Before crossing the alps I installed an air pressure sensor, taken from a 1125R, to my airbox. I made a simple voltage regulator (as shown in the Tuning Guide V2) and connected it to the 3-pin Deutsch plug, commonly used to trigger the exhaust valve motor. Contrary to the 1125 engines, where the pressure sensor measures air pressure outside the airbox, this one senses the airbox pressure.


I then went for a test ride from Frankfurt to the Grosser Feldberg, and back (logfile).

The elevation change is 2,500 ft, which results in a theoretical change of air pressure of -9.5 kPa. Absolute values at the time of this ride: Frankfurt 101.4 kPa, Feldberg 94.5 kPa, barometer readings from a device in the tank bag.



The image shows the dynamic pressure correction in Graph 4, Flags 5 (red line), dropping from 101% to 94%, which seems to meet real conditions quite well. The static pressure correction is shown as Flags 6 (orange line, not really visible, as it's not changing at all). It's value of 94% is a bit off, as is the stored baro value, (I would have expected to see it at 100%), so I 'fixed' it by adjusting the pressure sensor data to always show a pressure 4% higher than compared to the sensor's technical data. The pressure sensor voltage is shown as Graph 4, Unknown-63 (light blue line). The dynamic pressure correction follows the airbox pressure with a bit of delay and a staircase shape.

On my ride accross the Alps I did not notice any change in performance or mileage, so from that point of view the installation of an airbox pressure sensor is not a requirement. Seemingly the ECM is capable to handle air pressure changes with it's existing means (EGO correction, AFV) quite well.

When looking at a part of the log, where the bike was ridden with a higher speed (Graph 1, green line), a significant drop of airbox pressure is shown.



The bike had a stock 2004 model airbox (fully closed) and the 'snorkel' inside the intake flange through the tank installed. The pressure drop might indicate, that the intake system before the airbox is too restrictive to provide enough air for a high power ride.

To be continued ...

_________________
"It is often said that before you die your life passes before your eyes. It is in fact true. It's called living."
Terry Pratchett

Testride 2, changes applied: removed snorkel from intake, then smoothed out the intake flange to airbox base with a some gaffa tape. By removing the snorkel the area available for the air stream has almost doubled, so any restriction should be removed now. (The intake snorkel is part #4 in the image below.)


Went for the same ride (Frankfurt - Feldberg) as before, air pressure in Frankfurt was 102.1 kPa (logfile).



For better visibility of the AFV I hid O2 voltage and EGO correction in Graph 2. "AFV of the day" was 105.6%. When airbox pressure starts dropping (Graph 4, blue line) because of the elevation change (left quarter of the image), AFV is decreasing also, because the dynamic pressure correction is a bit delayed. Later, after the air box pressure correction caught up, AFV jumps back on it old value. So, basically, dynamic pressure correction is working, but requires a bit of fine tuning. There's a similar effect when riding downhill again, in the rightmost quarter of the image, for the same reasons.

Focussing on the high speed part again, the results are surprising:



Almost no change at all, quite contrary to my assumptions. Airbox pressure is still dropping significantly if engine speed is above 3600 rpm and throttle is more than half way open. At first glance, airbox pressure is dropping as soon as a WOT condition is met - seems even a bit counterproductive to add 10% fuel, when the air pressure is dropping so much.

If both test rides are compared in the MegaLogViewer, the air box pressure curves show almost the same run:

The comparison shows a fixed and constant offset of the two pressure logs (Graph 4, light blue and middle blue lines), caused by the different (static) air pressure of the two days.

This leads to the conclusion IMO, that despite all rumours, the snorkel is not a restrictive part of the intake system, otherwise air box pressure should remain stable now, independent of load or engine speed. Wether the snorkel is or is not installed, air box pressure still keeps dropping at high power conditions.

_________________
"It is often said that before you die your life passes before your eyes. It is in fact true. It's called living."
Terry Pratchett

Is this with DDFI3?

_________________
08 Specialized Langster

Part three of the airbox tests. This time a moderately perforated inner air box cover had been installed. About 30 holes with 1/2" diameter, mainly at the front, rear and upper-left side part. I also left the hole in the airbox base open and the intake de-snorkeled. The outer airbox cover remained the same, without the "breathing grill" above the gas cap, as were installed for 2006 and later models.

I skipped that part of the test ride with the large elevation change, as the dynamic pressure correction has been proved to work sufficiently enough, so made only the high speed part on the motorway, which still includes an elevation change of 600 ft. Air pressure in Frankfurt was 102.8 kPa when I started the ride (logfile).


Airbox pressure sensor data are still shown in Graph 4 (blue line), and the pressure fluctuation was not large enough to trigger the dynamic pressure correction, therefore the red line remains virtually unvisible at a constant 100%. Stored baro and static pressure correction (orange line in Graph 4) is still some few percents too low, but EGO correction and AFV would cope with that. "AFV of the day" is still 105.6%, as was the days before.



