ECMlink Help with tuning SD. What next?

[justacanuck]

can you post a log with this example?

I didn't capture a log including the knock at low/part throttle, but the log attached clearly show the temporary lean condition on TPS change scenario fairly well.

I would expect BaseTipInTPSAdj (aka accelerator pump in the old days) to be the right solution to this problem, but given we see very little TPSDelta in the logs, I'm not sure that changing those settings would make any difference.

I'm not sure that STFTAdjRate changes are the best choice/direction here either. Maybe O2FeedbackAdjRate would help for ClosedLoop scenario, but again, it just doesn't seem like the right solution, given that it also happens when switching to Open Loop.

Are these large spikes in AFRatioEstimate normal? It seems to run smoother when I don't see so many of those little spikes....

 

[We're on Boost]

It seems like the software has all the features we need to make this behave better, just no documentation or samples to demonstrate how/when/when those settings should be adjusted.

Yeah, lack of documentation for the Tip-in feature. They don't even show TPSDelta on the Loggable Parameters wiki page. They do have a few words about Tip-in on the Speed Density Setup wiki page, but it's just what we've already said.

I'm not so sure the Tip-in software has what we need to really do it, because in between 0 and 1 there needs to be about 10 more divisions. Or maybe 5 would do it. Anyway that's where most of the action is. We should have a non-zero result for squirts all the way down to a TPS Pos Delta rate of about 50% per second, I would guess.
If they could change it to be like that we might have something. But it worries me that the TPS Delta raw value goes from 0 to 1 and goes up from there in whole numbers only (apparently) 0 to 8. Maybe that's all they can get, from wherever they are getting it from. Maybe they can't add any divisions between 0 and 1. Well they could if they measured and calculated it their own way, but it looks like they are extracting it from something that already exists from the OEM.

Hope you get something helpful back from them.

 

[justacanuck]

Yeah, lack of documentation for the Tip-in feature. They don't even show TPSDelta on the Loggable Parameters wiki page. They do have a few words about Tip-in on the Speed Density Setup wiki page, but it's just what we've already said.

I'm not so sure the Tip-in software has what we need to really do it, because in between 0 and 1 there needs to be about 10 more divisions. Or maybe 5 would do it. Anyway that's where most of the action is. We should have a non-zero result for squirts all the way down to a TPS Pos Delta rate of about 50% per second, I would guess.
If they could change it to be like that we might have something. But it worries me that the TPS Delta raw value goes from 0 to 1 and goes up from there in whole numbers only (apparently) 0 to 8. Maybe that's all they can get, from wherever they are getting it from. Maybe they can't add any divisions between 0 and 1. Well they could if they measured and calculated it their own way, but it looks like they are extracting it from something that already exists from the OEM.

Hope you get something helpful back from them.

It really seems like a bug in the software implementon IMO. If mashing the throttle doesn't generate a significant TPSDelta, what would?

I have seen a TPSDelta of 2-3, but only once or twice in the ~100 logs I have collected in the last several weeks.

Reading other tuning guides on this forum, although it is not specifically mentioned as a related issue, my TPS adjust is currently set to 9. I think this could possibly be part of the issue. I'm going to fix this and see if it helps us achieve some additional TPSDelta on throttle position changes.

 

[We're on Boost]

I have seen a TPSDelta of 2-3, but only once or twice in the ~100 logs I have collected in the last several weeks.

Looks like you made 24 logs on Friday alone!

As far as I can tell, the existing turbo is very likely a td05-16b. The turbo is only rated at ~40lbs/min or ~505cfm.

I think we are flowing very near max already:
~300gm/s indicated in our logs is close to 40lbs/min or ~520cfm.

If it's a 16g, and if the airflow numbers in your Feb 23-24 log are accurate, then yup I think you are about one sneeze from the edge of the compressor map.
So from that log I'm figuring 22 psi of boost. You are at about 2,300 feet of elevation? About 13.5 psia atmospheric pressure then. Umm, half a psi lost in the air filter. So compressor inlet pressure of 13.0 psi. Maybe 2psi lost in the intercooler and piping. So to have intake manifold pressure 22 psi over atmospheric, you need about 24.5 psi gain in the turbo.
13 + 24.5 = 37.5 psia at the compressor outlet.
37.5 / 13 = 2.88 That's the Pressure Ratio you need from the compressor.
37.5 - 2 = 35.5 psia in the intake manifold plenum.
35.5 - 13.5 (outside ambient air pressure) = 22 psi boost

I put a yellow dot on this map at PR of 2.88 and around 520 CFM. No doubt this is a little off due to one thing and another but whatever, I think you are close enough to the edge!
Previous user of this photo had a dot at 512 CFM and PR 2.0, just ignore that one.

