Speed Density on the BA/BF Australian Ford Turbo 6

[rolls]

People do use Air Load as an indicator of Load, have logged the values and they are close but different. Load is from the SD tables whereas Air Load can be calculated from the MAP and temp. Paul Yaw's article actually uses Air Load to get to VE.

Where can I find paul yaws articles?

[DarrylC]

http://wpdev.injectordynamics.com/ar.../shelby-gt500/

Keep scrolling until you find the maths teacher

Also check out "slack dog Tune":

http://www.hptuners.com/forum/showth...Slack-dog-tune

and use the search engine on some of the pointers.

[IH8TOADS]

There is an accelerator pump table..... But its not called that

Tony ill help you out without giving u the answer because your a smart cookie

We use a calculation for transient fueling ,correct, its based off manifold volume correct ?

How does the ecu know what space/area/volume its working with ?

We need to tell it , correct ?

If we manipulate this input will it have a desired effect to what your asking ?

What is the accelerator pump table being referred to here?

[Romulus]

Tip-in pump shot.

[IH8TOADS]

I doubt it, that parameter isn't in HPT iirc.

[DarrylC]

Not on the BA/BF yet but on the FGs there is a Pump Shot Gain vs. ECT table under Engine -> Fuel -> Transient. Is this a different table to what you are referring?

I think OP was referring to the Manifold Volume as it was pre Eric's update.

[Romulus]

I doubt it, that parameter isn't in HPT iirc.

Can't say I looked for it in HPT, I know it's in Sniper.

[rolls]

Thought I would put these posts in here as it is good to have all the info in one sticky as it is a pain in the ass to search and try and find stuff. Took me ages to figure this out and would be way easier if it was already summarized.

On doing an SD recalc

If your AFR/LTFT is on point at the majority of MAP values (at a specific RPM and Cam angle) but then goes off at very low MAP values.

Eg

at 2500 rpm -10degree cam

8.61 inHg = 0.9 lambda (this should be the lowest map value you have measured in this cell but not an outlier, eg you should have at least a few measurements at this value)

10 inHg = 0.95 lambda

15 inHg and above = 0.99 -> 1.00 lambda

Then your MAP at zero is off. You can calculate the change required via this calc.

fuel mass in lbs = ((map- "map at zero airmass high res") / "map per airmass high res) / stoich_afr)
0.0000183059 lbs = ((8.61 - 3.3) / 19800 / 14.64)

now you want the ECU to inject 10% less fuel mass in grams eg 0.00001647531 lbs

Do some algebra

"map at zero airmass" = (stoich_afr * "fuel mass" * slope)

"map at zero airmass" = (14.64 * 0.00001647531)

"map at zero airmass" = 3.35668

voila, you change your "map zero airmass" value at 2500 rpm, -10deg cam from 3.3 to 3.357 and your AFR should be much closer to 1.00

Now you need to make sure your datalogging has LOTs of values at that cell before trusting this data, your wideband must also be calibrated correctly which unless you have a few to compare it to you don't really know this.

Also this only works if your slope is perfect, eg at high MAP values your AFR is exactly on point. You can do the same calculation for "map per airmass high res" if your lambda is off by exactly the same percentage at all map values.

Personally I would do everything in the following order:

Injector data dialled in, eg offset, low slope, breakpoint, high slope are set in that order.

Then get your MAP slope set correctly, only then should you go modifying your map at zero.

Something else to remember if you have DFCO (fuel cut) conditions met then your AFR will be off on these cells, you need to ensure DFCO is disabled if you want to optimise your MAP at zero cells.

