Welcome to CPGNation.com! Log in or Sign up to interact with the CPGNation community.

XFI street (retrofit on dual EZ EFI 1.0) Where to start with tune

Discussion in 'FAST Support Forum' started by Narler, Aug 15, 2019.

  1. Narler

    Narler Member

    Joined:
    Jul 17, 2013
    Location:
    Australia
    Hi,

    So I have finally received my XFI Street ECU after a large fight with FedEX who wanted to charge me twice for importing this item into Australia and now I've finally got some time to fit it. Thought I would start a new thread about the XFI rather than adding it as a post to my previous thread as this is a whole new topic.

    I read through all the help (html) files included with CCom WP software a few times and read through fastmanefi website and provided (.doc) files for xfi 2.0 (thanks for providing those), but my initial startup shows some unexpected output so I'm not sure what is the most optimum place to start adjusting.

    After going through every option in CCom and setting what I think is needed (except fuel maps), I started the car and as expected it sounded horrible and would not idle at all. Yes I know the car is cold but that is a good place to start because it will always be cold on start, so no point trying to drive it yet (imho) until I can get that sorted.

    I set closed loop and chose Speed Density Mode. See attached what I see from the logging and where the ECU thinks the load is on the engine. I've stopped the graph at engine 800 RPM and TPS is 1.2% but the ECU thinks it is under 82% load? Target AF is 13.31 but actual AF is 8.88

    Clearly my low idle vacuum is causing the ECU problems.

    I think I'll be needing to set the closed loop / learning to start at say 1,500rpm and then use manual settings below that so it doesn't have such a problem.

    So at this stage I think I should ignore the learning ability and just see if I can get this to idle (when cold).

    Should I set the VE vs RPM & Load table from say 700 - 1000 rpm through the 10% - 100% range to say 40? I look at the example 3D maps on fastmanefi website and I see values in that same range from 30 - 50.

    Or should I be looking at something else first. I read through the documentation about Alpha-N mode and based on what that said I think I may have to run that instead, however what confused me a little is when it says that to run in Alpha-N mode, I have to reposition my Map sensor upstream of the throttle blades so it can read ambient air pressure?
    Is that necessary if I intent to use the self learning at RPM above 1,500 (or whatever rpm anyone can suggest would be more optimal) and just use manual tuning below that.

    I set the mode to Alpha-N and then started the car and logged and now on the tables the green oval stays constantly along the 100% load row (load in the log lists as a constant 99%) and target and actual AF look the same as the attached.

    Also just on the logging I included Fuel lbs/hour on the dash as you can see it appears as an item on the log except it is always 0, but watching it live (during logging) it was showing 32 lbs/hr. I wonder why it did not log those values.

    Two other questions.
    Fuel Calculation Parameters. I'm using the standard injectors that were supplied with EZ-EFI 1.0 (back in 2012/2013). The manual for the EZ-EFI says they are 88lbs/hr at 60psi. I'm running dual (2 x throttle bodies) and I have my fuel pressure set to 43psi. so based on what the help said I have set my injector flowrate lbs/hr to 63 using this calculation 88 * (43/60). That ok?

    Setting Idle RPM. That was just a case of entering the target RPM on the hand controller. Now I see no option for that. Is the only way to set the idle now is via the Idle Speed vs Coolant Temperature and maybe also the AE Fuel vs Coolant Temperature chart?

    Thanks for any advice.
     

    Attached Files:

    Last edited: Aug 15, 2019
  2. Fastmanefi

    Fastmanefi Well-Known Member

    Joined:
    Jun 4, 2011
    Location:
    Sonora, CA
  3. Narler

    Narler Member

    Joined:
    Jul 17, 2013
    Location:
    Australia
    Hi Fastmanefi,

    Thanks for the reply and the offer.

    The largest problem I have is meeting your high speed internet requirement. I have no cable or fiber services available where I live, so I'm limited to ADSL2 internet speed. That means I only have 1mbps upload speed.

    I haven't had a chance to try adjusting the VE table yet, but hopefully will have time to work on the car this weekend.

    Is it normal to see the "football" so high up on the load axis in the logging when at idle, or is that just a side effect of having a low vacuum engine? Based on that it would appear I would end up with a large dead / unused section in that bottom left corner of the VE table.

    I guess I'm a little concerned that seeing it so high at idle, will the system be able to successfully learn or will I have to switch to Alpha-N, or will I have to set the learning to start at a specific RPM and then manually set all the idle VE values.

    I can see just from that first log as I increase the RPM just slightly the load drops (from 82% down to 58%).

    Thanks.
     
    Last edited: Aug 23, 2019
  4. Fastmanefi

    Fastmanefi Well-Known Member

    Joined:
    Jun 4, 2011
    Location:
    Sonora, CA
    When I remote tune a low vacuum engine I first rescale the load axis. You want the football to end up about in the middle of the vertical range.
    Turn off LEARN at idle and high vacuum conditions.
    Call me to duscuss tuning approaches.
     