You must be logged in to view this image or video.

 

[justacanuck]

I put a yellow dot on this map at PR of 2.88 and around 520 CFM. No doubt this is a little off due to one thing and another but whatever, I think you are close enough to the edge!

You must be logged in to view this image or video.

I don't really understand what this means... do I need a bigger turbo?

 

[justacanuck]

Took some more logs today. Removed some timing again to stop knock at part throttle.

 

[justacanuck]

I had to retard the timing a bit more in a few places to eliminate knock. Most of it was at fairly low RPM, e.g. the knock I complained about earlier.

Power is still really good. I could probably add a bit of timing up top, but I think I'll wait a while. I would rather keep it safe/reliable for now. I'll probably try to add a bit more timing, and possibly slightly leaner AFR, once the weather gets better.

Log attached.

 

[We're on Boost]

I don't really understand what this means... do I need a bigger turbo?

Well you are very near the maximum speed line, or at least what Mitsubishi considers maximum speed, of 145,000 turbo shaft rpm. That's determined partly by tip speed of the compressor wheel exducer and rapidly falling efficiency when you go past that line, and in some cases maybe the bearings.
You should be fine going up to that yellow dot. These Mitsu turbos can be "oversped" a little (past the max speed line). People do it. Usually they don't even know it because hardly anybody has a speed sensor mounted into their turbo.
Personally I would not ever knowingly overspeed my turbo. I like my stuff to last for a long time and not cause unnecessary damage. Oh wait, I guess we don't really know how much of your turbo is genuine Mitsu. But it seems to be going really well so far!

Putting a bigger turbo on would be cheaper and easier than doing a whole laundry list of upgrades to the engine and intercooler. The laundry list could come later. And maybe you want to keep using that intercooler. It doesn't mess up airflow to your radiator for one thing good about it.

A Forced Performance DSM 6851S aka 68HTA V3 would be a pretty easy and low cost turbo upgrade. It's designed to pretty much just swap right in where you had a DSM stock type turbo before, so you don't need to start all over with different exhaust parts, or different intercooler piping even. You would need to hook up boost control for it though, because when you buy it with internal wastegate, it comes with an 18psi spring in the actuator. So you would need to buy one of those Ingersoll Rand 3-port boost control solenoids that ECMtuning sells and hook it up, to get your boost going up over 20 psi again.
Anyway that turbo, the 68HTA V3, is rated at 51 lbs per minute and physically it's about the same size as a 16g.
You would not want to get the V2 by mistake. That version of it is also offered on a different page (and they don't say on the page that it's the V2!). The V2 has gotten a reputation for slow spool (for its size) with people saying they liked the V1 better, which doesn't seem to be offered anymore.

BTW the turbo you have now, I don't think I would change the spring in the actuator. I think it's good! Weak actuators were always the problem with internal wastegates back about 15 - 20 years ago. Even now I don't think they are standardized very well so getting a different one would be pretty much a guess. You'd probably spend a lot of time just trying to get the boost back up to where it is now all by itself with no help!

In my car, right now my fuel is almost straight gasoline. It's 17% ethanol at the moment. I've set the boost to about 24 psi. The timing map in use with that fuel mix is the "TimingMinOct" table which is pretty low timing - it's very similar to your timing. My gasoline is 92 octane. My pistons are 9:1. That fuel mix and low boost is how I use it in the winter when I don't want much ethanol in the tank.

You can also get the 68HTA V3 with a different turbine housing that is configured without a wastegate, for if you want to put an external wastegate in the O2 housing instead. You'd need a different O2 housing obviously.

 

[justacanuck]

I also really appreciate the insight into your timing table and feedback on ours. Also thank you for all that turbo information and advice. I am a novice when it comes to this DSM/4g63 stuff.

The other car will probably get the Turbonetics back that we got with it. It seems as though it is a Garett clone T04E "54 trim" clone. I believe it has Turbonetics "stage 2" compressor/assembly based on the inducer and exducer sizing. From what little data I can find, it looks like maybe it could flow about 20% more than the TD05-16g, which should be good to make just over 400hp in the stroker setup. But I don't think that quite meets my son's goal for the stroker (i.e. +500hp) so the T04E-54 might end up in this car eventually. I also have a boost controller from ECMlink already, just waiting for a good reason to want to use it but so far the TD05 is working pretty well, exactly as is.