Good luck!

edit: Like IH8TOADS said if you lock your cam angle (set all values to a certain cam angle) then you can get better data as the cam isn't constantly moving interpolating between cells.

on injector slopes:

Read this:
http://injectordynamics.com/articles...racterization/
This is also interesting though not directly related
http://injectordynamics.com/articles...-and-dipshits/

Here is the formula for the slopes:
You can plot the slopes on wolfram alpha or a graphic package, change the breakpoint/offset/high/low slop and see how the curve changes.
IF (fuelmass >= breakpoint) y (fuelmass_lbs)*=**highslope_lb_sec * (x (injector_sec) - *(breakpoint* (1/lowslope_lb_s - 1/highslope_lb_s) + offset_sec))
IF (fuelmass <*breakpoint) y (fuelmass_lbs)*=**lowslope_lb_sec * (x (injector_sec) - *offset_sec)
fuelmass_lbs = airmass_lbs*/ commanded_AFR

Where offset is your injector offset (eg how long it takes for the injector to open and actually start flowing fuel)
Breakpoint is where the two slopes change, usually this is when the injector starts to flow linearly

Now the ecu actually feeds in desired*fuel mass and then gets injector ms, eg the opposite of how the equation looks. You can rearrange the formula if you wanted to do that.

Basically the whole point of the two slopes is to approximate the flow of a non linear injector. With the standard injectors this works pretty well, with large injectors it works not bad however you will have an error as can be seen in pauls graph of injector flow vs approximate flow. This isn't a huge deal but you can see why a nice lookup table would work better, problem is the ford ECU doesn't work that way.

edit: Increasing the high slope means a steeper graph, this means you will open the injector for less time for the same requested fuel mass. Bigger injectors flow more fuel, hence you want to open them less for a given air mass. Eg you want to lean the mixture out.

[IH8TOADS]

Been a while since I read the Speed Density section. The way I have always done it, in the example above you would end up with a higher MAP@zero airmass but a lower MAP per airmass. Ie if your actual lambda was as per commanded at MAP point 2 (high MAP) but not correct for MAP point 1 (low MAP) then you end up with a new slope and offset...not just offset.

[rolls]

Been a while since I read the Speed Density section. The way I have always done it, in the example above you would end up with a higher MAP@zero airmass but a lower MAP per airmass. Ie if your actual lambda was as per commanded at MAP point 2 (high MAP) but not correct for MAP point 1 (low MAP) then you end up with a new slope and offset...not just offset.

Yes you would need to recalculate both to get it perfect, depending on the error you have correcting just the offset is good enough a lot of the time.

[mrharvey]

This was one of the things I kept trying to find on the net but no one seems to have written it up so I have done a write up on Ford Speed Density. It is probably too long and wordy and I would appreciate feedback (particularly by tuners who have worked in this area) but hope it gives some information to other people trying to understand how it works.

It is for the BA/BF turbo 6 but you can use the concepts on other Falcons. I have not included the correction factors that get applied as it complicates the description and I am not sure all the tables are available in HP Tuners.

At this stage I have left it as a pdf attachment but once it has had some feedback I can put it in the post. I had to take the tables out because you are only allowed to upload ~195kb for a pdf (yet you can upload tune files much bigger).

Feedback welcome, especially where there is further information/corrections.

Darryl.

Thanks for such great information

One quick question as i was testing out your equations (page 2), one of the calculations you have for constant difference in the amount of fuel required based on +4% fuel trim, which can justify the adjustment of map per airmass being divided by 1.04. This value states 23048.5, as i manually tested the numbers i got 24038.5. Is this me making an error or just a typo?

[DarrylC]

mrharvey, good pick up. I picked the wrong cell (1750/24000) instead of (2000/25000). So yes 25000/1.04 = 24038.46 as you correctly calculated.

[LA Automotive]

Darryl

Awesome write up mate, love your work !

Thank you

[Markfraze]

This was one of the things I kept trying to find on the net but no one seems to have written it up so I have done a write up on Ford Speed Density. It is probably too long and wordy and I would appreciate feedback (particularly by tuners who have worked in this area) but hope it gives some information to other people trying to understand how it works.

It is for the BA/BF turbo 6 but you can use the concepts on other Falcons. I have not included the correction factors that get applied as it complicates the description and I am not sure all the tables are available in HP Tuners.

At this stage I have left it as a pdf attachment but once it has had some feedback I can put it in the post. I had to take the tables out because you are only allowed to upload ~195kb for a pdf (yet you can upload tune files much bigger).

Feedback welcome, especially where there is further information/corrections.

Darryl.

Thank You

[dat111]

when tuning my ba xr6t

when logging, which histrograms is the best to use to build my VE table?