  5. Narler

    Narler Member

    Joined:
    Jul 17, 2013
    Location:
    Australia
    Hi Fastmanefi,

    Thanks for the reply and the offer to discuss tuning ideas over the phone. I've not worked out the time difference to you yet as I'm in Australia, but I want to see what I can achieve at the moment and I want to see how the engine is responding so I have a better idea of what is what and therefore can be more informed if we do get around to having a discussion. The more I can understand as to how the engine is operating and at least get a reasonable baseline the more productive any discussion will be.

    Thanks for the tip. I have re-scaled the axis now and I've had a chance to drive the car. It ran much better than I was expecting, infact it was pretty excellent considering I've not adjusted the A/F table yet but as you can imagine I'm now left with more than a few questions.

    The largest question is how do you know when you have the correct A/F target. I have not made many adjustments to the A/F table yet as I've been focusing on the VE table. I also can't seem to get the system to log the Fuel lbs/hr (L or H). I select it, I see it being displayed on the logging screen and showing values, but the log itself always shows 0 when I play it back (as you can see from any of my attachments). Can others log this value or maybe I have too many values being logged and it is ignoring the last ones added? Even though the engine seems to be running fine I still feel it's using way too much fuel. Is Fuel lbs/hr the best way to indicate that, or use injector DC %? Is there a normal amount that you'd expect it to use at idle etc.?

    Do you just keep leaning out each section until the engine starts to have issues or you feel a drop in power? What is the best approach. This isn't a race car so I'm looking on the economy side (certainly can't be worse than my last fuel tank using Ez-EFI 1 ECU of 5.7MPG (US) !!! and that was primarily just cruising too), but I don't want it to lag if I do want to get moving quickly. (every mans dream I guess)

    Next question. You can see from the attachment on a 20 minute drive that consists mainly of slow / light / constant throttle usage and a few medium throttle applications up to only 4000rpm (engine rev limit is set to 6,500rpm) to see how much of the VE table is in use should I stick with Speed Density Mode or switch to Alpha-N Mode. From what I understand I should run in closed loop as much as possible and then see what the O2 correction value is and then adjust the VE table? However as the load is not accurate due to low vacuum I can't see what the best strategy would be for each cell. Example:

    engine RPM 1650 TPS 0.8 target A/F 14.13 actual 12.13 O2 Corr -20.3 load 47%

    So on the VE table that is cell 5 up and 5 right and bordering on the 6th cell. But now on that exact same cell region (currently has a value of 43)

    engine RPM 1700 TPS 3.1 target A/F 14.25 actual 15.13 O2 Corr 3.9 load 54%

    So same cell region under slightly different throttle load yields two totally different O2 correction values. How on earth do you know which one you should be adjusting for especially when both are only extremely light throttle usage?

    Also is there any correlation to the VE value vs O2 correction value. eg. if O2 correction is -20 and VE is 43, if I change that to 42 how much change in O2 would that see under exact same driving condition would that be a 50% drop to say -10 or maybe only see it drop to say -15? As I have to adjust the VE table after driving I don't know how much I need to adjust it by, however this doesn't take into consideration my above concern that this cell is giving me large negative and light positive O2 correction values for different TPS.

    This making me think I should stop using Speed Density and use Alpha-N which is more RPM vs TPS so that each cell is more relevant to actual usage rather than this erratic load value from the MAP. Anyone have thoughts on this?

    Next question is starting when warm. There seems to be a large increase in RPM up to 1,600 (3 seconds after starting) then slowly dropping back to idle after about 12 seconds and there is no input from TPS (remains at 0.0). I see a large increase in AFT correction highest reading is 28.1% What do I adjust to reduce that because at that time Target A/F is 14.13 but actual A/F is 9.75.

    Still on the topic of idle I do not seem to be able to get the idle down lower or smooth. I've adjusted the Idle speed vs coolant temp and from 182 degF the idle should be 760, yet it is often bouncing between 700 to 900 when out of gear. I suspect that might be due to cam size. IAC target is around 5 - 7. Is there anything else I should check ? Once in gear engine rpm jumps between 650rpm and 750rpm. When at 650rpm IAC target is 13. Load is listed at 86% target A/F is 13.69 actual A/F is 13.13. I think idle is running too rich (VE table for that cell is set at 36). Even at 650rpm there is no signs that the engine is struggling at all which also makes me think it's running too rich. Oh I should mention having the car idle in the garage and it throws black stains out both exhausts for a good meter or so, making me think it's definitely idling too rich.

    Thanks for any input / suggestions or comments.
     