We also have a Holset turbo that I think is ruined, but we have the housing and all the accessories to make it work so we'll be looking for a deal on a clean/unbroken unit. I expect it should be good for over 500hp.

Those brand new Force Performance turbos are reasonably priced and I hear they are really quite good. I'll look into them some more. I might have to buy a nicer car to put such a beautiful turbo into.

I need to post a picture of this beater we've been talking about/working on... you'll understand why I haven't spent much time trying to make it look nice.

 

[We're on Boost]

I need to post a picture of this beater we've been talking about/working on... you'll understand why I haven't spent much time trying to make it look nice.

Yeah I just looked through your Specs & Photos page and your build thread, looking for photos because I wanted to see what your intercooler piping looks like - and didn't find any pictures!
Or pics of your mystery turbo. Maybe it is something known, like a Kinugawa.
The pics they show now have a very fancy looking wastegate actuator that stands out very obvious, but maybe the older ones didn't have that. Their logo looks like a demon with like horns and all that demon type stuff. That might be cast into a casting somewhere.

BTW on the 68HTA, I don't mean that you'd get 51 lbs/min with it on your present engine and 93 octane pump gas. I really don't know how much more you'd get out of it than what you have going right now. But the compressor and turbine sections would both be more efficient than what you have now which translates to less heating of the intake air and better flow through the turbine section for less backpressure in the exhaust manifold. Going higher from where you are now without knock on 93 octane gets harder and harder.

 

[Steve93Talon]

I am going to post some dyno info I've saved from posts made by a guy in here, @Steve93Talon who went quite far with a 14g, using stock cams, stock head, and stock intake manifold, but it was on E85 so that makes a big difference. Anyway you can see how early in the rev range the power and torque drop off. This is the same thing I found with my 14b (which was always on gasoline).

For reference, I also made 369/403 on a 2G head and EVO3 intake with GSC S1 cams. Still drops off just as hard over 6k

You must be logged in to view this image or video.

 

[We're on Boost]

For reference, I also made 369/403 on a 2G head and EVO3 intake with GSC S1 cams. Still drops off just as hard over 6k

Hey nice, thanks for the input!
I'll add that to my Word document for my own records here on the ol' computer.

 

[justacanuck]

For reference, I also made 369/403 on a 2G head and EVO3 intake with GSC S1 cams. Still drops off just as hard over 6k

I feel like some of the drop off after 6k is not just the possibility that the turbo might fail to maintain adequate boost and/or airflow, but also that our injectors are _really_ running out of juice anywhere after about 4500rpm. The injector time simply does not stay linear after approximately 5000rpm and ~80% duty cycle. After 90% duty cycle, it appears to start exhibiting some strange behavior, and does not appear willing to move any additional amount of fuel. I guess it could also be our fuel pump running out of capacity, combined with the fact that our AFR is set fairly rich/conservative.

I would really like to put some bigger injectors in this thing and run it up to 7k rpm and see if it might make a little more power after 5500rpm, where we currently hit peak HP at ~85% injector duty cycle.

I wonder if Steve has insight into his injector duty cycle at 5-5.5k rpm?

The fact that Steve was able to get even more power, with a smaller turbo, makes me think we could be missing out on some of the power/potential of this setup.

 

[justacanuck]

Once I got my SD table sorted out a little better, it became clear I needed to take my global fuel a little lower. I subtracted 2% more from Global.

It was able to pull to almost 7k rpm without running out of Injector duty and power didn't start to drop off rapidly until nearly ~6500 now. (Pull1 attached)

I also did a little work to better calibrate the Autometer Wideband. I had noticed a small discrepancy between what the meter was logging, and what the fuel trims were doing, so I think this small discrepancy was throwing off my efforts to get the SD table nailed down better and I believe we have our global deadtime set a little more accurately now.

I made a few more small tweaks to SD table and made another pull, but this second pull we encountered a little knock near the top so I quit early. I'm really not sure why...

Is this combination so sensitive to knock that I might need to keep AFR below 10.5 at WOT to prevent knock? I am really starting to suspect this engine may have higher than standard 1G compression ratio because of it's apparent propensity to knock. Does anyone know the expected compression ratio of a 6 bolt with 2g head combination?

Now that SD and global seem quite accurate, I should be able to experiment with the AFR and timing a bit. I believe my timing it still quite conservative.

