Understanding Power Enrich logic

Sorry for the long post but I wanted to include as much detail I could to hopefully make getting help from forum members as easy as possible.

I need some help with understanding the PE logic in the factory naturally aspirated USDM Subaru ecu's. They are made by Unisia Jecs and I believe they are a 32 bit ECU, mine in particular being for a USDM 2002 Subaru Forester.

The most relevant parts of the build consist of a used USDM factory built sti shortblock mated to the original SOHC Subaru ej251 heads, a rotated gt3076, 2014 sti top mount, and ID1000s along with a full bugeye WRX suspension and 5 speed.

I'm tuning with software that is essentially open source Romraider but for the na ecus called Project Lambda Tuning.

The car is factory speed density with a ve table and a few different Fuel - Power Mode tables (and many more) that seem to operate inconsistently.

Those consist of the following:

Power Mode - Initial Fuel Ratio

-Final Fuel Ratio

-Blend Delay

-Enable Delay

-Throttle Enable

-Blend Step, Engine Speed Index

-Blend Step

The biggest problem I'm having is understanding how to enter Power Mode everytime it needs to as well as what the Blend steps are doing and to what.

With the Blend and Enable Delay modes set as low as possible and the Throttle Enable (rpm vs tps%) set where I want it to switch from closed loop/stoich afr targets to open loop and a WOT afr target it doesn't always do as it should.

Usually if starting off in 3rd gear and rolling into the throttle it will go quickly to Initial Fuel mode (IFM), then into Final Fuel Mode (FFM) and follow the FFM target closely to redline. This is how I believe it should work.

Then most of the time if I continue accelerating into 4th gear and WOT it will quickly go to IFM, then either never go into FFM or very slowly go there in which case it never targets my FFM lambda target of .75 more less actually gets there.

Other times it works like I think it should thru both/all gears.

My Throttle Enable settings should make it enter Power Mode at anything > 31% tps and >3200 rpm.

If anyone doesn't mind to help I would sure appreciate it. The software for Project Lambda Tuning is available for free on their website and I'm using the ECU type: AG407.

I can provide my tune and datalogs as requested along with anything else I can do to make it easier to help but I think having the actual software will make it the easiest.

Thanks!

I'm not familiar with that ECU or software, but I would look for closed to open loop delays, minimum enrichment for open loop transition, confirm your closed and open loop fuel targets at the RPM/load area you want this transition to occur are rich enough to overcome any minimum enrichment requirements for power enrichment as a starting point. The tables you listed sound straight forward so there are likely other tables involved if you aren't getting consistent behavior.

Hi Mike I was hoping you'd see this. Unfortunately I have either done what you've mentioned or I don't understand the tables or what they effect.

Like for example the blend step. I have an A, B, & C referencing engine speeds which I get that part but the parameter is just a percentage and in stock form is set to .2, .3, & .4%.

What is that a percentage of? I don't see any noticable increase in fueling that corresponds to those %'s and rpms anywhere when looking at a log but probably because of such small numbers. Would that be a % change of pulse width or ve or something else?

Another one is the blend and Enable delays. Are they in Ms?

With them set to zero for both and the throttle enable set to >30% tps and >3200 rpm and I can be at 99.6% tps and 4k rpms and it still not start targeting the power mode final fuel ratio by the time I'm to redline.

If there are other tables involved that aren't defined I assume I'm just SOL with this ECU and/or software? Maybe I could reach out to the developer to see if there are other tables they could define?

I'm not the best with computers but I have I've invested a lot of time, effort, and money to learning what I have and feel like I have a pretty good tune on the car. It just sucks I can't seem to get the last little bit figured out especially since I don't know if it's my own understanding or just simply not possible in the software.

Sorry, I've been supporting a race car at Pikes Peak all week.

When someone reverse engineers an ECU, their understanding of calibration comes into play. For example an RPM axis on a main fuel table may be easy to scale and display as sane values, but other tables like you're describing may simply be scaled and displayed incorrectly, units may be wrong as well.

Increasing a value will likely still cause an increase, but the values themselves may be incorrect and that may be causing your confusion. For example, maybe 0.2% is really 20%, or 30% may really be 1.65.

When a reverse engineer finds a raw 32 bit value, then they guess what the unit might be based on what they think the table might do, then guess how to scale it into something sane, but it's often not obvious.

Yes if the tables you need aren't defined, unless you can convince the developer to find them, or find them yourself, you may not be able to adjust what you're hoping for. I would exhaust your options with the tables that you have before giving up though.

Will do. Thanks again Mike and congratulations on your success at Pikes Peak!

Thank you!