Poor Starting 912iS

The ECU didn't arrive by the weekend, so I didn't get the chance to eliminate this component. I did re-examine air in the Fuel System, I placed an Analogue Pressure Gauge in line with the fuel pipes, the pressure was good at 3 Bar and remained at this value even when starting it. After the pumps were switched on for approx. 40 sec the air in the system and the gurgling noise stopped. I tried starting it even though the wind was gusting 40kts in Lincolnshire, England today. It started first time and stopped 2sec later, it took 4-5 goes before it ran Ok. I did notice that the Fuel flow from Lane B wasn't reading but was Ok on Lane A, Fuel Pressure on the EMU varied between 3.0 - 3.5 bars. Once the engine was going fuel pressures were Ok. The Lane A&B Lights were Flashing and remained On after starting, they cleared when toggled. I am not going to try anything or second guess until I substitute another ECU. 

Reviewing the magic equation again. Fuel + Air + Squeeze + Spark = Bang.

The Air is mechanically controlled by the Throttle.
The squeeze is fixed by the design.
The spark is controlled by the ECU.
The Fuel supply appears to be good but the injection is controlled by the ECU.
It would appear that the ECU is NOT detecting some event that it expects to see and shuts down the fuel and spark.
On the second start, the missing event occurs and all continue normally.
That missing event should be in the ECU Log file.
- - -
While we are waiting for the ECU swap ( I do not have a warm feeling this is going to be helpful, but I have been wrong a lot lately)
Let's try this...
Turn on the fuel pumps, Start Power, and both Lane Power switches.
Then wait an actual full minute before cranking.
If it a sensor that is slow in initializing, this start would be occurring at about the same time as a normal second start.

We might also try starting with only a singe Lane A or B powered up instead of both and see if it makes a difference.
The engine normally starts on B so Try A ON, B OFF First.
- - -
Also if the ECU Swap solves the problem, do not call it solved until you reinstall the "BAD" ECU and confirm the problem returns.
The problem may just have been a marginal connection that was resolved during the swap.

Re-reading your previous post...

"I did notice that the Fuel flow from Lane B wasn't reading but was Ok on Lane A"
This means that the "B" Fuel injectors were NOT Injecting. (No B Flow!)

"The Lane A&B Lights were Flashing and remained On after starting"
This means there were identifiable Faults that were most assuredly recorded in the ECU LOG.

You need to get the LOGs downloaded.
Having access to a BUDS Dongle and Software is NOT an Option if you own one of these engines.

- - -

I would recommend looking at the connectors on the ECU to confirm that they are fully seated and secure.

It might be worth a shot to swap MAP sensors. When I had a Lane B MAP sensor error (shown on the BUDS program) I also had indicated fuel pressure problems . A new Sensor cured the issue. When ordering the sensor I noticed that the part number has been changed, indicating a possible problem with the first production run of the sensors, my engine was an early one SN 4417255 so I expect it had the first run of MAP sensors. 

I have a BUDS Dongle and it doesn't show anything wrong apart from the usual errors which have always been there since the engine was new. I sent this data to the UKRotax dealer because they won't let us lowly mortals have the full suite of software, they reckon the Voltage to the ECU was dropping below 10v. I have removed and cleaned all the Plug & Sockets. I have measured the Voltage to the ECU with an oscilloscope and it holds up at 12v even during cranking. I really need to swap the ECU just to eliminate it, before trying more complex and time consuming things. The sensors are ok. We had an ECU fail on the previous engine and again it was a bizarre fault that took several weeks to trace it to the ECU. The support given on these engines is poor, I wouldn't recommend anyone to buy one. Ideally the BUDS system should help but in practice it falls short of the mark. It would help if Rotax would supply the full software suite and circuit schematics - I don't understand why they are reluctant to supply this information. I will try and get an ECU for next weekend and report back. Thanks for everyone's suggestions and advice.