Tuning ITBs - AN or SD Only

Home > Projects > Megasquirt > Tuning ITBs
Up

 

This page describes my attempts to use either SD only or AN only tuning with ITBs.  Neither approach worked all that well and I describe each approaches short comings.

SD only tuning

My first attempt at tuning was to use SD only tuning, relying on the MAP signal to determine how much fuel to provide for a given RPM.  My reasoning was that in the low to mid RPM area of operation, the 15% to 20% MAP range was greater than the 5% to 10% throttle range that I would be using.  Since I wanted the best fuel control possible below about 3500 RPM this seemed like a good choice.  I configured my VE table based on recommendations on the MS forums so that over half the table was dedicated to the MAP region above 80%.  In fact, I split the MAP range from 92% to 100% into 5 rows of 2% increments.

This method worked ok at lower RPMs and small throttle openings.  The mixture was maintained fairly well under these conditions.

Over about half throttle or high RPMs this method was not ideal.  In the upper RPM portion of the table, I had more than 10% changes in fuel required from just a 2% change in MAP.  This made the tune incredibly sensitive to noise on the MAP signal.  As a result, the VE table would bounce around several bins creating inconsistent fueling and mixture.

Here is the last VE table I used for SD only tuning.  This was run on the 45mm ITBs with my 2.0L engine.

 

VE table for SD only tuning.  Notice that 2/3 of the table is dedicated to the MAP region of greater than 80% with almost all VE increase occurring only in the upper 15% of the MAP range.

 

A 3D view of the same VE table.  This view really shows the nature of the ITBs relative to MAP.  You are trying to tune an engine using only 15% of the available MAP signal.  Notice the steep slope of the curve where large fuel changes are made over just a 2% change in MAP signal.
 

MAP signal noise with ITBs

I learned a lot about the MAP signal on ITBs while trying to make the SD only tune work.  The collecting of the individual vacuum pulses from each TB creates a pretty stable MAP signal but some operating regions are smoother than others. 

The combination of throttle position and engine RPM effect the amount of noise on the MAP signal. 

  • The more open the throttle, the more noisy the MAP signal.
  • The higher the RPM, the smoother the MAP signal.

This means that the MAP signal gets noisy in exactly the areas of the SD only table where the slope is the steepest and most sensitive to MAP noise.  The area above half throttle and below about 3000 RPM was the noisiest.

You can attempt to correct this by increasing the filtering on the MAP signal within Megasquirt, or by adding vacuum reservoirs but these techniques reduce the MAP signal responsiveness and cause problems with acceleration enrichment because the MAP signal changes lag to far.

AN only tuning
The 2.3L engine was originally tuned using AN tuning.  This method uses the throttle position to determine fuel, MAP is not used.  AN tuning works well with ITBs except for the low RPM, low throttle position region.  In this area AN tuning has difficulty due to the large changes in fuel required from very small throttle position changes.  AN tuning works well for race engines that don't need to operate in this lower left portion of the table.  This specific car needed to run well across the entire range and was experiencing problems maintaining a stable mixture in this light load area.

 

VE table for AN only tuning.  Notice that this table is set up almost the opposite of the SD only table.  The largest portion of the table is used for the smallest throttle positions.

 

 

The 3D view shows that AN tuning has the same kind of sensitivity problem but in a different part of the table.  Very small throttle movements create large changes in fuel at low RPMs near closed throttle.  This is right where you operate when driving through a town at 30MPH in 4th gear.

 

 

 Top

 

email me

This site was last updated 05/09/10