Raceghost

iTrader: (4)

Hi all. Been a minute, but been busy. I decided to add this thread as to some new data I have analyzed over the last year, in my many hours of seat time, driving time, tuning time, and playing with various Injectors, Size Scaling and Latencies. This seems to shed some light, on why some of us, who think we have the same injectors as others, and try to use their scaling and latencies, seem to not be able to achieve the numbers as starting points people have posted here with the INjector Latencies Super Merged thread. Here is some info why.

I know that most people are going with High Z injectors, and this does not speak to those, for this is for the old fashioned Low Z injectors.

First we need to understand something, there are (3) known injector types, that I know of, and you should be aware that 2 injectors of the same size, of course don't exactly flow the same, but if they are of a different type, they definitely won't flow the same.

The (3) types I am talking about are:
Ball and Seat
Pintle
Lucas Disk Style

If you were able to get all (3) sets of the same size injector, i.e 1000cc, but each one is one of these three variants, all (3) would flow different.

Watch this video:

These are just examples of the various spray patterns that injectors have.
Various Data shows, that Ball and Seat design seem to be what the evo likes, and easiest to tune, they only have one spray pattern.
Lucas Disk Style, atomize better, but flow less than there variant of another type. Lucas Disk come in a 1 hole pattern, 3 hole pattern, and 5 hole pattern. ALL if were the same size, I.E. 1000cc, would flow different.
Pintle I do not know enough about, because well, I haven't had a set.

I have had (2) sets of Ball and Seat, of various sizes, and they have been easier to tune with the numbers recommended by others on this forum for Size and Latency. However, the set of Lucas Disk style I have, have been a pain in the **** to tune, but I have numbers now on them and can offer a starting point on all of them as to how to size scale and calculate a starting latency, via a mathematical equation, that seems to work.

I have also played with other cars with both Ball and Seat as well as Disk Style, and found these following statements to be mostly true, in all my data of tuning I have accumulated over the last several years.

Here is Merlin.oz 94170715 ECUFLASH XML definitions to give you an understanding of the actual calculation numbers for the EVO 8 ROM. I think you might be able to find these definitions for other ROMS as well. Notice the two first Scaling calculations and how they work. These are the actual calculations done by the ROM, which represent the lamens terms of 15~20% reduction. The first one represents actual real size of the Injector in Mitsubishi terms. SO example, the PTE1200's can be entered at 1219 or 1177. The Second Scaling is how ECUFLASH sees those numbers. So the 1219 Mitsu Scaling shows up at 1044, if you enter 1177 for Mitsu, the ECUFLASH sees 1008. The ECUFLash is the definition most of you have, that is in the ROM definitions when you download from ECUFLash.

ALL current test and data is based off of U.S. Pump 91/10% Ethanol blend.

If you have a Ball and Seat style injector, then I have found like I said, that the threads that have Size scaling and latency listed, seem to be close, and of course you have to fine tune.

With these style of injectors, we have often used this equation for size scaling:
Take the rated size, and try a 15/20% decrease in size and plug that into the injector size scaling tables of ECU FLash.

With Lucas Disk Style, I have found that size scaling has to be this equation:
Take rated size, and try a 25/30% decrease in size.

So I tested this theory, and borrowed a set of PTE 1200 Ball and Seat Injectors, and I have a set of PTE 1200 Disk Style injectors.

Changing nothing but Injector, I found the comparison I just listed to be true.

Ball and Seat PTE 1200's on my car, show that in this thread(https://www.evolutionm.net/forums/ec...00-cc-pte.html) that the Size Scaling and Latencies offered by John Bradly or Aaron of English Racing, hold true and will get you close.

The Disk Style Injectors on my car, are Size Scaled at 860/886, and a calculated latency based off of stock latencies.

This might be of some value to those of you who struggle to tune various types of injectors you may come across.

If your struggling with Latency starting points, here is a basic calculation based off of stock injector sizing and latency.

