Kevin has some other maps you can download and try. My tuner tweaked one marked original. Think the MWR one is for an Ingen intake. What engine and mods do you have?

 

Oh, yeah. Oops.


1zzfed with stock internals, EFR turbo, 700cc/min Deutschewerks injectors with 13 ms latency at 14v, 3" OD maf tube with honeycomb screen 7" upstream. My MWR map is turbo map which is supposed to be good for 630cc/min faster injectors and a 3" OD maf tube with a GT28. I paid them for it because at the time I thought it was a quality map, and they deserve to either be paid their asking price for it, or have it not be used, and I thought it would get me in the ballpark faster.

 

To auto tune fuel, you need to define these:

Min Base map (you say you have one)
Max Base map (ditto)
Target AFR map (14.7 in every cell is the default, but if there is 14.7 in every cell, the auto tune algorithm won't run because it doesn't think you've defined Target AFR)

You can display Curr(ent)-Min Base, which will show how much auto tune can change each cell downward (less fuel), and Max-Curr Base. Or maybe it is New-Min Base and Max-New Base - it has been a while and Current and New maps should be the same unless you are working offline and edited the New Base map. The point is that you can tell if auto tune will be able to change cells.

There's also a Sample/Cell map - how many samples you need in each cell in order for auto tune to make a change in each cell.

Water temp has to indicate engine warmed up.


So you are getting samples in the Logged AFR map and yet it doesn't make any changes? If all those things above are OK, have you checked the obvious, like enabling auto tune fuel (check box)?


----

Oh by the way, your best bet it is to save that MWR map you paid for and beg someone for a better one. Is the Inj map all 1's?

The latest and most helpful (step by step) How To topics on Copilot are over on mr2spyder.net/spyderweb.

I don't have a Power FC or even a Spyder and stopped all this work. Will help a little bit.

 

I have done all of those steps. I set the minimum number of samples per cell at 10. I set the inj map to 1 in the very low load cells, and set it to quickly taper down to 11.5:1 everywhere else as a safe temporary target baseline.

Some cells have minimum afr logged of 17:1, hundreds of samples, and still have no changes applied. It does retard timing every time I hit the rev limit.

I'm wondering if it is something really basic like having to click the star in the upper left of the map before logging or something like that that I missed. Also thinking maybe if I have cells where an 80% error correction would exceed the max or min, maybe that disallows any corrections?

 

The rev limit is a fuel cut, there is no timing retard on a PFC. Don't hit fuel cut on a turbo motor if you can help it, bad things happen. Once you get copilot up and running the program will retard the spark in a high knock cell by rewriting the map. Sort of a poor mans J&S if you hit bad knock.

 

Did you click the AutoTune button under copilot configure? Did you start a new autotune folder?

Is copilot logging and showing the difference between logged and targer in each cell? Can you manually make a change to a map and save?

 

Yes to all of those questions.

 

Actually, I just checked, which I created the autotune folder in copilot called test2, no such sub-folder existed in the copilot folder. Does that indicate a problem, or was i supposed to manually create the folder and then direct the program to save data there?

 

This screen capture was taken from before any folder called test2 existed on my computer....

 

Can you do a screen capture of the setup tab when car is warmed up and logging?

Can you save the map so you can reload it. Then put 1s in the min sample require to see if it copilot will change the map?

 

You will find "test2" under a folder named Autotune. That's where all your files will be. You don't have to create any folders. When you select New Auto Tune Folder from the file menu (I'm just going by memory here), it creates folders for you.

 

I noticed that you have "Min" selected way over on the right, not "Avg", for Logged AFR. Copilot can display Min and Max Logged AFR, but all it cares about is the average logged AFR in each cell.

 

There is another auto tune fuel setting that can explain it not writing. I can't remember what it is called, but it is the minimum number of cells to be changed before a write to the base map. The default value for that is 0, which had s special meaning (and serves as another way to disable writing). The reason for having this parameter is at times I wanted it to write as soon as it has good data for even one cell and at other times (usually in fact) I want it to hold off until there are several cells ready to be written. When the auto tune process does a write, there is quite a delay because it writes the entire base map and reads it back to ensure the current base map is in memory. It also updated the display (redraws the map) and queues a request to another software module to write the current and new base maps (in memory) to the files on disk. The time it takes to do all that is less than a second, but meanwhile many more AFR samples could be read and responded to. Normally I set that parameter (Min Cells to Write or whatever I named it) to 3. I hope that's it. At the very least, I explained the parameter.