Thanks in advance for any feedback and/or advice.

 

[justacanuck]

It's been a while, so I figured I should post an update.

I had the car running well but still not satisfied with some of the log data, I continued to tweak and refine the tune. Each time I spent some time tuning, it ran just a little better and a little stronger, so I persisted in working on it and refining the tune.

I think I have the VE table close to a few percent from perfect now, and the results are apparent: It runs so much smoother now and makes a little more power. Even those weird spikes in the AFRationEst are significantly reduced now that the tune is in better shape. The MAP sensor signal would jump by >1/2psi beween samples, which I believe was making tuning effort more difficult. I turned ON some of the smoothing of the SD signal data, which probably helped.

The car demonstrates a nice/fat powerband that makes it a pleasure to drive. Boost comes on fairly quick, really just a couple seconds to go from vacuum to +16psi, and hits max boost of ~23-4psi after another second and it continues to hold +20psi past 6k rpm where power and efficiency start dropping. I also set the OpenLoop thresholds lower, so it switches to open loop as soon as we hit more than a few pounds of boost (it was holding on to open loop for quite a while at 20-30% throttle, with boost >6psi. That didn't seem like a great idea.

I advanced the ignition by 2 points again and according to the virtual dyno, the car appears to hit a peak HP of ~320HP between 5500-6k and peak torque of ~320ftlbs between 4500-5000. I am seeing +300ftlbs from ~4000rpm to almost 6k and I'm seeing values around "9" now in the HPbyAir. (Thanks Justin, that is a helpful metric/target for tuning.

I must say, I am very impressed with the 4g63 platform. I am also very happy with the current tune, and quite amazed at the performance of this little 2L motor. I am not sure that I will bother making any significant tuning changes for a while. I may take another attempt in spring at making more power, after we get the Cyclone intake working. I might try messing with the CAS and adjustable intake cam timing as well.

I want to say a big thank you to all the people on this wonderful forum who offered suggestions and provided feedback which helped me get the car running so well. I would add, that I now have a much greater appreciation for the level of effort that goes into tuning on this platform.

I have attached a log from today's effort, and a picture of the log data for anyone interested, as well as a spreadsheet I built that some might be interested in using to visualize the VE, timing, and ignition tables. (*pictures of 3d charts from the spreadsheet below.

I am looking forward to doing this all again next month, assuming we can get the stroker finished and installed in the 93FWD, along with bigger turbo, bigger injectors, and some stickier tires.

Also a couple teaser pics of that stroker.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

 

[We're on Boost]

It's lookin' good!

Does anyone know the expected compression ratio of a 6 bolt with 2g head combination?

Not sure but I don't think it makes any difference.
I noticed that STM gives the "head cc's" on the pages where they show Wiseco pistons for the 6-bolt and the 7-bolt engines. It's in the charts they show that look like they came right out of the Wiseco catalog. They are down the page a bit. They don't say otherwise so they must be figuring 6-bolt heads on the 6-bolt and 7-bolt heads on the 7-bolt.
Anyway they show 47 cc's for both. So the combustion chamber volume according to that, is the same.

Wiseco 6-Bolt DSM 1400HD Pistons (88mm Stroke for 150mm Rods)

Wiseco 7-Bolt DSM 1400HD Pistons (88mm Stroke for 150mm Rods)


I'm a little confused by one thing in your last log, the "...V10" log. In the first second of it, it is in closed loop and your real O2 (Front O2) is cycling like it should, so the AFR should be about stoic there, but the Wideband shows 10.1. That's very odd. Can't see what came before though. It's just that one second.
Is this happening a lot, all over the place? Your previous log called GlobalMinus 35.5 has a similar clip but that part of it looks normal, Wideband showing mid-14's like it should.

 

[justacanuck]

I'm a little confused by one thing in your last log, the "...V10" log. In the first second of it, it is in closed loop and your real O2 (Front O2) is cycling like it should, so the AFR should be about stoic there, but the Wideband shows 10.1. That's very odd. Can't see what came before though. It's just that one second.
Is this happening a lot, all over the place? Your previous log called GlobalMinus 35.5 has a similar clip but that part of it looks normal, Wideband showing mid-14's like it should.

I noticed that. We had an intermittent wiring issue on the back of the WB gauge. I think it goes to a minimum of 10.1 when it isn't working properly. Thankfully we aren't simulating the narrowband signal with it.