The perforated inner airbox cover obviously solved the air flow problem, as no (significant) drops in air box pressure are shown any more, even running at WOT and higher engine speed. Reasons? I can't tell. Maybe the flow at the left air scoop is not designed really well and creates a local spot of low pressure.

The overview shows two deviations in AFV, shown in more detail below:


This one (shown in the left side of the overview) is probably caused by a glitch in the fuel map. The AFV increases to it's original value only little later.


The second one (shown in the right side), is a good example why open loop learn should never be switched off. In this case, the WOT row of the fuel map apparently has a few "lean" cells, so after acceleration enrichment is finished, the mixture is "lean" (from the narrow band O2 sensor's point of view), and OL Learn increases AFV by one step (5.6%) after a delay of 3 seconds. As soon as the calibration starts again, this increase is corrected and AFV drops back to it's old value.

_________________
"It is often said that before you die your life passes before your eyes. It is in fact true. It's called living."
Terry Pratchett

The bike has still the original BUEGB installed. As I swapped the muffler for one without a valve, the AMC feedback pin (black - 9) has been reconfigured to now read baro sensor data instead. I copied the sensor and correction data from a 1125 ECM, then adjusted them to what seemed to fit better IMO:

_________________
"It is often said that before you die your life passes before your eyes. It is in fact true. It's called living."
Terry Pratchett

Cool. A bit like in the tuning guide on the ECMSpy site.

_________________
08 Specialized Langster

Interesting read

I am wondering whether the position of the sensor is in a local area of low pressure and with the change in the entry flow pattern this goes away. I was thinking of trying to mount something similar on the top of the cover at the centreline of the throttle body. However, if you did this, you would not be able to isolate the effect of the filter from the box.

But going back to your data, the fact that the AFV rises suggests that the increase in flow (higher airbox pressure) is real

Turning the AFV OLL off and running with a WB O2 sensor would confirm how much more flow there is.

Cheers

Nick

_________________
08 Specialized Langster

That's what I was thinking about also. The perforated airbox has some holes just above the pressure sensor (the place where a tank pad would go), and I felt a bit unhappy with that position, because in the case of a local low pressure spot this would simply equalize pressure difference, but without a real significance in respect to air flow. But, OTOH, all measurements would share this problem. As sensor readings are evaluated at the same crankshaft position each time (controlled by the blips on the camshaft rotor cup), it might also have been just temporary low pressure waves that had been recorded. But I doubt that pressure waves from the stack would make it through the (stock) air filter.

Anyway, thanks to the many holes in the inner airbox cover, I could easily route a tube from the sensor to where I would like to measure air pressure. Might be worth to try recording pressure inside the left air scoop or at the front side of the airbox.

I never mapped that bike using WB-O2, just tried to make sure the mixture is always rich at WOT, so it might well be possible, I simply missed one or two cells and that lean spot shown in the log is not caused by a change in air flow, but was always present, just went unnoted. A first estimation will be, that it's less than 5% lean, as AFV changes 5.6%, and this provides enough fuel to make the mixture go rich (in NB-O2 terms ...).

_________________
"It is often said that before you die your life passes before your eyes. It is in fact true. It's called living."
Terry Pratchett

A résumé after a few tankfuls: mileage has slightly increased, about 8-10 mls per tankful. This should be caused by less fuel consumption in open loop only, so I assume, that baro corrections is better graduated than AFV.

_________________
"It is often said that before you die your life passes before your eyes. It is in fact true. It's called living."
Terry Pratchett

how about combining a 2007 airbox with the lower part of a pre 2006 airbox base ( dual air admission ) on a pre 2006 bike ?

i also be thinking about putting a 2'' aluminium 90° tube in the left airscoop to make a better transition for the air admission

can't we have a more precise correction by putting the baro sensor between the airfilter and the butterfly ?

Good question. I don't know. I assume, watching if/how O2 voltage changes with rising engine speed at WOT would give a first impression. I'm not sure if pressure waves would impact the baro sensor readings, when installed so close to the velocity stack.

_________________
"It is often said that before you die your life passes before your eyes. It is in fact true. It's called living."
Terry Pratchett

The ECM records pressure at the same time(s) each cycle so pressure waves shouldn't make much difference.

_________________
08 Specialized Langster

That's what made me think about a possible impact. The pressure sensor is always read at the same camshaft angle, but the pressure waves run with (more or less) constant speed, so the sensor might get hit by a pressure wave at different crank angles with engine speed changing. I assume powerfull pressure waves are generated when the intake valve opens (from any remaining pressure in the combustion chamber), and the pressure sensor is evaluated at the end of the compression stroke, so pressure waves will be gone by then.

_________________
"It is often said that before you die your life passes before your eyes. It is in fact true. It's called living."
Terry Pratchett