You all place too much stock in the capability of the electronics to make these mood swings.
The absolute best way to monitor this car's fuel control is to view the EHA current. Everyone that owns a KE car and wants any info about control should creat the proper test harness. Once wired the test can be done with little effort. The hardest part is removing the aircleaner.
What you must make is a harness with a male and female connection that allows you to unplug the connector at the EHA. One part of your rig will plug into the EHA and the other back into the connector you removed. The harness is wired such that the two wires that jumper the circuit have one wire broken and wired to plug into the ammeter sockets + and - . This is so that the current flowing must pass through the ammeter.
Make the jumpers long enough that you can view the current from the drivers seat. The added capacity of this testing is that by pulling one of the + or - leads you have opened the circuit and have assured zero current (which is the limp home position - coincedentally).
Now once you are monitoring you will notice that it will take a huge number of ma to make much difference. The only instance I can imagine would be the equivalent to decel cut-off. In decel cut-off the current goes to -60ma. This shuts off the motor by dropping the differential pressure to zero. This is easy to view. Go watch decel cut-off..... When the event occurs it will take a similar current (I would presume at least 30 negative ma) to get the drop in performance I am imagining from your description. If the current isn't close the elctronic fuel control is not the problem. You will notice that all other driving when warm never causes more than a 10ma change. The greatest change occurs on quick full accel. A combined effort of full throttle and rapid movement of the airflowmeter position potentiometer.
Get some perspective by viewing this, watch the road. All aspects of electronic control including closed loop can be monitored. I wish all aspects of control could be monitored so easy as KE.
If you need more info on this jumper/break-out lead send me a fax number and I'll draw it out. A picture of its use on a cheap RS meter is in my DIY article.
Anyone making diagnostic decisions on this system needs to do this testing. Not only is it the easiest test imaginable, it gives one of the best systematic views of control I can think off... everything in plain view.
Bosch Master, ASE Master, L1
33 years MB technician