Not seeing the gain I expected. :-( What's going on?? **solved**
 

To avoid wasting anyone's time, I copy post 18 in here:

---------------------- Original post----------------
I did 4 pulls today on as perfect flat road as they get - 2 in each direction.

I loaded the logs with Virtual Dyno, compensating for IAT and ambient pressure as read out from the logs at 4000 rpm. Compensated for wheel size (own measurement) to compare with my old logs (winter wheels).

So here are the results:
- Red, blue, dark green, yellow (first 4, are todays pulls). This is with a Gruppe S header and Wayno's most recent stg 2 tune and OFH timing.

The light green:
- A bit unreliable from when everything (incl tune) was stock but the best I have and fairly similar shape and number as an actual real life dyno pull I had made during a dyno day.

Here is the bummer:
- The purple pull is one of Wayno's most recent stg 1 tunes, with some extra timing. I don't think that it's more power than some of todays pulls as virtual dyno suggests. But just the fact that it appears to be in the same ball park I find truly disappointing. :(

I was expecting a 10 hp gain from installing the header. While the 4000 rpm torque dip improvement is nice, I'm not sure I think it can be right?

Butt dyno says there is more power obviously especially in the torque dip. But now I am starting to doubt if the rest perceived improvement is placebo? I'm a man of science, I tend to believe more in number than feelings.

Don't know if it's screwing up because of the wheel sizes? The old pulls where with AD08Rs in 225x40R18... Todays pulls with stock Primacys 215x45R17.

I'm close to putting the AD08Rs on to see if it makes a difference, though it might not be the best idea with temperatures below freezing at the moment.

If these are indeed true numbers, the header is being put up for sale!

https://s24.postimg.org/7058w0mkl/dyno.jpg

Edit:
Here are the logs in datazap (3 logs):
http://datazap.me/u/tor/ft86club

No IAM drops in play or otherwise strange.

As mentioned IAT taken from 4000 rpm and pressure from an average of the highest and lowest in the logs, wheel sizes 24.88 for the old pulls, 24.6 for todays pulls. 2769 lbs weight for the car and 165 for the driver.

Just go FI lmao

I can't... Living in Germany. It's fucking monopoly here.

There are only 2 legal kits. A turbo and HKS, and they can only be installed by one of the 2 companies that provide them respectively. Naturally they charge a shitload of money. Converted 8000$+. No fucking way I am going to spend that much money on a car that's worth 15-20000$.

Then I'd rather buy a Porsche.

What about engine swap?

Technically I'm not even allowed to swap the header. :bonk:

But a header I can swap back in 30 mins every 2 years when the car has to get it's mandatory TÜV inspection.

I'm not quite sure about the orgins, but I think they introduced TÜV in the 1940'es. :(

So stock was 156 hp/170 Nm, and pull 4 was 175 hp/190 Nm with a catless UEL header and tune. You gained 19 hp/20 Nm overall, and a lot more area under the curves. I'm not sure I understand the problem. Most UEL headers and tunes will get you in the 175-185 range. If you want to push 200 hp you need one of the nicer EL headers such as PTuning or ACE. Or go with E85, but I don't know how feasible that is where you are.

awe shit. Well that blows (kinda sorta pun there)

I don't know how they do a smog test in Germany, but in Belgium I passed the smog test with an OpenFlash UEL Header installed. Compleet exhaust-setup was OFH UEL, Milltek overpipe, OEM Frontpipe, Milltek CBE (the non EC version ;) ). Passed without any problem :)
You can also use E85 to lower your emissions if necessary and E85 gives you a nice bump in power also.

The problem is that without the catless header and just a tune I got 170-172 hp (the one used here is not a fluke, all my previous pulls on flat road was in that ballpark).

Sure there is under the curve improvements but only below 5000 rpm. I track the car, I was hoping to see improvement on the top as well.

I was hoping someone would post that Virtual Dyno is shit, but seeing the consistency between pulls I don't believe it can be discounted totally.