 

[HR][/HR]Another reason for a potential write to the base map to not occur, one that may not be obvious, is that the number the auto tune algorithm calculated for a cell would be too radical a change. This is to guard against bad input from a wideband O2 sensor. Let's say all of a sudden the wide band goes out to lunch (I have had that happen a number of times) and Copilot sees something like 10.0 no matter what - every sample is 10.0. If not for the Min Base Map and the way it is treated, very soon Copilot would make your base map go way out to lunch. So it isn't treated as a limiting value for changes as in any calculated value for the new base map being set to the limit if it is too low, it is treated as a test for reasonableness. It will not change a cell that is outside the limit. A warning message is written to the message log and there are visual indications. Same kind of thing for the Max Base Map.

 

I noticed that you are using version 6.5.8. This could be important, too. I gradually, then completely, stopped Copilot development (upgrades and bug fixes) because it was taking up too much of my spare time and it got to the point that I did not update the main Copilot Download topic over on www.mr2spyder.net/spyderweb. Here's a much better version, the last one:

http://home.comcast.net/~kevinbeane/Copilot/Copilot_7_5_1.msi

 

You should uninstall Copilot (don't worry, your files won't be deleted), then download and install the latest one. Sometime before this latest one, I changed the way I/O Settings are done. So you will have to set up your wideband again. Here's some screen shots of the newer I/O Settings window.

The main reason for doing all that was so you could choose which interface to use for wideband input. I had a Datalogit, then switched to FC Hako (fewer analog inputs), and also had a Phidgets 8/8/8 USB device that had 8 analog inputs, so it made the I/O Settings much more flexible.

Spyderweb has all the latest documents and topics.

I apologize for not updating that link to the lastest version in the first post of the Copilot download topic on Spyderweb. I just took care of that.

 

I am brand new the the PFC (imagine my surprise when I saw myself credited on the splash screen for the first time the better part of a decade after the last time I discussed this stuff with Kevin). I am running an MWR base map (I think, not impressed with it at all).

Initial questions:

1.) The MWR base maps are not rescaled, the VVTi map is 95% full retard, MAF tube size and injector size and latency seem to be just tuned in the base map with no compensation settings (which would make load values make more sense and allow injection at the correct time, etc). Is this really an MWR base map?

MWR likes to put everything in the base map. If they use the Inj map at all during tuning, they end up wiping that out in a Recalculate Base operation (don't do that - it destroys information). They don't seem to like using the injector size correction factors, either, although I've seen MWR maps where they did use them. Very confusing. In fact, the whole reason I developed the Auto Tune function (and Copilot in general) was that I got frustrated by the fact that my Inj map was all 1.0 so what the hell does this all-in-one Base map tell me? What numbers to use?

2.) I have O2 feedback on for now to make it so I can use the car to get around before i tune it. With one of the sensor wires disconnected, it seems to work better, but of course I have only have one wideband so it is essentially taking an average value. Does the PFC do the dual bank thing that the stock computer does and I am screwing with the individual trims majorly by feeding only one signal, or is feeding only one signal actually a viable way of halving the swing amplitude of the control?

The PFC works like the stock computer in a lot of ways. It will use both O2 sensors in the header (before cat) during closed loop (O2 Feedback enabled and being used because of conditions e.g., idling or cruising and warmed up). In a Celica GT-S Power FC (2ZZ folks like I was), the Power FC only has one O2 sensor input. In an MR2 Spyder 1ZZ Power FC, you may have 3 O2 sensors, but the Power FC only has 2 inputs (for the ones before the cat). When in open loop e.g., WOT, the Power FC just uses the maps and settings to determine the fueling and that's not any different for one cylinder or bank. I don't really know about the stock ECU in open loop (does it do individual cylinder fuel adjustments or not). If you have a bad tune, definitely use O2 Feedback when not tuning because it won't run well. Note that when you go into auto tuning (start it) while O2 Feedback is ENabled, Copilot will first throw away all the samples it got previously because the AFR was influenced by O2 Feedback being enabled, then disable O2 Feedback for you, then start auto tuning. If you leave O2 Feedback disabled (like when you have a decent tune), you can go in and out of Auto Tune however you want... any samples it logged before (which are averaged in the Logged AFR map) are still there and still usable.

3.) Copilot seems awesome, but the learning curve is steep. Since I essentially need to remap every fuel cell, it would be awesome to use Autotune to get a rough map. My problem is that it does not ever analyse data, even when it has hundreds of data points per cell for some cells. The max and min maps are set, I followed the directions from a dozen different Q/A threads and the latest copilot manual, best I could understand them. Any ideas what I am missing?

Oh, it is very steep. That's why I wrote all those topics. It was also a Moving Target because I was constantly improving it. Some of those old help topics are outdated.

 

Thanks Kevin, I also am just starting to work w Copilot. Awesome work.

