The CCM problem differs from the Husky, on the CCM the motor does not suffer from flame-outs, but does have a violent surge at "technical riding" speeds ( 3000 - 4500 rpm) when hot. When the surging starts if one is careful with the throttle the lcutch can be pulled in and the engine will cycle between 3000 and 4500 by itself with about 1-2 seconds between the peaks. It will continue to do that until something chnages, i.e open or close the throttle.
Adding more ( a lot more!) fuel in the lower ranges of the CCM has improved that a lot, but there is still a bit of a stutter. I have added upwards of 30% more fuel in places. I was initially concerned, but when I cross-checked with the PCV maps intended for the CCM I saw that they too add about the same fuel in those places.
One of the biggest problems is getting to see the relationship between the F and L map points, and when the switchover between these maps occurs.These are some things I have learnt whilst "playing" around with TuneECU on my CCM, and based on a lot of work by the forum user Ama. on the ADVRider forum. There is a really great ( and sadly overly long) thread on ADVRider titled "KTM 690 E & TuneECU"
http://advrider.com/index.php?threads/ktm-690-e-tuneecu.685657/. The KTM 690 uses the same ECU as our Husky/BMW/CCM, but has a second ECU to control the main butterfly. There is second set maps for that, but it cn be set to follow the twistgrip exactly. Much of what I write below is based on my understanding of that thread and my own experiences first on my KTM 690, and then on my CCM. Let me stress now that I am not a FI, nor tuning, expert. I bow to the better knowledge of others and welcome theor input to learn from them. I was (un)fortunate enough to be stuck in a boring yet well paid IT job that allowed me to indulge in my toys; and gave me hours of internet time to read articles whilst waiting for test routines to run or for senior personnel to actually make a decision...
There are 3 tables in a map set, the F (or F2 when Power Plug is fitted) map uses TPS and RPM to determine how long the injectors should stay open. Injector flow rate and fuel pressure are fixed, so the amount of fuel is directly related to the opening time.
The CCM maps have less RPM rows than the Husky or BMW. In the mid-rpm range the CCM tables do not have any rows for 3,250 , 3750 or 4250, but I believe the other two do.
The L map ( or L2), uses the Absolute Manifold Pressure Sensor on the throttle body to measure the air pressure inside the throttle. It then uses the table with pressure and RPM to determine the injector opening time.
The I map ( or I2) uses the TPS and revs to define the ignition timing. As the air/fuel mixture moves to very lean it will burn much slower, so the ignition may be advanced to get a more complete burn out of it before the exhaust valve opens. On the CCM the I map is very different to the I2 map. I think that this is where the CCM departs from the Husky/BMW camp. IIRC the latter use the same table for both I and I2? As I have been adding mor fuel, which will then burn faster, I decided to copy the I2 table in its entirety to the ! table.
The L map is generally more useful at lower engine speeds, as revs rise small changes in RPM alters the pressure less so the ECU uses the F map at higher engine speeds.
There is a table that determines at what point the ECU will switch from using F map to L map. At a few RPM positions the table defines at what TPS position the switchover occurs. E.g. At 5000 it may say "18%". That means that if the engine is running at 5000 RPM and the throttle is open only 5% (a trailing throttle on down hill section) the L map will be in operation. The ECU will take the MAP value and the revs and look up in the L table how much fuel to add. If we open that throttle to 20% open, the ECU will switch to the F map.
On my CCM, after adding fuel in the lower ranges, and even taking some away in the mid range, I found that the surging had diminished, but not gone. When the engine got hot enough to engage the fan it could be back almost as bad as ever. I then found comments from Ama. in the ADVRider thread relating to the use of the L map. He believed that he had found that the L map only comes into play when the engine has attained full temperature.
Adding more fuel had resulted in a cooler running engine, and earlier experimentation with the F-L switch had seemed to show that it had little effect. I revisited that but with a view to getting the motor hot first. My findings seemed to match what Ama. had said, the F-L switch only comes into play somewhere above 82degC. Without first hand information on how a specific TPS/RPM combination can be converted to a pressure value at the same RPM, I feel that my changes to the F-map have resulted in big differences when stepping between the 2 tables. I need access to more information to align the two tables, which is just not possible in my current circumstances. I decided to push that switchover out of the way by setting the values really low, that reduced the remaining surges further. I am comfortable with this, my Sherco 450 has lovely smooth power and torque curves from an FI system that has no manifold pressure sensor at all. The Sherco system uses RPM, air temp, atmospheric pressure and TPS only - as do many other systems. The demands of emmissions regulations require engines running to very tight specifications...
Looking at the 3-D chart posted by Padowan (neat!) I can see that there is a big step in the 3000 - 4500 region, any big changes are going to be noticeable. During the time that I have been looking at the CCM maps (and the PCV trims for those maps) I cannot help but think that soemone has been focussed only on the areas that they felt were important, mostly WOT or near-WOT behaviour, and meeting the specific targets of the EURO-III testing. I do not mean to belittle that work, I know that it must have taken a long time; but I do wish that they had been able to work on the areas outside of that.
You may wonder why I did not use the Power Plug and work on the F2, L2 and I2 maps? Well, one of the things that the power plug does is to disable the O2 sensor. "No bad thing" you may think, but the idea of the CCM is to replace my KTM 690 with RR tanks, i.e. a longer distance riding machine where fuel economy has a degree of importance.
The O2 sensor is generally ignored by the ECU unless the engine is at full temperature, the O2 sensor itself is at full temperature and the engine is running at a fixed speed with little perceptible variation in RPM or TPS. At this point, the AFR is measured and the mixture adjusted. The sensor itself is a "narrow band" type, the output swings rapidly either side of the ideal point. Small changes in AFR result in a huge swing in output, it is effectively an "on-off" type switch. The ECU can only really know that the mixture is leaner than that point, or richer. It can then either add, or subtract, from the injector opening time until it crosses that point again. I am guessing here, my assumption is that the ECU tries to work out how much to fine tune the opening time to attain stoich (AFR 14.7:1) , then reduces it by some percentage to try to get it somewhere around 15.4:1 for best economy and minimum CO2. As a feeback system it would need to keep cycling through that test and adjustment. Keeping the sensor also helps me feel that I have not entirely abandoned the thought that the atmosphere can be saved ;-)
The CCM surging seems, to me at least, to be a result of the lean mixture, advanced ignition curve and the stepped secondary throttle behaviour. Sorting out the first made the second more noticeable. Fixing that highlighted the differnce between F and L maps,moving that out of the way has left the steps in the secondary. That was the last mod that I have made, to smooth out the secondary, with gentler transition to open, and a minimum value of 3% above 2000 RPM ...