I don't want it to sound like I am bitching. Rather, I just don't get it! I run 5 deg more timing up top. What's wrong here??

Also, I am not bitching about the tunes. Rather I'm surprised how good and sorted Wayno's most recent stg 1is. If it outputs similar numbers at the top with 5 degs less timing that's truly astonishing.

The gained torque in the dip is nice, but I'm not sure I'm going to bother putting it back on next time I have to take it off for TÜV inspection.

Exactly, I was expecting to see 185 hp. Since the Gruppe S/Tomei header supposedly is regarded as one of the better UEL headers. That's why choose that.

I seriously doubt the 200 hp claim of the Ace headers. That's nice marketing from CSG. Probably they are max 190. 5 hp more wasn't worth more than double the price for me.

I'll buy your headers then...

Well I will only be selling one header. ;) And because people here are scared of ordering overseas I could probably sell it with a profit even after using it a year until the next time the car needs inspection and I have to take it off.

It's not just a smog test. They look at the vehicle too and I wouldn't want to risk going there with the header installed and praying they don't notice. They are nuts here, driving without a cat is considered tax evasion, since you pay a yearly tax fee to have the car registered based on emissions.

Oh dear, your German is dominating your application of English.

In Belgium it's the same. It's illegal to remove any cats, only a CE approved CBE is allowed. But the header isn't visible from underneath the car and if you have a stock (looking) enginebay you don't see it from above also. The soundpipe blocks the view a little bit. With my car they only looked to see of the engine was in the enginebay :p . Guess I was lucky. But then again here are inspectors who try to smog test a Nissan Leaf.

Verstuurd vanaf mijn Xperia L met Tapatalk

You need "i know someone" who works on that smog inspection

Haha :D I just find it funny that a lot here write headers like it's a V8. And I'm not even German, I just occupying some of their land as retaliation of the past.

The Gruppe S sticks out a bit in front of the engine, the front O2 sensor is clearly visible and clearly not going into a cat and I deleted the sound generator that blocks the view. Though I wrapped the header it's still pretty obvious. I'm not sure I will risk it, the swap is really fast.

But we are drifting off topic which is "where are my 10 hp". :D

I was clearly a bit annoyed at the logs earlier and calmed down a bit. The improvement in the torque dip is nice for driving the car on the streets and for sure will have some benefit on track too. Also, it doesn't sound that bad though I prefer the EL sound. I was just hoping for more.

Now that I have it, I'm going to keep it and keep tuning. I was already thinking about E85 before (though it's a pain to get here), so this is extra motivation. I now expect a 20 hp gain from E85, since I got cheated of 10 here. :D:D

You could try making it run a bit leaner and see if you can get a few more ponies from that. The OFH ignition advance is already pretty aggressive, though you could try adding ignition advance incrementally where there is no knock corrections.

The tire size you inputted will have an effect on the Vdyno calculations and total weight to a lesser extent. The numbers might be slightly off between different wheel/tire setups, so that might have thrown your numbers off a little. Though 174Hp with a stg1 tune and stock header is pretty impressive. It's quite possible that the stg.1 pull you did was reading high. Good luck. :)

For your reference, my stg 1 tune got me about 165-168hp on 91 fuel. Stg 2 with Tomei EL my pulls were getting between 178-181hp pretty consistently.

Guys, I must admit I this thread was a mistake. Thank you for all your input. I apologize much for wasting your time!

I didn't do my due diligence and jumped to conclusions. I relied on memory and didn't load up all my old logs. I used one perhaps too optimistic log as a reference.

Anyway, loading up a bunch of logs, sorting and discarding based on shape and deviating number in either direction I came up with this. Admittedly it's biased towards the better numbers - but hey that what the pros do too :D

https://s23.postimg.org/5vsi25d17/dyno2.jpg

So in reality, I may be around 175-177 hp now vs. 168-170 before. More important than the actual number, this gap is present in the whole curve now. So I'm more inclined now to go through the hassle of keeping it. :)

It's already running pretty lean at 11.7 and there is already some slight FLKC, so I don't think it will be more headache that gain going in that direction. Also, this needs to be a track capable tune.