Stock injectors on the EVO 8/9 have been determined to be 600CC injectors, scaled to 513. This is how we determined the old thought of 15/20% size scaling decrease for ECU Flash.

Latencies are given in the stock map.

If your having trouble with latencies, do a calculation based off of size scaling of the actual size of the injector.

Here is what I have come up with.

Take the stock Injector Size, 600cc. You know have 1000cc injectors.
Get the ratio of percentage of how much larger, 1000's are to 600's.
So take 1000 and divide by 600.
So 1000/600 is 66% larger of injector.
Now take the 14 volt latency of the stock injector, which is .432
Take .432 multiplied by 66%, and you now have the 14V latency for the 1000CC injectors.
So in this case, .432 x 1.66 is .717, you would enter this for the value of the 14v latency of the 1000cc injector.

Now your not done, how do you get the other latencies.

Well, take your stock latency table, highlight the entire table, and increase all values in that table at the same time, until the 14v column reads roughly .717 in it, and then the other latencies are now elevated, and if you look at the graph, you know have the latencies, and they follow stock calculated curving.

This gets you close with both styles of injectors, Ball and Seat, and or Disk Style.
Car should idle and start, and then watch trims. Before playing with trims, use the standard new way of tuning injectors, by going on a freeway cruise, give it a little boost, and then drive for approx 20 minutes. Then pull over, look at things, but don't make adjustments yet.

Drive to a place where you can do a 3rd gear pull, watching AFR's and see if your lean or rich. OF course practice safe methods by letting off if you are more than .5 lean. At this point, pull over and make adjustments to size scaling only. To richen or lean out, depending. Get size scaling done first, since boost doesn't care about latencies that much. Then once your there, now play with Idle and Cruise, since they depend on latencies, and make further adjustments based on that. Combine this method with the other methods discussed, and you will be where you need to be in no time.

One other thing, is the Small Pulse Width. With LOW Z injectors, of High CC, 1000 and above, require more Small Pulse Width. LOW Z injectors require less and actually go pretty linear. If your interested in tuning that, please see MFREDS thread on Small Pulse Width --> https://www.evolutionm.net/forums/ec...ion-table.html I have Small Pulse Width Data on PTE 1200's both styles, FIC BlueMax 1250's, 1350's, and 1450's. All 3 FIC injectors use the same Latencies for which I have from FIC, and Size scaling as well. Surprisingly, with the FIC's, there latencies are actually pretty spot on. Small adjustments per ride of course, but pretty spot on. I find that my calculation works better for the PTE 1200's, rather than the info I got from PTE on there suggested latencies.

Hope this helps you all. It has been a staple for me to get a baseline for both of these injectors and latency values. Of course tuning, usually changes them a little, but were in the ball park with these calculations.

Good Luck.

sandman121383

iTrader: (15)

i'll be in touch with some info on this! surprised nobody has chimed in with any new info!

AlexEvo26

Bring this back up, that’s really good info (besides numerous other threads).
I’ll try to discuss some points here, sharing info is the real strengh.

First, just want to thanks Merlin for his « bible guide », and all of you digging for years and this code and sharing info.

Well, Now just sharing my thoughs and (quite small) experience even if I’m playing with this stuff since years now.
About the method, I’ve adapted along the years while reading and figured out new information on this forum.
So, before I mess with any kind of ign/fuel/boost table modifications
1- I first update the « injector flow linearization PW adder » table (thx mrfred https://www.evolutionm.net/forums/ec...ion-table.html, and https://www.evolutionm.net/forums/ec...ing-notes.html ). In my case, I always choose to buy injectors for which the info is available. I don’t really get any « method » to refine those numbers (but will be glad to read about it !)
2- Tune injector scaling then latencies. I use already shared values as starting point (ex https://www.evolutionm.net/forums/ec...upermerge.html ...), and I always need minimum tweaking around that.
In this phase I use « method 1 » from Merlin’s guide. I Flash a fuel basemap I’ve made with 11.5 AFR all around for load 120 and above. The left part is gently smoothed to ramp the values. I first concentrate on inj scaling logging AFR, AFRMAP and LT trims, then on latencies with LT and ST trims. I also record battery voltage to have a better match between trims and left Column of latencies.
3- once I’ve done my best with injectors related stuff, I refine MAF settings (which only need small adj IMO). At this point I’ve seen two methods, one let the MAF scaling as it is and tweak MAF comp table, the other set MAF comp table to a flat value and tweak MAF scaling. I don’t really know which method makes more sense from a « code » point of view, and I’m not enough good about code and disassembling to figure it out.

