Found this:
Setting the Throttle Potentiometer
This is an extract from a service bulletin issued by the factory in Oct 1992.

Model 907 & Cagiva Elefant 900 IE & GT models

remove the filterbox cover and the air filter; unscrew the anti-sticking screws i) and starter adjust ii). Screw up to the moment that the throttle is totally closed. Check this position by touching the filter side and having the throttle operated many times. (open the throttle a little and snap it closed ounce or twice)

Connect a digital voltmeter (accuracy=1mV) to the pin no 11 (-) and no 17 (+) of the central box terminal board iii) Then check that the voltage reading (switchboard On) to be 150 mV ± 15mV. Otherwise loosen the potentiometer iv) fastening screws setting their angular position up the the mentioned voltage figures are got. During this operation horizontal cylinder throttle must be forced to stay closed. (be careful what they say about being forced, what they actually mean is do not allow the turning of the potentiometer to open the throttle body butterflies)

Using the anti-sticking screw i) of the vertical cylinder (master) adjust the throttle opening up to a voltage reading of 300mV ± 15mV. In the same way adjust the anti-sticking screw of the horizontal cylinder up to the contact with its support, or when the voltage figure previously stated starts to increase. At this point the air flow rate of the throttle at idle rpm ratio is in condition to allow a positive and negative adjustment (rpm increase or decrease) using the by-pass screws.

Adjust the by-pass screws at one opening turn and after the usual preparation, arrange engine starting; under these conditions (20°C approximately for air, & oil temp.) rpm range must be stabilised at 900-1300 rpm approx. On the contrary, ( if this is inconsistent) check again with anti locking screws loosened, that throttles are resting on the duct indeed. (throttle body butterflies properly closed)
Go on with the usual procedure of adjustment (see our service bulletin No. 149 dated 20/06/90) specially paying attention to the mixture strength balance of the two cylinders (CO equal figure 3-6%) As a final practical test, alternatively disconnect the two ignition spark plugs, noting (using and auxiliary digital rev counter) that the engine rpm ratio decrease is repeated at the same figure.
Adjust the start opening screw ii) in a way to get an rpm rate of 3000± 500 n/1 with hot engine and starter totally turned on.
Notes for the above.

Anti-Sticking screws = throttle body butterfly stop screw
Starter Adjust screw = warm-up lever or choke
Central Box Terminal Board = ECU
potentiometer = in this case the throttle position potentiometer
The standard running configuration for these tests is to make sure that you have adequate ventilation for the exhaust. If you run the engine for extended periods you must have a cooling blower and the engine will have to be run using a remote fuel tank that holds just a small amount of fuel, remembering that the remote fuel tank must have a free flow return pipe as well as the normal fuel feed. The fuel injection system has a bypass that must be fed back to the tank. as an example I have a small lawnmower fuel tank with a fuel tap for feed and a return pipe welded into the top air space for the fuel return

these instructions are exactly what I have used to set the TPS previously - I have also "played" with the zero position (throttles shut 150 mV +/- 5mV) to make the mix richer / leaner...

they are good instructions and this will result in the engine having the perfect idle...and (in theory) good balance / mixture throughout all throttle settings.

BUT what do you do when the manifold vacuum is balanced at idle but different by 10% or more for each cylinder at 4,000 RPM??? This is where the bike spends most of it's time - running, not idling...

in my case i have had to adjust the butterflies from both set "fully shut" to about 1/2 a turn out-of-synch to get the cylinders balanced at 4,000 RPM (unloaded) which is probably about 3,500 RPM under load.
then turn the idle adjust screw to 300 mV / 1,200 RPM...
the idle manifold vacuum balance was then achieved with the air bleed (by-pass) screws.
I know there is a second idle stop screw in there - I just can't set it with the air box in place...so I just use the one closest the TPS.

I'm not going to throw the $$$$ at flow benching the throttle bodies - I can't afford that and don't require that level of performance...might be a different story if I was racing...

do you have a copy of the service bulletin referred to?

Ducman - you're one step behind me...lol
I emailed Brad at moto one yesterday...and sent a link to this forum.

My questions are:
how much does the MAP sensor output affect the ECU outputs - compared to the TPS? [I had expected a greater response delay with the damped air line]
and
could the ECU get upset by a wide-tolerance signal from a sensor [caused by instantaneous vacuum fluctuations]?

Brad indicates he owned an 851 [which has the same injection system as the 907ie].
what I did note is that his 851 air box was run with the top off - which would have some effect on the vacuum the MAP sensor sees.

he seems to know what he's talking about and has all the good gizmos to test results - hopefully he will respond.

Great information from that link, DucMan.

Man, this sounds like a real key to finding a solution to this problem --- the mapping possibly needs to be fine-tuned. Imagine that. So why did I buy an expensive FIM chip for the 907 that still leaves it stumbling and coughing??? I thought they were supposed to design mapping that got the engines to run and respond better than stock, wih open pipes, etc. I have the USA chip along with both Euro and FIM versions. And CObra F1 pipes and Termignoni carbon fiber pipes. Bike still stumbles.......

As a side note, I posted LT Snyder at DesmoTimes and asked about this problem. He said the 907 was one of the only bikes he had run up on that he could not completely sort out. He thought it was maybe the P7 ECU that is the problem. But evidently the 851 and some 888s also used that unit? Didn't those bikes run good? They raced , yes, but were also on the street. And are 4-valvoles also.

Please post any updates you guys have concerning what you find out. Thanks.

100 km`s today and I must say the stumble has gone from unbearable to acceptable w the ciggie butt. :thumbup:

More ign examples, 907 compared w the Grand canyon on Brad`s site:
F e at 2000 RPM, 10-12 % throttle: 8-9 degrees preignition on GC, 20 on 907!
2,8k/16 GC, 2,5k/30 907, almost full normal advance for a 2V 900!
Have 2k/14 on my carbed Monsters programmable ignition.

No ie expert but what do you think?