Yes, the above now better match those numbers. I'll do some more runs at different temperatures later on, maybe the compensations cause deviations too. Best thing would be a pull at 20 deg C and with the other wheels.

How good is your fuel over there TOR? These 86 motors really need good fuel.

It's awesome. :party0030: 100 Octane RON = 95 AKI. We have 102 as well = 97 AKI. The brand fuel I am using is 100 RON and allegedly they deliver higher octane than advertised.

It's not even complaining much at 4,5 and 6th gear WOT:

https://s29.postimg.org/8hts0fv93/Sc...t_04_50_59.png

another thing that I'd do to try to be more consistent converting logs to Vdyno is look at the RPM's in the logs throughout each pull. If there are any unusual dips or spikes in any part of the pull, I would not use that data for Vdyno. It will read falsely high in most cases. Small dips and spikes in the RPM during a pull indicates wheel slip, even small ones, will cause the converted information to be inaccurate.

Also, I noticed you using smoothing factor 3. I generally like to use 2 smoothing. The higher the smoothing factor selected the lower the converted numbers, typically. But I suppose it helps smooth out inconsistencies between pulls.

Glad to see someone using this tool correctly, as I've seen people post some pretty wild looking Vdyno pulls on this forum. Pretty much all the ones I've seen on this forum looked horribly wrong besides yours.

Maybe try a stage 2 tune ? There are changes in them specifically to take advantage of losing the catalyic converter, and may give you a bit better result.

Sent from my LG-H812 using Tapatalk

I believe multiple tuners have tried extracting extra power from my header, but all have failed. Also, it's noticeably peppier than a stock header and tune. So I'm confused. But anyway...

Thank you for your help. I didn't consider minute grip differences could be a factor, but that would make sense. But wouldn't wheel spin cause a lower reading? I mean we measure the time it take to accelerate a certain mass to a certain speed, and if it takes longer to get there it should show an overall lower reading? I understand it may locally bump the curve up where the spinning is taking place

Thanks for the tip. I probably went to 3 at some point where I looked at curves that didn't have the best data. I now tried factor 2 and even 1. They both give good curves with pulls that are reasonable, not so much with pulls that are questionable.

It seems I used the worst data and manipulated it wrong in the first post.

Thanks. I messed up the first ones and it really confused me not to see an improvement. I am going to change the subject to "solved". Thanks for your help! :)

Here smoothing 1 (just to make me happy). :D This is just another confirmation to me to take any actual dyno charts from anyone that is looking to sell anything with a grain of salt.

https://s24.postimg.org/lhsc9jw11/Sc...t_05_17_50.png

It already is. Stg 2 vs stg 1 vs. stock. That's why I was confused not to see an improvement.

If you're getting that FLKC every pul, you'll likely make more consistent power by getting rid of it.

I'd also address the rich dip at 3500-4000RPM.

I'm running the same header and Wayno tune and I have exactly the same rich dip at 3500 rpm.

Thanks, yes I will at some point when the LTFT settles, I know from experience that it will go slightly richer with this MAF scale.

I don't know if I should wait with the effort of adjusting the timing until it gets warmer? At the moment I have IATs of 5-15 degs C, I assume I will need to remove even more with more humane 20-25 degs, which is what I have on average through the year?

What speaks against touching it is that in OL, that MAF scale produces so consistent number that I know what it will learn before the itself ECU does. And once the LTFT settles they have very minute corrections (like +0.8 +1.2). And as long as it's in closed loop it doesn't matter anyway?

I did let the Vgi tool chew on my MAF scale but decided not to correct it. What it suggest is almost impossible to smooth out meaningfully. And it's right in the critical spot 2.5 to 3.2 volts that it want to have the MAF look like a rollercoaster. The reason I put it through the Vgi tool was because it was slightly lean on top (0.2 higher) after switching to stg 2 tune. But I decided just to add 2 percent above 3.2 volts instead, and I guess it will settle at 11.5 once it's finish adding LTFT (it's 11.7 during the pulls for Vdyn).

