XMLWRITE

[xswrex]

I've been using Ecuedit, it works like a charm. XMLwrite created a darn good XML. It doesn't include the injector scaling factor conversion to CC (because it isn't in the rom that way), but the map editing is really good.

it should have the 2450000/[value], are you sure?
Can you create again the XML with your rom and let me know? If there is a problem i should fix it!

I have created xml's for the following roms and the scaling was included

A2ZJ500J (04 us sti)
A2ZJ710J (04 us sti)
A2ZJB10J (05 us sti)
A2ZJ500I (04 forester XT with sti injectors)
these had 456cc

A8DH200Z (06 eu sti )
this had 498cc

[gregsachs]

I've been using Ecuedit, it works like a charm. XMLwrite created a darn good XML. It doesn't include the injector scaling factor conversion to CC (because it isn't in the rom that way), but the map editing is really good.

it should have the 2450000/[value], are you sure?
Can you create again the XML with your rom and let me know? If there is a problem i should fix it!

I have created xml's for the following roms and the scaling was included

A2ZJ500J (04 us sti)
A2ZJ710J (04 us sti)
A2ZJB10J (05 us sti)
A2ZJ500I (04 forester XT with sti injectors)
these had 456cc

A8DH200Z (06 eu sti )
this had 498cc

usdm sti should be 550cc injectors, per everyone in the industry. Freon's test with perrin 820cc injectors suggests the constant should be ~29546000 to make this field in the units we'd like. (see, i knew it was wrong, I was just wrong about how it was wrong;-)
stock value: 5372, flow is ~550cc, scales to 456
perrin: needed hex(for 14:7 idle): 3602.941, flow is ~820, needed to enter 680
Using those points i calculated the 2.9m value.

Oh, additional note/idea for xmlwrite:

Would it be possible to duplicate the temperature and speed calcs to english units? I know I'm more able to quickly think that way.
like so:
<scaling name="DATA-13-Deg C" units="Deg C" toexpr="x" frexpr="x" format="%.0f" min="0" max="0" inc="256" storagetype="float" endian="little"/>
<scaling name="DATA-13-Deg F" units="Deg f" toexpr="9*x/5+32" frexpr="(x-32)*5/9" format="%.0f" min="0" max="0" inc="256" storagetype="float" endian="little"/>

[Freon]

Yes, it works out perfectly if I change the scaling to assume the stockers are 550cc. That would put my current 680cc value at 820, which jives with Perrin's quoted 816cc

~2950000/[value] is probably correct.

[Hitokiri]

The stock injectors are not 550cc. Just like WRX injectors are not 420cc as everyone once thought. The big issue here on injector size is that 565cc or whatever the quoted size for STI injectors is, is rated at higher than standard fuel pressure. they might be 480 at 43psi and 565 at 55psi.

The other thing you must consider when playing with other than stock injectors is the injector latency. There is a compensation table in the ecu for the time it takes the injector to respond and open. Depending on the internals of the injector itself this time varies from .3ms to more than 1.5ms in some cases. At idle the car might be sending ~2ms signal to the injector. If in the extreme case it takes 1.5ms to open, then you are only getting .5ms of fuel rather than 2ms. Thats why the latency needs to be corrected for. If you change the make of your injectors (WRX and STI top feeds have the same internals and thus the same latency) you will need to figure out what the latency of the new injectors is. If not, you can forget about all your injector size calculations.

This is the reason that none of those UTEC cars with other manufacturers injectors never idle right. The scaling needed at idle will always be different then higher rpm simply becuase UTEC doesn't account for injector latency.


so those Perrin injectors might be 820cc, but I garantee the latency is different. So in order to truely be able to compare the injector scaling numbers and the conversion with the stockers you must correct the latency.


And please. grab ecuexplorer and log the fuel trims at idle. When you get the scaling and latency times right, the fuel trims get very low, and the car will never have any trouble idling because it doesn't have to correct for anything. you did it already.

[Freon]

The Perrins are modified stock STI injectors. The latency should be the same, as there is no change to moving mass, the coil, etc. That's precisely why I went with this solution rather than other aftermarket injectors. I honestly can't tell the difference now in idle quality. All they do is mill off the little screen that has the nine holes it in, leaving pintle needle visible. It probably degrades spray pattern a bit, but that's a discussion for another thread. I slapped 680 in the injector scaling and I honestly can't tell any difference in the way the car drives.

See here:
http://freon.shackspace.com/car/injecto ... jector.JPG

Either way, increasing the inverse value by 50% seems about right for Perrin modified STI injectors. Either the stockers are 456 and Perrins ~680, or the stockers are 550 and the Perrins ~820. Take your pick. I do remember seeing someone post results back on their Perrins and seeing ~805cc. I'll see if I can find some relevent threads on NASIOC.

565cc is the quoted spec for STI pink top feed injectors, what all the stage 4 people run. I have heard 440, 480, and 550 for the USDM STI yellow-top sidefeeds, and now with the 2450000/[value] equation it solves to 456cc. Not sure how much arguing it is worth. I know +50% is about dead on for my injectors.