 

[Justin DuBois]

Youth is king…
You picked up on the concepts and got this done quickly! Nicely done

 

[We're on Boost]

Just want to make sure there's nothing gone wrong in the Autometer wideband settings.
As near as I can tell from the Autometer manual, you can change the volts to lambda relationship in what they call the "BGD Range Setting Mode".
Then you can change that again in the ECMlink window "LinWideBand Preferences".
In your LinWideBand Preferences I see you have the max voltage of 4.0 set to be 1.10 Lambda, which is kind of a unusual setting. That's a gasoline AFR of only about 16.0.
That's ok if you have it setup the same way in the Autometer BGD Range setting. But the default BGD setting would not have been that, I'm pretty sure.

I've had AEM and Innovate widebands and none of them have this variable volts to lambda in the wideband itself. They give you a chart that shows you what the volts to lambda relation is period, you can't change it. Then by each lambda point they show you what the gasoline AFR is for that lambda. So the other thing I'm wondering is if the Autometer wideband is really giving you this option - or maybe I'm just reading the manual wrong.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

 

[Justin DuBois]

Huh - that's an interesting way to describe it. I use this dialog to pick two points - 0v and 4v in this case, but it could be anything, 1.5v - 4.5v. you pick. And then, you define the lambda value for each point - and a linear mapping between the values will be used to display what is coming from the sensor input.

The values to use typically come from the sensor manufacturer. Most of them publish a lambda to volts table.

 

[We're on Boost]

Huh - that's an interesting way to describe it. I use this dialog to pick two points - 0v and 4v in this case, but it could be anything, 1.5v - 4.5v. you pick. And then, you define the lambda value for each point - and a linear mapping between the values will be used to display what is coming from the sensor input.

The values to use typically come from the sensor manufacturer. Most of them publish a lambda to volts table.

So I just looked at the user manual for the Zeitronics ZT-3 Wideband and man they do a good job of explaining this thing I'm talking about that is vague in the Autometer manual. They show that you do it with software on your laptop which you connect with a serial cable and a splitter that they give you in the kit. They say "Output voltage versus AFR can be customized in the Zeitronix Data Logger software." And they tell you that the default is "9.6 AFR at 0 volts and 19.6 AFR at 5 volts".

I think you can customize voltage versus AFR within the Autometer too, but you do it by fishing through menus with the 2 buttons that are on the gauge. The Mode and Select buttons. And I know from my old AEM boost controller which also uses 2 buttons to fish through menus, it's about as much fun as doing the worksheet to figure out how much of my social security income is taxable.

Anyway my main point for post #94 was to say that the volts to Lambda relation you have set up in LinWideBand Preferences has to be the same as what you have in the Autometer BGD Range setting. I think. That's it really.

Here's how they show volts to AFR in the ZT-3 manual - good job. The white wire from the "Signal" port is what would go to our ecu. Otherwise it's as shown here:

You must be logged in to view this image or video.

 

[justacanuck]

In your LinWideBand Preferences I see you have the max voltage of 4.0 set to be 1.10 Lambda, which is kind of a unusual setting. That's a gasoline AFR of only about 16.0.
That's ok if you have it setup the same way in the Autometer BGD Range setting. But the default BGD setting would not have been that, I'm pretty sure.

This gauge is supposed to have a range of 10-17.

I did change the 4v Lambda value a tiny bit lower than the factory says it should be, to help get my AFR readings in ECM link better aligned with what the engine was doing with fuel trums under closed loop conditions. I found with the "normal" settings, my AFR in ECMLink, were always reading a little leaner than it actually was. The log data also better matches the gauge itself this way. I believe this could be caused by a small discrepency in the voltage supplied, or ground point/wiring etc. I normally use fuel trims to adjust VE rather than wideband log data anyway so I figured this should be OK.

I would be more concerned about changing the lower range setting, because I have nothing else to compare to at WOT, and the guage obviously has no problem providing 0 volts.

I am not sure I can adjust this one (autometer phantom) but it might be worth checking to see if it can be adjusted in the meter, or perhaps even even better idea would be to try and get it calibrated against another sensor.

 

[justacanuck]

Youth is king…
You picked up on the concepts and got this done quickly! Nicely done

I feel pretty old... but l'll take that as a compliment.

 

[Justin DuBois]

I feel pretty old... but l'll take that as a compliment.

Yeah - compliment - I'm the old one on this.

 

[steve]

Yeah - compliment - I'm the old one on this.

Not by any means...