EFI University

[Johnny_9]

I don't have access to a steady state dyno, so I would like comments on tuning ignition curves with reference to VE curves. I have overlaid my tune for a WOT dyno run

Any comments on the shape of ign curve vs torque?

-as rpm increase, increase advance
-as VE increase reduce advance
-as swirl increase reduce advance

I got access to log knock and afr; and maybe later EFT, no steady state dyno in my area!


As the torque curve drops at 4200-4300 rpm, should ignition be bumped up a bit? c.f the upward slope of ignition at 6200rpm as torque drops

Assuming the AFR ratios are equal - would the EGTs across the torque curve be the same?

Just doing some part-time on the road tuning (nothing serious)

[MrDomino]

Replot that with a y-axis that goes to 0. That way the torque won't be fluctuating all over the place.
_________________
I am a Mechanical Engineer.

[Johnny_9]

So what about street tuning? can anything be done on the street? The answer is yes, you can get the low load ign in the ballpark on the street but you'll need a buddy to drive while you tune on a level stretch of road. Here's how: select a gear and a speed to get you into the desired RPM column. For example 3rd gear 50MPH then monitor intake air and record how much intake air is required to achieve this speed. Have driver target the 50MPH speed then increase ignition by 2 deg and speed will most likely increase slightly if it does not increase ign again by 2 deg. When speed increases above the target have driver back off on throttle till 50mpg is achieved again. Then record new air flow and repeat till you find sweet spot. The idea here is a little different than on the dyno, here you are finding the least amount of intake air required to achieve a given speed. Once you understand this process simply select different speeds and different gears to target different columns in the PT IGN table. You won't be able to tune the PT hi-load section of the table this way but you can certainly tune most if not all of the lo-load sections of the table. Don't try to force tune a section of the table using this procedure, simply target a speed and a gear to achieve an RPM then let the chips fall as they may. Meaning just tune whatever cell the car wishes to be in for the given condition, then you can simply interpolate the cells you may not have hit.

[TurboNova]

[Johnny_9]

Timing is generally the inverse of the VE map.

So once we plot the engines VE map/surface via AFR mapping, then invert it, it would give a guide to touching up the shape ignition map surface at different loads.

Without a load dyno and that data in-hand, a good strategy would be to advance the whole map and do WOT and part-throttle pulls, can see where power was gained and lost and adjust accordingly should come pretty close to a steady state ignition tune.

The use formulas to go a rough regression on (delta VE/rpm vs timing) at different loads.

On the same engine, say change the cam-timing, tune the afr, plot the inverse of the VE, set timing based on the above regression, and your are more or less done

Views?

Should be able to do this in excel?

or better in the tuning software could do this.....

[coz]

[Johnny_9]

The post is meant to address alternatives to those WITHOUT access to load dyno As stated in above. The parameters are one single engine, with AFR tuned and a VE map derived thereon.

The key word here is regression co-efficients on a surface function (you may have missed that) on a proven VE map with inputs of differential torque from part throttle pulls

[Johnny_9]

[Johnny_9]

Once the data is imported into the toolbox and the model structure decided, the two-stage model is fitted automatically and becomes available for analysis. For this particular engine response, the data typically has two distinct regions on either side of the maximum recorded torque. This shape of response is hard to fit with a single polynomial type curve unless a high order is used. A polynomial spline is selected from the available model types to fit this behavior. This curve type has two polynomials joined in a continuous way at a 'knot' and allows many fewer coefficients to be used to fit data that has two or more distinct regions.

Tools are provided for automatically or manually identifying outliers, (anomalous data points). The plot shows a fit to a local spark sweep (lower graph) with one point automatically flagged as an outlier based on the statistical measure of studentized residuals.

The lower plot shows a good curve fit using the polynomial spline. The upper plot shows the very high residuals for the last point, explaining why this point has been automatically identified as an outlier (red circle). Click on image to see enlarged view.

[Johnny_9]

...the concept is "analytical calibration"

http://www.mathworks.com/company/newsletters/news_notes/oct02/statistical.html

[Johnny_9]

This concept of analytical calibration (to supplement - not supplant) time consuming dyno tuning is gaining ground in OEM calibration development - not saying it gives a 100% spot-on answer, but it cant be dismissed out of hand either.