Any suggestions how to go about that?
https://s28.postimg.org/6w20eljlp/Sc...t_13_06_42.png

By the way, one of the things I love about Wayno's new tunes is the lack of learning required. With the uniform injector ratio, what the learning does is just to move the whole MAF scale up or down a slight bit when the LFTF finishes learning. If the MAF is good it won't move much at all. I feel absolutely no difference in performance between a fresh flash and driving several hundred kilometers, because all it will do is move the whole MAF scale about 0.2 down.

I went through a lot of my old logs and even the stg 1 tunes do it right at the same spot, just not quite as pronounced. Also the old tunes before the uniform injector ratios do it. I guess it has to do with the AVCS?

I don't have a lot of knowledge on this topic. However I did compare a couple of logs from my friends, running the same header and tune (122.1), and the AFR looked like this, three different shapes:

https://lh3.googleusercontent.com/jE...=w1091-h323-no

You can clearly see my car with the pronounced rich dip.

On my car with the 122.5 tune it looks like this (uncorrected for SAE), it's the 122.5 100 stg 2 UEL +2% (default) Wayno tune:

https://lh3.googleusercontent.com/vZ...A=w978-h643-no

As you can the shape of our AFR is very similar. I played around with some MAF scaling (generated by a friend) , the lean hump at 4 - 4.5k rpm was then resolved. However the rich dip at 3.5 - 4k rpm remained.

.... Just take some fuel out the fuel map. If MAF scale is good, then don't use that to fix it. There's more than one way to do it ;)

I believe Vdyno measure how quickly RPM's build for a given gearing/weight to make it's calculations. Therefore, a spike in the RPM will be falsely calculated as an increase in power.

Yes it's difficult because it all happens between 3 and 3.4 volts when you only have 2 point on the curve to manipulate. I could try and richen up 3.4 a notch, but I am guessing it would create another problem.

You mean instead fix it in open loop fueling so that AFR becomes correct and AFR_Command has the tops and valleys instead?

Yes, that's what I meant with locally more power. But once the wheels grip again I would assume the subsequent data would then be calculated as a decrease and looking over a bigger part of the graf RPM would build slower because time was wasted spinning the wheels?

It doesn't necessarily mean it's running rich, just that there's more fuel blowing by the valves and past the O2 sensor, so may actually be lean. It's not even below 11.5. Unless making it leaner produces actual measurable power, leave it as is.

Yes, in open loop you don't need commanded and actual AFR to match.

Reverse that logic ;) Unburnt fuel isn't measured by the sensor, so it can't read richer than the burnt mixture. It can read leaner than the total fuel injected, but not the other way around. A more stable AFR is ideal, you want it to gradually richen up. For the sake of a quick tweak to the fuel map, it's an easy fix. Helps with EGTs etc, the timing curve I'd imagine is smooth so you don't want it too rich when the timing advance is less. You want them to almost be the reverse of each other.

I made a video of the drive when I was logging. The second of the 4 pulls is at 1:50 into the video with datazap overlay. As one can imagine from the rest of the video it's pretty difficult to find a flat level road here. The one in the video runs along a river, that's why I'm pretty sure it's level. There are a lot of roads here that looks level but has a lot of gradient, which becomes apparent when doing pulls in both directions.

[ame="https://www.youtube.com/watch?v=Hw0JiDWUpBc"]DIY tuning: Performance check... - YouTube[/ame]


Thanks for the explanation. It's good to know there are other options than messing with the MAF. Which I would rather not touch anymore except to richen up the top a bit if LTFT doesn't take care of that eventually (crap weather here again).

I had to read this several times and my head is about to explode thinking about what unburnt fuel may do to an O2 reading. :D You say "it can read leaner" i.e it's actually richer, where Wayno say it "may actually be lean". If the O2 sensor just reads the oxygen ratio, isn't all what unburnt fuel can do just to dilute the remaining composition of the exhaust gasses? Which in terms would mean the sensor reads less oxygen (rich?) - In that the reading is rich and the actual combustion is leaner than the reading? Or where is the fallacy?

