alpha-N (TPS) vs MAP load source

In reply to mke :

Getting a little off topic here, but I'm curious if that should apply to other sensors as well. For things like most fluid temperature and pressure sensors it's not a big deal. I don't need to be precise to within 1 deg F or psig on my oil temperature However, I do notice that using Haltech's calibration file for my (GM) 1 bar MAP sensor gives me an indicated 0.5 psi MAP with the engine off. I did some digging and for some reason Haltech's calibration file is slightly different than what I've found elsewhere for the GM sensor. Haltech uses a simple linear relationship from -14.5 psig/0V to +0.5 psig/4.72V. According to "manufacturer data" 0 psig/101.3 kPa should be about 4.91V. I need to go check what my sensor is showing for raw voltage but for something calibrated at SL I'd expect mine to show slight vacuum to reflect the slightly lower atmospheric pressure. In this case would I just shift the calibration up/down to zero the reference point? I suppose in the end it doesn't really matter as long I'm being consistent.

There are couple pieces there.  First, sure, you want the most accurate data possible so if you have data for your senors that is different from the canned cal curve there is not harm in updating the curve.

Part 2 is exact reading from sensors only matters on production cars where 1 tune needs to work on every car.  For you, the tune is for your specific car with the sensors you have and if a sensor is off its corrected for in the tune.  The tuning table are really error tables, if they were anything else you wouldn't need them.  So if you read MAP -0.5 but its really say 0 then you're tuning point that sets fuel for the point -0.5 is really 0....and its tuned right.

Last  just looked thought the ESP software and was honestly a bit disappointed in what I saw.  I looked at both the ESP and NSP and they seem very similar....it looks like while it will let you run both MAP and BARO corrections I'm not sure there is any point as the best I can tell the baro is just a less good version of the MAP correction.  That made me sad.

In reply to mke :

I got the car up and running on the new ECU over the weekend. Aside from having to change the trigger offset for some reason everything went smoothly. The onboard barometric sensor reads 14.7 psi. In reality it should be a little lower but at least I know the baseline and can adjust accordingly. Using the "manufacturer's" MAP sensor calibration caused it to read -0.4 psi at atmospheric, which is lower than I expected. Ideally it would be -0.1 psi if the sensor was nominally calibrated at SL. Since my whole objective so far is just to get the car running as it was before I'm going to leave it with the Haltech calibration. If/when I switch to alpha-N load based tuning I'll play with zeroing the MAP sensor calibration.

From what I understand the ESP and NSP software is functionally the same. There's some GUI differences but no real capability differences when using an Elite ECU. The only benefit of the NSP software is faster loading, connecting, downloading, etc. times. So far ESP isn't bothering me too much so I don't know if I'll upgrade. I'm always wary of being told things are "automatically" translated. I agree with you regarding barometric vs MAP correction. Seems like they should be doing the same thing. Maybe MAP is more for boosted applications where pressure will be significantly higher than atmospheric whereas you wouldn't expect to see much in the way of higher barometric pressures.

In reply to infernosg :

congrat!

So looping back around. The car is up and running on the new ECU as well as it was before. However, I really want to try out alpha-N again since most seem to agree it's better for my setup. Is there a reliable/safe way to convert from one to another? I'm not so concerned about fuel since safety can generally be guaranteed with closed loop O2 control and some caution. Not so much for ignition advance and other load-based functions. I've tried using my histogram to match RPM, MAP and TP steady state points and convert the tables but I get some pretty wonky curves. Think weird undulations when generally advance should increase with engine speed and decrease with load. Am I overthinking this? Would it be better to just try roughly approximate TP to MAP even though the relationship isn't constant over a range of RPM?

In reply to infernosg :

I would suggest reviewing logs and try to make the base maps for there....the more steady statre points you have the better, thos will transfer directly.

In reply to mke :

Yeah, that's probably best. I started to look into it. In addition to steady state points I also looked at slower transient conditions. Say, a slow acceleration at constant MAP/TP from 2000-5000 RPM. With this I was able to come up with some rough MAP->TP points based on things like starting and ending ignition advance. The weird undulation in advance is technically something I'm already living with assuming I'm accelerating at a constant TP. This keeps getting back to the fact my engine pulls less vacuum at low RPM/idle than it does at higher speeds. I imagine this is true of all engines to a degree but probably not as severe as what I see: -6 psi/-12 in. Hg at idle (1000 RPM) but it'll be -10 psi/-20 in. Hg at light cruising conditions. The base ignition table I started with was for a MAP-based turbo car so it likely doesn't account for this. From what I can tell most street driving, outside of acceleration, is generally between 5-15% throttle, which encompasses half of my MAP table. By 30% TP I'm already through 3/4 of the MAP table. It's looking like my initial TP-based tables will essentially by the MAP-based ones, but squished to the left with little change above 30-40%.

infernosg said:

From what I can tell most street driving, outside of acceleration, is generally between 5-15% throttle, which encompasses half of my MAP table. By 30% TP I'm already through 3/4 of the MAP table. It's looking like my initial TP-based tables will essentially by the MAP-based ones, but squished to the left with little change above 30-40%.

Right!  That is why alpha-n is so nice on track cars, TONS or resolution in the hihger power ranges  and why its more challenging on the street.  What I do is use very non-linear TPS axis points, like 0, 2, 5, 8, 12, 15, 20, 25, 40....like that depending on the setup.  I'd try to put at least 10 points idle to cruise, i know their are a lot of schools of though on table sizes but I've never been happy with a tune using less than 16 and alpha-n needs the most because its inherently non-linear along both axes .   I've never tuned a rotary so maybe they are different but I'd expect similar behavior.