[gregsachs]

The stock injectors are not 550cc. Just like WRX injectors are not 420cc as everyone once thought. The big issue here on injector size is that 565cc or whatever the quoted size for STI injectors is, is rated at higher than standard fuel pressure. they might be 480 at 43psi and 565 at 55psi.
(snippage)
And please. grab ecuexplorer and log the fuel trims at idle. When you get the scaling and latency times right, the fuel trims get very low, and the car will never have any trouble idling because it doesn't have to correct for anything. you did it already.

I guess my thoughts on this are that the scaling factor is for _our_ convenience, and is (somewhat) arbitrary. Both values will work fine for the usdm sti. However, everyone shorthands the injector flow rating to 550, etc. So, lets get a constant that gets the ecu number to be useful for us, and allows us to enter an easy, mindless number. 2950000 is probably not exactly it, but it is probably pretty close. At least ball park to get the ecu to do fuel trimming. The 2450000 doesn't do that. So, since it is pretty arbitrary, lets change it. Freon, you have a pm.

[JRSCCivic98]

There are xml issues with ecuFlash xml files being created with xmlwrite. Will they be fixed in the next version? It has to do with the width of the cels in various maps.... the values are unreadable because the min and max values are not defined in the xml files. So basically 10.25 would look like 10... you cannot make out the rest. Fix please.

[jpmarotta]

The stock injectors are not 550cc. Just like WRX injectors are not 420cc as everyone once thought....

Hey All-

FYI, just dialed in my STi pinks yesterday, scaling came in at 485. Have an ECUTek'd ECU on hand flashed for the STi, scale value is set for 485.5.

-Jason

[JRSCCivic98]

Should be 478.

[xswrex]

There are xml issues with ecuFlash xml files being created with xmlwrite. Will they be fixed in the next version? It has to do with the width of the cels in various maps.... the values are unreadable because the min and max values are not defined in the xml files. So basically 10.25 would look like 10... you cannot make out the rest. Fix please.

JR this was answered before

"min & max attributes are not included in the output XML, i thought this would be done automatically when you opened up a map like the other editors.
There is really no way of doing this properly unless XMLWRITE reads and evaluates every map's min/max values...."

Easy way is to use a text editor, mark the scalings lines
find................ />
replace.......... [1 space]min="-9999" max="9999"/>
and then the numbers would show fine

This should do the trick, but now the gradient coloring will not work

Colby - i am sure this can be done dynamically when you open a map. I suppose you do that already for the gradient colors!

[xswrex]

Should be 478.

No, it should be the number that makes the car run properly, that is A/F and fuel trims

[JRSCCivic98]

There are xml issues with ecuFlash xml files being created with xmlwrite. Will they be fixed in the next version? It has to do with the width of the cels in various maps.... the values are unreadable because the min and max values are not defined in the xml files. So basically 10.25 would look like 10... you cannot make out the rest. Fix please.

JR this was answered before

"min & max attributes are not included in the output XML, i thought this would be done automatically when you opened up a map like the other editors.
There is really no way of doing this properly unless XMLWRITE reads and evaluates every map's min/max values...."

Easy way is to use a text editor, mark the scalings lines
find................ />
replace.......... [1 space]min="-9999" max="9999"/>
and then the numbers would show fine

This should do the trick, but now the gradient coloring will not work

Colby - i am sure this can be done dynamically when you open a map. I suppose you do that already for the gradient colors!

I'm going by Colby's input below in another thread... so, we know there's an issue... the question is... who's going to fix it first?

Regarding min/max, consider it a difference in philosophy. Soon we should get to a point where we have decent metadata from the various ROMs were are looking at. At that point:

1) I don't like to see the cell sizes of my tables changing based on what is in them. It makes for ugly UI if they change while you are editing. They should just be the right size to start with.

2) The problem you are seeing really only comes up when you use XMLWrite, because it sets the min/max to 0, which is incorrect. I will add code to detect this case and size the cells with a different strategy.

3) Similarly to (1), I don't like to see my color scales autoscale to the data. The color scaling in EcuFlash uses the min/max values. An AFR of 12 always looks the same color when using the same colormap. If you autoscale, the color keeps changing, and you could even have ridiculous values in the map and not notice.

4) The min/max serves a purpose. It helps indicate what the functional range of values are beyond simply the maximum range of the datatype (the two can be radically different in some cases).

[JRSCCivic98]

^^^ I've been helping out a few other people with generating xmlwrite files for ecuFlash to use on their cars. They obviously ran into these same issues. We need a fix on this... ecuFlash xml files are basically useless comming out of xmlwrite. You might as well remove the capability if that's the case. I know I sound like I'm bitching, but you know what they say.... the sqeaky wheel gets the greese.

[xswrex]

^^^ I've been helping out a few other people with generating xmlwrite files for ecuFlash to use on their cars. They obviously ran into these same issues. We need a fix on this... ecuFlash xml files are basically useless comming out of xmlwrite. You might as well remove the capability if that's the case. I know I sound like I'm bitching, but you know what they say.... the sqeaky wheel gets the greese.

So you have been helping and it is still useless

[JRSCCivic98]

^^^ Don't be a f***ing smartass.

By useless I mean that the guy can't read any of the cels properly because your cel scaling is off so all you see is 8..., etc. He was having issues even getting the xml file to generate, that's what I helped him out with. Obviously nothing difficult, but some people have problems with even that. Anyway, I'm just letting you know AGAIN that you need to fix this if you want people to use the ecuFlash xml def capabilities of xmlwrite.