ECMlink Locking in Open Loop, wideband switch point

[Calan]

The screen shot at upper left shows max lambda of 1.36 @ 5v, but your wideband properties in Link are set to a max lambda of 1.29 @ 4.67v. That is ok as long as your 4.67 value falls on the line, but I don't see a transfer function to check it against so I don't know. You should probably change the Link properties to match just to be safe.

BTW - Why did you change the output curve from the factory one?

 

[LiquidX]

I changed it earlier. The curve is so I have a narrower sweep within values I would actually use and see. No sense in having a less accurate sweep by having a larger range.

 

[Calan]

I changed it earlier.

Gotcha

The curve is so I have a narrower sweep within values I would actually use and see. No sense in having a less accurate sweep by having a larger range.

Makes perfect sense.


*********

How is the car running now? Any new logs to look at?

 

[LiquidX]

It's running excellent at idle right now. I'll get a log a little later, playing some xbox

 

[Calan]

Ok.

 

[LiquidX]

Just for you Craig, I went and made a log

Overall I'm pretty happy with the results. For the VE table, I'm only getting half points of adjustment in 1 or 2 cells from where it's sitting now. I'd rather leave them alone otherwise I'd have to go and adjust the Injector battery table some more and you know how fun that is...

 

[Calan]

Nothing jumps out at me in that log. Looks like it's time for some light cruise calibration.

BTW - You should set the display min/max properties for AFRatioEst to match your WB, so they are scaled the same and superimposed over each other in the graph. I usually use 4.8 and 18.

Same thing for boost and BoostEst. (I currently use -20 and 60)

 

[LiquidX]

How does the offset uSec work under AFRationEst? It's says it is normally 0us. Changing that value to 30us the AFREst and my wideband match up even better. I'm assuming its for a delay?

 

[Calan]

I don't know, but I've wondered that also. I've never found any info on it though.

 

[LiquidX]

Oh what's this, the master fuel tuner doesn't know this feature ???

I think I may be going back out later to re do my dead times and VE table. I think I'll upload either Sams or Coreys and go from there. Mine is too butchered up to save

 

[Calan]

Oh what's this, the master fuel tuner doesn't know this feature

Just stop it... I'm about as far from being a "master tuner" as you can get.

I think I may be going back out later to re do my dead times and VE table. I think I'll upload either Sams or Coreys and go from there. Mine is too butchered up to save

Your deadtimes still just "feel" too high to me....but if it works, it works.

Probably not a bad idea to steal Corey or Sam's table and go from there. Just set the low RPM/load values to match your current table, since it seems to be idling pretty good. Of course if you change the fuel numbers, all that goes out the window anyway.

 

[LiquidX]

I used the latency values given by FIC and my global dead time is 210uSec.

 

[Calan]

Well, either Jens gave us two different sets of numbers for the same injectors, or one of us is using the wrong numbers...and it could be me.

I'll see if I can find the original datasheet he sent me and double check it.

EDIT:

Found it. Here are the latency numbers I have:

[U]8 Volt[/u]     [U]10 Volt[/U]     [U]12 Volt[/U]     [U]14 Volt[/U]     [U]16 Volt[/U]
1750us     1560us      1350us	   1170us      980us

Using 12v as an example, FIC shows 1350us, whereas you are running 1700us according to the last log you posted.

 

[LiquidX]

I have the same numbers as well, I needed to up it due to the fans turning on and my car wanting to stall out

12v I'm at 1580+210=1700, I see how it works now

 

[Calan]

Where are you getting 210? That last log you posted shows 120 for global deadtime. Granted, I've been up all night, so maybe I'm a bit delirious.

 

[LiquidX]

Dyslexia I suppose haha. I thought I had 210, not 120. So then it would really be 1580+210=1700uSec like you said.

Maybe once my VE table is fixed my global will come down some more. I'd like to get it to 0 global and just use the injector battery table as it's more precise.

 

[Calan]

So then it would really be 1580+120=1700uSec like you said.

Fixed.

****

It's the same either way. As long as the total adds up to where it should be, it doesn't really matter what is in global and what is in the table.

People just use the global because it is meant to be an offset from the stock injectors, and is easier to adjust on the fly. When someone says something like "FIC 850's have 330us of deadtime", they are really saying "FIC 850's have 330us more deadtime than the stock injectors". So you enter 330us for the global deadtime value and life is good....at least in theory.

 

[LiquidX]

That's just it, I could understand that working on a 100% stock voltage vehicle (I can make that assumption now) and global would be fine. However when you start adding in power draining devices, like my 2 SPAL fans, it has a significant voltage drop that global won't do anything about.

The global adjusts the global scale of all voltage points and unless you single out the voltage points which we did then you will still have the problems regardless. Perhaps your 14v is dead on using global but other values are now too high because of the global addition.

Changing 1 will affect the other for what I'm trying to do anyway. I'll need to get out there and work with it some more and figure out an optimum condition. My issue is the 13v range is where I see the most fluctuation and I cannot adjust it unless both 12v and 14v get touched then it changes everything else.