 

Thanks for saying that. My whole life has changed since I was doing all that work. Now I have a different job (not even developing software), I've lost a parent and the other is doing quite poorly, I am selling my house and moving to a different state, and I'm getting a divorce (although that's all very amicable). I still work on simulators, but instead of developing software (that work dried up due to government funding for an outdated aircraft) and I am just doing hardware troubleshooting/maintenance for simulators... not as challenging or interesting as software development, but hey, they pay me well enough and I'm kinda over the hill (old dog, new tricks). I got really serious about my health, having been quite sedentary for years and got a wake up call when I found out I had high blood pressure and oh by the way, a fricking tumor (but benign, thank God). That was before all the other things happened. So I pretty much stopped doing software for a hobby like flipping a light switch and got my ass on a bicycle. I've also taken up kayaking. So health-wise I am SO much better than before, but wow, what a lot of change. I've also gone back to my first love, which is painting. That's what my degree is in (never finished a Computer Science degree - didn't need to, either, never slowed me down or affected my salary I don't think) - Fine Art. If I may say something important here, I don't regret having spent all that time developing Copilot for two basic reasons: 1.) I'm obsessive and creative and this served as a positive outlet for that, 2.) I helped people. Time is precious. Never waste it. You never know how much you will be blessed with. Helping others in areas that you can is, in my opinion, a responsibility. You have to give back. I'll keep trying to help as time allows, but woah Nelly, have I been through a lot. I try not to let it show, but damn - the stress sometimes! I am just glad I'm not working nights anymore. That about killed me - no REM sleep. banmepleasebanmepleasebanmepleasebanmeplease that shit I am 56 years old.

I've always been somewhat embarrassed about how difficult Copilot is to learn, but was used to developing very complex software which was used by pilots and other officers to train flight crew and partly because it was their job to learn my user interface, they never really had much trouble. But yeah, it is complicated. Once you master Copilot, though, you will appreciate what it can do. Even if you don't tune your engine with it, nice displays. I miss having that tablet PC in my car and watching all the gauges and graphs and stuff. Now I have to - doh - pay attention to driving.

 

Not a lot of time for a response at the moment, but I wanted to say THANK YOU!

I installed the latest version. I reconfigured the wideband by manually typing in factors for my plx. As soon as I turned it on and got into the throttle it started showing cells written immediately. The changes seem much smaller than I expected, but it is working, and I can probably figure that out myself once I dig into it more (so far I only ran it for about 1 minute sitting in the parking lot at work this morning).

 

Set your min and max maps +\- 5 percent from your base map for now. If you get a cell to rewrite you are up and running, watch for the initial trend then you can manually increase or decrease the entire map manually with just a few key strokes. Autotune only tunes the cells you visit so getting the bottom left and top right portions of the map in tune may take some manual changes. Fuel autotuning takes quite a bit longer than ignition. Get the fueling map where you want it before you turn on the ignition autotune.

 

Kevin, you ARE The Man :biggthumpup::biggthumpup:

 

w00t!

 

Holy Crap, god of copilot returns! How did I miss this thread? Forced retirement in 8 months for me Kevin, hope things are going well I have not seen you in a few years.

 

Thanks, Blackcreek. You are really good at tuning and gave good advice.

Forced retirement doesn't sound that bad if you get retirement bennies. I would think you'd miss the flying, though. Maybe airline pilot?

 

Started using replay mode. AWESOME!

 

Bad things with turbo motor on fuel cut : it goes ultra lean

Replay is fun, yeah. Oh things you can do with GPS, too, although it doesn't work with Garmin GPS. Copilot is a big time-sucker. Enjoy,

 

Why is ultra lean an issue? A little lean is scary....

 

The whole point of fuel cut is to go to infinitely lean... so there is no combustion. That can give you 1-2 'dangerous' cycles at high RPM/load (incomplete injection on several cylinders), but is generally a safe (but harsh) rev limit. Cutting spark entirely is probably safer, but not emissions-friendly. Soft-cut can sometimes be neither safe nor emissions-friendly.
I have never heard of a case where fuel cut grenaded an engine, NA or turbo.

 

Ok, so the minority that think it is bad news bears think so because it gives the knock sensor fits, not because of something I missed.

 

So I've been playing around in the settings. I decided to define the knock sensor response curve such that the values from 10 rpm below my rev limit up are divided by a high number. This should make it so that any knock detected as a result of hitting the rev limit will be pretty much ignored. That should in turn stop causing non knock noise at rev limit to retard my high rpm cells, without loosing high knock retard safety on the actual last cell in the ignition timing map.

 

That sounds like it might work IIRC Copilot knock retard will lower a cell even if it was in a write protected row. Are you getting high knock readings at the rev limiter? What are your autotune spark retard settings?
If you want to use playback to tune your car off line don't turn on auto tune, simply log what you drive. If copilot auto tunes even one cell while driving it won't offline tune any cell after that in the log. When you replay the log copilot will then show you what cells it thinks need a rewrite, you can save them to the offline map then install when online.

The dyno program works pretty good, it is accurate enough to see a 2 or 3 HP difference.

Curve editor works really great for tuning the parts of the map you don't visit much.