Using this motorcycle (French and rare!) Much more often than my car (see Mercedes 300TD experimentation) I started a doping mount with water vapor.
Voxan Scrambler Engine Features:
72 ° V-twin engine of 996cc developing for this version 85CV (instead of 100CV for the other models).
Magnetti Marelli injection (IAW26V calculator), sensors: potentiometer on the throttle shaft, N engine revs, engine phase, water temperature, air temperature, barometric air pressure. So no measurement of the real air flow nor regulation by lambda probe.
Consumption without modification with optimum adjustment of the throttle potentiometer (150mV throttle fully closed).
stabilized speeds in 6ème (reduction of origin)
These measurements are close to those given by BMW for a R1150RS for example.
It is, I think, important to leave with values before modifications and on a sound basis, engine well adjusted etc.
Changes:
1 ° consumption measurement:
I used a fairly simple process, I count the cumulative injection time with logic circuits (CD4060 and 4011 pulse generator with a quartz, divider and a logical AND with the injector control), every X ms I send a pulse to a bike counter (SIGMA) that gives me Liters / Hour (instead of Km / H). So I have the totalization, the average, the maximum, and the instantaneous value.
Since 6 month I could compare with each full and the error is at most only a few 1 / 10L.
2 ° measure of wealth:
A preheated narrow-band type lambda probe (automotive recovery) is positioned on the rear cylinder tube.
A small montage based on LM3914 see: wealth controller
allows me to display the wealth on the dashboard with a leds baragraphe.
3 ° wealth setting:
I use a variable resistor (dimmable potentiometer on the dashboard) that modifies the value of the air pressure sensor and can allow me to deplete up to about 40% (high altitude simulation and low air pressure) .
4 ° steam generator:
It consists of 2 copper tube meters of 4mm inside diameter (6mm outside) wrapped around the exhaust tube of the rear cylinder. The temperature is very high (the copper changes color) but I noticed that some small drops of hot water were still passing.
There is an intermediate piece that allows drops of water not to pass easily (copper Z-piece).
The arrival of steam is in the air box so above the throttles, the depression is low but the steam comes under light pressure.
5 ° water injection:
No bubbler but a water injection with a minipump water coffee! Very small, good pressure (up to 3,2 bars).
Cost 25 € retail at a dealer in my city, we can find used in a machine HS.
This type of pump operates on impulse, with 220V alternating with 50 Hz we have one pulse every 10ms. But this pump can not run continuously with 50Hz (heating) it is intended for temporary use.
It is controlled here thanks to the consumption measurement circuit with a dividing stage (CD4060) and therefore proportional to the fuel injection. The pump is rewound in 12V. Attention for the command I use a MOSFET IRF840 because there is a strong surge of more than 400V (zener diode and capacitor of protection).
The command pulse is 10ms (at 5ms this is not enough). The pump does not heat because the pulses are spaced apart (it is not 50Hz) a few seconds between each pulse at idle.
At the moment I do not use a reactor in the exhaust I do not think it is useful for a gasoline engine and I admit that I do not want to drill my stainless steel exhaust tube whose price is certainly enough important.
I would like to get 10 to 20% as shown here: better-mileage simply by giving the engine water vapor.
Findings:
The first measures tell me that, without change of wealth, there is no drop in consumption.
Some pictures
Pump :
Pump disassembled (piston with spring valve and seals):
Pump (12V winding) with control circuit:
Steam generator:
Exhaust with cache:
Dashboard with info and settings:
Lambda probe added:
Circuit for counting consumption:
Michel
by a small box that is of quality and originality, but which is alive despite all his efforts 