Don't get me wrong, the global deadtime works great to get the injectors to a ball park, I just think it's better to get that to an area that the car will run at and then adjust the battery table to make it precise.

On top of that, after getting them dialed in right for the voltages, it would then make it SOOO much easier to determine which individual injectors need a different dead time whether it's positive or negative since not all are made equal.

 

[Calan]

Global deadtime doesn't mean anything to the ECU though... not really. The ECU is calculating pulse width based on the latency at the current voltage; it could care less if that latency is taken directly from the table or by adding something to what's in the table. The final number at each voltage is all that matters.

Here is how I came up with the numbers I used for the BM 1250s when I first set things up in Link. I took the FIC latency at 12v (1350us), and subtracted the stock value (915us). That gave me a difference of 435us, which is basically the global value. I then subtracted that 435 from the other FIC voltages (and interpolated the missing ones) to come up with the numbers to put into the table. So I still end up with FIC's latency numbers, which just happened to work for me.... but I offset them by 435us so that global deadtime would still represent an offset from the stock values.

 

[LiquidX]

Post up your .eda file so I can see a visual of what you are referring to.

 

[Calan]

Ehhh... it won't really help you much, because I've tweaked it a bit (not much.. but a bit) and the numbers won't match what I was explaining. I used your 2G numbers for my example, which are different compared to my actual values.

I'll run some numbers based on your values and post up a table that would look like what I was talking about. Give me a few minutes.

 

[LiquidX]

That's right, I keep forgetting you are on a 1g. Your values will differ due to pressure difference and all that jazz.

 

[Calan]

That's right, I keep forgetting you are on a 1g. Your values will differ due to pressure difference and all that jazz.

********

Ok...using the stock deadtimes from your log and the FIC datasheet, here is where I would start and how I would get there.

These are the raw deadtimes, covering the range in ECMLink's voltage table. The matching voltage values were plugged in right off the datasheet, and the missing ones were extrapolated based on the curve being linear...which it isn't exactly but it's close enough. So to use FIC's data, you would plug these values into the DA table (except for the ones in red, which you can't) and use a global value of 0us.

Voltage - 7     [COLOR="Red"]8[/COLOR]     9     [COLOR="red"]10[/COLOR]     12     14     16     19
Latency - 1845  [COLOR="red"]1750[/COLOR]  1655  [COLOR="red"]1560[/COLOR]   1350   1170   980    695

If you want to use a global value and have it represent an offset from the stock values (which I would recommend), then you pick a voltage point to calculate it with, and then adjust the rest of the numbers. I use 12v, because it's common to both FIC's data and ECMLink and is a logical choice. So:

1350us (FIC @ 12v) - 915us (stock @ 12v) = a difference of 435us

Subtracting that 435us from the other FIC numbers gives us an adjusted table that looks like this:

Voltage - 7     [COLOR="Red"]8[/COLOR]     9     [COLOR="red"]10[/COLOR]     12     14     16     19
Latency - 1410  [COLOR="red"]1315[/COLOR]  1220  [COLOR="red"]1125[/COLOR]   915    735	 545   260

So you could use that table ^ with a global of 435us, and get the same results as the first table with a global of 0us.

Keep in mind that:

1. Tom has shown that most injectors behave differently with our ECU drivers, and they don't always match up with manufacturer's published data.

2. Those numbers are just starting points and you may need to tweak either the global value, voltage points in the table, or both.

3. This is all based on my own approach to it, and may be completely out in left field.

I would start with the second table and a global of 435us, and then tweak the 435us as needed to get things close while operating at a normal 13.5-14v. Once global fuel, VE, and everything else looks good... then I would go back and make small changes to each voltage if necessary to compensate for fans, etc.

Remember also that other things besides just the injector latency are affecting your AFR when the voltage drops, such as a drop in RPMs, less voltage to the fuel pump, less voltage to the ignition, etc. In fact, you'll probably notice that you drop into a different VE cell due to the drop in RPMs.

 

[LiquidX]

Changing to those numbers, my VE table will increase I believe so I don't stall out. I'll still have difference values at the voltage points due to when the fans kick on and what not. I'll give them a shot however and see how it works out. working with a 0uSec global.

 

[Calan]

I wouldn't worry about the fans yet; just use the adjusted FIC numbers across the board, and then tweak the 435us global and your VE table to get AirFlowPerRev and AFRs in the ballpark. You also need to make sure your global fuel is as accurate as possible..which means using known 100% pure gas or testing whatever you have available for ethanol content.

Like I mentioned, it may be something other than the injector latency causing the AFR change. You might find that as you get the VE and your fuel dialed in, any dips in the AFR during voltage drops will even themselves out.

Based on the earlier discussion, I think you'll also find that the 435us global value will need to increase...for whatever reason on your setup.



****

50 freakin' posts, and somehow we still have our own private thread going here. .