this makes, from my opinion, a very good starting point to deal with fueling.

please feel free to comment and share your experience. Once again I don’t pretend to be experienced, I’m trying to confront my thoughs to improve my understanding and method.

Biggiesacks

This is my take on when it would be appropriate to play with the linearization tuning knob if you will. Disclaimer I just like to tinker around with this stuff (obsessively), I'm not a pro tuner or anything.

Linearization is a very fine adjustment, it comes into play only at the very bottom of the IPW spectrum and Injector scaling and latencies need to be as close as you can get them before even considering messing with these. Since both scaling and latencies are global settings, in that they effect the entire fueling strategy they need to be as good as you can get them over the broadest range, scaling being more important in the higher IPW domain and latency in the smaller, but still a balance of both across the whole range. Once you have those down you can start really dialing in closed loop at different MAF Hz to try and tighten up those LTFTs. When logging maf hz and IPW you will notice that at lower frequencies IPW might be the same or even lower at higher Hz then lower. This probably has to do with engine VE etc. If you notice that at 2 different Maf hz you are logging basically the same IPW and they are both wrong (LTFT adding or subtracting, but in the same way) then this might be a good time to try adjusting the linearization table. The linearization table only effects small pulse widths (these PW's are pre-latency) so if your out of that range then adjusting the MAF table would be more appropriate, or maybe even latency comp depending, So to sum that up, if you are trying to adjust fueling within the range of the linearization table, and your fueling error seems to be more IPW dependent then Maf Hz dependent than it might be more appropriate to try adjusting the linearization table to deal with that error.

I'de love to be corrected on my thinking about this from anyone who knows better, this is just the way I have come to understand it after a lot of trial and error.

AlexEvo26

Thanks for your answer and input, helps me thinking in a slightly different way.
since the linearization table was made « available », I started re doing a couple a things (to get more accurate but also for « trial and errors » purpose.)
I consider linearization table a little bit differently from injector and MAF scaling, since this value should be a solid basement given by the manufacturer whereas inj scaling and latencies (but latencies are also provided with injectors..) and MAF scaling are more « tweaks » to be adjusted by users to keep the ecu behavior as consistent as possible.
then I understand with your post that playing with linearization values may not really make sense (or maybe think about it at the very end as you said) because manufacturer values should be pretty good and car adjustement could be done via our other « tools ».

Biggiesacks

Having that data from the MFG is truly an asset and makes for a great jumping off point. The reason I say jumping off point is because that data is created under a very specific set of conditions. When it comes to using the product in your environment often times there are many extraneous variables that cause the performance of the product to deviate from the manufacturers values. Things like fuel pressure, voltage at the injectors, or even Rds On of the transistors in the ecu etc. can effect the performance. That's where the tuning comes in to bridge that gap and achieve the real world results you are looking for. The important part is knowing what *tools* you have available to you, and when its appropriate to use each one. This is where I believe you really just have to roll up your sleeves and get your hands dirty to build up that knowledge/intuition. It's why good tuners are worth the money, because it takes countless hours to really get a good handle on everything that's going on under the hood. I'm not claiming to have succeeded in doing that or anything, its just the goal I'm always working towards.

/rant

AlexEvo26

fully agree ! I’ll keeping posting when I’ll be able to make more « trials »

Noobi

Interesting, thanks for sharing.