    Attached Files:

  6. Fastmanefi

    Fastmanefi Well-Known Member

    Joined:
    Jun 4, 2011
    Location:
    Sonora, CA
    Narler
    I can't begin to try and teach XFI tuning from a forum - so I won't try. But a few comments first. If someone else wants to jump in - more power to ya.

    * Don't datalog any more channels than you need. The old ECU chips are restrictive and very memory limited. If I didn't know better I'd guess they had vacuum tubes in there. When you reach the capacity some channels will not log.

    * NEVER use Alpha-N (Note NEVER is in caps) on a street engine. If you think cells mean different things at different times with Speed Density what would you say in Alpha-N which is TPS position vs RPM. Don't do it!

    * RPM goes up on a hot start because your Start IAC Position is too high at those higher temps. There's nothing more to it than that. Engines should rev like that to quickly burn off left over cranking fuel.

    * How to find the best Target AF ratios? Do the same as you would with a carbenator (spelling intended). Turn off CL and manually adjust the fuel table lean until you start losing vacuum, and then do the same rich, and then pick the happy place in the middle. It's not rocket science. For cruise run as lean as you can as long as it doesn't hesitate or stutter. For WOT leave it at 12.8.

    * Learning. The system never runs on one cell - it interpolates a group, yet learn will change only one cell at a time. Over time it will build a fuel table that hits the Target AF as best it can. Yes the CL correction will vary depending on the direction in to, and out of, a cell. Just ignore it. Customers with stick shift transmissions really don't like this. So - first get a base VE (fuel) table to run well, then smooth out the areas you don't like the looks of, and then decrease the learn gain in those areas.

    * ALL AE TUNING IS MANUAL. CL and Learn can do nothing about that. So if you are puffing our black smoke on a throttle hit - your AE tables are too rich. Fix it by trail and error. That's the only way to do it.

    * Timing has a huge affect on all this but the Street system cannot control that. So some things have to be assumed you have correct in advance.

    Good luck.
     
    Last edited: Aug 27, 2019
  7. fabr

    fabr Member

    Joined:
    Jan 23, 2017
    *Banned**Banned**Banned**Banned*!!! The best info I have ever read in such a short and to the point of the basics of tuning the thing. I wish I had read that a LONG time ago when I was completely clueless on what to do. I made a LOT of mistakes (and still do!).
     
  8. fabr

    fabr Member

    Joined:
    Jan 23, 2017
    Richard ,am I correct to assume you ,in general,consider a street car as any engine that is not being used in drag racing. No matter how radical a combination,if actually being driven with widely varying rpm/loads it is a "street" car? My question is referring to an application such as Baja/desert type running with 7-800 HP .
     
  9. Fastmanefi

    Fastmanefi Well-Known Member

    Joined:
    Jun 4, 2011
    Location:
    Sonora, CA
    Not quite. A "street engine" to me means:

    A) It has full exhaust, so the O2 sensor can be trusted further down in the RPM range.
    B) It runs on gasoline or E85.
    C) The customer cares about how it idles - in gear.
    D) Water cooled block.

    I'm okay with radical cams, turbos, blowers, nitrous, etc. on a "street engine". I have many customers running over 1500 HP one their street engines.
     
  10. fabr

    fabr Member

    Joined:
    Jan 23, 2017
    Concerning full exhaust, my O2 sensor is about 30-36 inches from the exhaust outlet right where the tubes merge into the collector. Not open headers but the exhaust is slightly less than 4 foot after the collector with tailpipes same ID as the collectors. I don't think there is any reversion reaching the O2 sensor as it idles great in or out of gear. Is there any real possibility of reversion reaching the O2 around 3500-5000 rpm? Cam is fairly big with .800 lift and 108LSA. I have been chasing a lean condition in that area under mild to hard acceleration that I haven't been able to get rid of. (I guess I should start a new thread and not hijack this thread.)
     
    Last edited: Aug 28, 2019
  11. Fastmanefi

    Fastmanefi Well-Known Member

    Joined:
    Jun 4, 2011
    Location:
    Sonora, CA
    Sounds like what I call "camshaft chop". It's a duration and LSA issue not related to lift. It happens at around 2500 - 3000 RPM when the engine is under low load conditions. Something that helps is to change the timing in that range only.
     
  12. J.K.

    J.K. Member

    Joined:
    Aug 7, 2016
    Richard,
    That was some fantastic reading. I dont want to hijack this thread either but I just had to throw two questions at you. One in regards to turning off closed loop and manually tuning, is that for idle and cruise or just idle? For a bigger cam would you leave closed loop off for "idle" or would you turn it back on once done manually tuning?
    One other question is in regards to gain adjustment. I have a manual shift car and often find a lean stumble upon reapplying the throttle after the 1-2 gear change at anything below 2700 rpm. I did lower the gain number in the second and third row a touch, seemed to help a little. Would gain tuning be the correct way to tune out this gear shift stumble?
     