For a DIY tuner at least it would get the ignition is the right shape (if there was no access to a load dyno

The question posed here

1 Single engine at single ECT and AIT & fuel octane
2 Fuel Curve Tuned by Wideband giving VE
3 Derive a formula for ignition as function of rpm, map and VE

Formula can be derived via regression

[TurboNova]

No matter how you calc it the result will be just a guess at best. There have been several engines I have tuned that have had the exact same parts in them and they netted a different tune to make them run correct. All parts the same you would expect the same result but you don't get it..... that is why all the mail order tuners don't ever get the tune correct ever.

For timing to be something based on your VE table, the problem is junk in = junk out.... so if you didn't tune the car on a steady state dyno then your VE table isn't correct either, so how can you base another table on info that isn't right to start with?

For all the time you have spent thinking of a "better way" to do this you could have actually tuned a fleet of cars the proper way.
_________________
Brian Macy
EFI University Instructor
Horsepower Connection
www.horsepowerconnection.com

[Roberto Arano]

[quote="TurboNova"]No matter how you calc it the result will be just a guess at best. There have been several engines I have tuned that have had the exact same parts in them and they netted a different tune to make them run correct. All parts the same you would expect the same result but you don't get it..... that is why all the mail order tuners don't ever get the tune correct ever.

For timing to be something based on your VE table, the problem is junk in = junk out.... so if you didn't tune the car on a steady state dyno then your VE table isn't correct either, so how can you base another table on info that isn't right to start with?

For all the time you have spent thinking of a "better way" to do this you could have actually tuned a fleet of cars the proper way.[/quote

Regarding VE estimation (or calculation of), I believe you need good data ,yes (actual AFR , Intake AIR density ,MAP etc , actual fuel flow) , but this data CAN be obtained without a dyno at all...? (NOT talking about estimating timing from it here.)
_________________
www.circuitse7en.net
http://stores.ebay.com/Circuit-Se7en

[TurboNova]

[quote="Roberto Arano"]

[MrDomino]

You can actually get a pretty good approximation over the engine map provided that your VE model is good enough.
_________________
I am a Mechanical Engineer.

[Roberto Arano]

I think there is no need to measure the power/torque output of the engine to calculate VE.You need to datalog actual AFR vs fuel flow. So a dyno is not specifically needed for a VE measeurement. Fast enough datalogging is needed vs the rate of change of the data being measured. Achievable on the road with most datalogging systems I think if you don't try to do it in first gear going downhill (hyperbole their lol).
_________________
www.circuitse7en.net
http://stores.ebay.com/Circuit-Se7en

[Johnny_9]

Wideband lag is 0.2-0.3 secs - the way round that is holding the engine and data-logging AFR at constant rpm and map for 2-3 secs (each) at 2000 rpm to 6000 rpm in step 500/1000 rpm increments on and open deserted road.

Since VE/MAP is linear at each rpm for loads bounding approx. 50%-100%, IMO you can get a pretty good interpolation of the AFR/VE map.

If you look at a ignition map tuned on a steady state dyno, at each specific rpm, at which the VE/Map is linear at that specific rpm, the spacing of the the ignition point at that rpm band is inversely proportion to the delta MAP. e.g

3000 rpm - VE/Map (m) = a1m + y
4000 rpm - VE/Map (m) = a2m + y2

i.e evenly spaced assuming tuned at same AFR (for which high load sites aren't in the final tune, but that can be adjusted for)

Admittedly won't be exact, but close and could be time saver (on the dyno) - i.e. analytic calibration of timing may supplement the dyno process

the analytics can be done a hack level (excel) or mathlab etc. won't be exact, but close - for a NA the 99% of MBT is about a 2 degree margin anyway?

[Johnny_9]

http://delphi.com/pdf/techpapers/2010-01-0158.pdf

[MrDomino]

[Johnny_9]

http://redhardsupra.blogspot.com/2009/06/parametric-ve-explained.html