is determined by the Throttle-Position-Sensor (TPS).

If the idle-switch is closed, then the DME will only use the idle tables.  If the WOT switch (944NA), or full-load signal, is active, then only the WOT tables will be used.  Finally, if neither switch is active, then the DME defaults to the PT tables.

Lets take a look at the ignition timing tables for a stock 1987 951 DME:

The first thing that should be of interest is the difference in the sizes of the three tables.

The Idle and WOT tables are 2-dimensional.  These tables have only one axis, RPM, and only one possible value for each respective RPM.

The Part Throttle table, however, is 3-dimensional.  This tables has two axis, RPM and Load.  The PT table is where the majority of driving is done in.

Now this three-table strategy is the main reason the 951 feels sluggish before it gets "on boost", even more important than the long crossover design.  Because the WOT table is 2D, it cannot account for different load situations, so the DME must assume you are full boost anytime you are in the WOT table.  Obviously with a turbocharged car, full boost is not achieved the instant you depress the throttle.  During the time the turbo is building boost, fueling, and timing will not be ideal. 

For example, if you are driving down the highway at 3300rpm, your ignition timing advance will be ~35° BTDC (before top dead center).  Now, you go wide-open-throttle, immediately the DME transitions to the WOT table.  Still at the same or similar RPM, your ignition timing advance is now only ~16°, nearly 20° difference!  This extreme reduction of ignition timing feels like a flat-spot or sluggish.

So why not put more advance in the WOT table?  Well as mentioned earlier, the WOT table is 2D, and does not understand different loads.  The WOT table must be tuned according to the maximum boost pressure, and quickest boost response theoretically possible.  If the tuner puts an extra 10° of ignition timing in the WOT table, the car will feel better during the transition from the PT to WOT table, because there is less of a change.  BUT, if the car is in a higher load situation, such as going uphill or a higher gear (3rd-5th), the extra 10° of timing is too much and will cause the engine to knock - forcing the KLR to pull out timing, and possibly causing engine damage!

Unfortunately, adding this extra timing to the WOT table, just to make the car feel more responsive is what some tuners are and have been doing.

So what is the correct solution?

The correct solution is to enhance the WOT table by changing it from the flat-response 2-dimensional, to a true 3-dimensional table.  Only by doing this can the WOT table account for load as an axis (the other axis being RPM).

For example, lets take the situation from earlier, on the highway at 3300rpm, and you go WOT.  Instead of the flat-response 2D WOT table from earlier, the DME goes to a true 3-dimensional WOT table like this:

Now instead of ~20° drop in timing as before, there is only a few degrees - much more appropriate for the actual load.  As load (boost pressure) increases, the DME selects from the higher load columns, lowering ignition timing advance as needed.  Of course this 3D table, with load and RPM as axis, gives the DME a much more accurate table to choose the correct value for performance, responsiveness, and safety. 

This is the solution that we provide in every Rogue Tuning package.  As far as we know, we are the only ones who have rewritten the DME software that has 3-D tables.