  13. Fastmanefi

    Fastmanefi Well-Known Member

    Joined:
    Jun 4, 2011
    Location:
    Sonora, CA
    Hijack has been completed ... LOL...

    I leave CL on at idle IF (and that's a big if) you can trust the O2 at those lower RPMs. However, most tuners turn LEARN off at idle. Once the VEs are set in the idle area, no tuner whats the system to change them. I turn Learn on >1300 RPM.

    High gain can cause Learn to incorrectly lean-out those areas because all engines are rich after the release of the throttle and high gain can incorrectly decrease the numbers in those areas trying to get back to the Target AFR when it should just let it go. So, the cure is simple, First use Learn in all the steady state conditions to get a good base VE table. Temporarily turn CL off and set the AE tables manually so there's no stumble. Then change the LIMIT TABLE, not the gain table, in those areas, although decreasing the gain in those areas won't hurt anything if the base tune is good to start with.. Set the negative limit in those areas to ZERO. Turn CL back on. That way you have full learn in most areas and if the O2 sees a rich condition after a shift it can't do anything about it.
     
    J.K. likes this.
  14. J.K.

    J.K. Member

    Joined:
    Aug 7, 2016
    Yes....sorry for the hijack Narler! LOL!!!
    Started following this post and found it to be super informative that I just had to jump in.
    Richard those were very helpful tips. Much much much appreciated. :)
     
  15. fabr

    fabr Member

    Joined:
    Jan 23, 2017
    Would you explain that a bit further?
     
  16. Narler

    Narler Member

    Joined:
    Jul 17, 2013
    Location:
    Australia
    Howdy,

    Thanks for the replies and hijacking is totally fine. It's all information and been interesting reading as some of the posts have been on things I've not asked about yet. So all good, no complaints from me.

    In regards to blowing "smoke" or leaving black powdery stains on the garage floor. I just idled the engine for about 2 minutes in the garage today after totally cleaning the floor, and there's spotted black lines behind both exhausts and that's with no throttle hit, just starting the car and leave it idle for 2 minutes. I currently have set the idle in the A/F Table to 14.3 but engine was not warm. Is it just running so rich due to having dual TBI? I guess I'll continue to keep leaning it out and see how it idles.

    Timing. I put the car on a dyno to set the timing. I'm also running a MSD 6AL2 digital / programmable ignition box (MSD6530), so I can adjust the retard curve on my laptop. Distributor is Mallory Dual Point (but it has been converted and only has 1 set of points).

    I'll continue to adjust and log and see how it goes. Oh I did a quick run today and after dropping some items off the logging, it is now logging fuel lb/hr H and L, so yes it seems that there is only so many channels it can log which I think is quite ironic because if those channels are displaying values on the laptop (dash) then it would appear to be a CCom WP software issue as to why it is not logging / saving the data from all the channels. If it was an ECU issue I'd expect those channels to show no "live" values, so if values are being displayed then it's CCom that is not saving them correctly.

    Thanks.
     
  17. J.K.

    J.K. Member

    Joined:
    Aug 7, 2016
    Richard,
    For finding idle AFR you were saying to adjust the VE or Target AFR table?
    My example my target AFR set to 13.5 with VE value around 28. I played with that VE value and found lowering the value to 20 actually gave me the lowest map reading. Lowest map reading I'm taking it means the same as highest vacume (don't have an actual vacume gauge). So lowering the VE to 20 gave me lowest map, any lower and it went in the opposite direction and started stumbling. This gave me an actual AFR of 14.0.
    (Quick note: this is NOT in closed loop).
    If I go back to the AFR target table and change it to 14 so it matches actual....then my actual changes to a much leaner number and would require raising of the VE numbers again to bring it inline with the new target. Doing so however just brought my map numbers back up where I started. So my question is, is this a case of where the O2 sensor cant be trusted? This is a 248/248 @ .50, 110 lobe separation camshaft if that helps any. Is this a case where I might have to ignore the fact that target and actual dont match, but rather shoot for the lowest map?

    Also for learning....someone once said manual shift cars should only learn or tune in a middle gear due to learn constantly tuning cells that are shared with other gear loads. Do you share that opinion or would you be good with having learn on for all steady state driving even in a manual car?
     
  18. Fastmanefi

    Fastmanefi Well-Known Member

    Joined:
    Jun 4, 2011
    Location:
    Sonora, CA
    If you trusted the O2 to get the ideal 14:1 with CL off, then it's the same with CL on. You might have a CL offset in effect that you don't know about.

    The negative CL limits should be zero (or at least a small number) around the shift points. I leave Learn on everywhere above idle (I never let Learn change my idle settings) but I limit its control with lower gain and the CL limits I just mentioned.
     

Share This Page