Arggg my head is slowly exploding in a lean, knock filled detonation! Got to think of something else quick!!

https://s30.postimg.org/6i0tw0bld/beer_intro.jpg
https://s29.postimg.org/5b9ubuchz/sunbathing.jpg

(Ahh, now I'm feeling better again. :D )

As you can see the car hauls pretty good now. So on one hand, I am inclined to leave it as it is as Wayno suggests. On the other hand, I am curious by nature and yes, being a perfectionist, the bump bothers me. If I was to apply corrections to OL: Would this look in a reasonable order?

https://s24.postimg.org/ownab5jth/Sc...t_13_40_43.png

https://s30.postimg.org/vmcmipz8h/Sc...t_14_46_34.png

Markers at 3200, 3600, 4000 and 4400 rpm:
Log here...
https://s28.postimg.org/cjl8xtinh/Sc...t_14_52_20.png

One of the things I did was completely change the MAF scale so that I have a lot more points in the critical region around 3V.

I don't have the numbers in front of me but my MAF scale tops out at 4.25 V since I am not boosted. I also changed the spacing of the voltage points down in the very low region where it is very linear. The end result is I have a more points to work with in the 2.5V to 3.5V region.

Same as the MAF, calculate the fueling error and multiply the commanded AFR as required. Need 5% less fuel, multiply the table value by 1.05.

The Lambda sensor is a gas sensor. Unless you were to literally flood the exhaust with fuel, unburnt fuel is not detected. Think about it like this, if you have a lot of cam overlap there can be air bypassing the cylinder and therefore the reading can be lean. But you can't have a reading that is richer than the mixture in the cylinder. Unburnt fuel reaching the exhaust will be almost always be when there is cam overlap and therefore a lean reading will occur.

I suppose if it is caused by the AVCS that I could adjust the whole rpm row where it's occurring, in all loads? I'm going to start out with the small change I showed above next time I flash and see that it doesn't mess anything up.

I am doing it just for fun and out of interest. I think if the timing doesn't make any problems when it get's warmer that I could just leave it as it is. The car runs great.

Not sure I understand that. If the fuel is not in the cylinder during combustion (because it bypassed the cylinder through the valve over lap). And the fuel in the exhaust is not being detected. Why would it read lean? Why wouldn't it read accurately? I mean the fuel is not there, and it's not being detected.

On the other hand, wouldn't some of the fuel burn in the runners before reaching the sensor? If that was the case (don't know if that is possible), it could read richer than the combustion taking place in the cylinders?

Still, it would need to be scaled to a weird shape (in my case at least) so I'm not sure that would help. I think Kodename47's solution is more elegant as it keeps the MAF scale linear (or exponential or whatever the correct mathematical terms is for smooth :) ).

Perhaps.

I managed to get my AFR / commanded AFR pretty close and flat with my last round of load limit and MAF changes but then it got cold up here so I can do any more changes unti it get warm again.

I would just do it in the load regions you're seeing in that pull and then taper it in the table if it's vastly different. Don't go changing the whole row with a blanket change. Don't forget that you can give yourself more accuracy where you want it by changing axis etc. There's a fair bit of redundancy is the stock fuel map ;)

What I meant was that if there is unburnt fuel in the exhaust, it's likely that it's in air that has gone right through the cylinder without being part of the combustion cycle. This unused air will make the O2 sensor read leaner as there will be a higher ratio of oxygen gas.

If the ignition event is retarded enough that it continues to burn as it exits the valve, then this will still be converted to gases and read more normally. This is more common, but your solution here is less fuel or more advanced ignition timing as this is bad in every way as it causes high EGTs, can damage valves, creates more emissions and usually makes less power.

86Inches on Romraider has just released a spreadsheet that gives a perfect exponential curve, I have yet to flash the result but a well respected user says that it makes the car smoother..... will feed back and post my version of the tool if it works. ;)