|Topic Review (Newest First)|
|August 18th, 2019 08:16 PM|
Doin' more work. Actually making 15psi this time. Starting at a 15 mph roll and lazily shifting at 5000 RPM, but it pulls hard. I'm disappointed with how it doesn't show how hard it went or how fast it got going, but you can kind of see how it picks up the nose. Might look into alternate camera angles next time as well. The downside to mounting it outside is you don't get anything for sound. I'm open to suggestions.
|August 16th, 2019 05:20 PM|
|Snorky||Glad it was something simple like a down pipe slip fit coupler. Before i welded on ball and socket junctions that would happen all the time. I hate the way open down pipe turbo mustangs sound. At least single turbo mustangs anyway. I dont see how guys run them on the street.. they sound like a tractor|
|August 15th, 2019 04:36 PM|
|1C2||Sounds good! If it were winning car shows it wouldn't be a dumpster fire|
|August 11th, 2019 07:13 PM|
Always new shit.
Sliced my hands to ribbons changing up the boost reference lines to the wastegates. Rather than 5psi pre-throttle they're set to 15psi at the manifold. And to make sure I taped a known good pressure gauge to the column to see what it would do with the configuration.
Well hot damn does it make 15psi now. Didn't get to see how fast it was with keeping my eyes on the pressure gauge and wideband, but it sure performs. Didn't get video of it just yet.
The shutdown didn't go quite as well. After letting out the wideband gave me an error, and the engine sounded like dogshit. Uh oh. As the car slowed down it sounded awful, and loud. I had a bad feeling, but there was no smoke, no faults, no vibration, and with how it sounded I had a feeling it was an exhaust leak. Either way, I cut the testing session short and limped it back home, getting dirty looks the whole way for this car that sounds like...well, a straight piped V6. I figured one of the wastegates or something came loose, something that I touched yesterday, but when I parked it, it was pretty obvious where the issue was, and wasn't something I messed with.
Just one of the slip fit connections at the downpipe popped out. And pretty significantly. Happy it's a pretty easy fix, but surprised it blew it out that bad.
After that debacle I'm kind of glad I'm not dumping the exhaust out the bumpers. It would sound absolutely terrible. Not just volume wise, but tone. I guess that's what you get when there's a turbo four cylinder sticking out each side. :
|August 7th, 2019 09:41 PM|
Switched around the bypass valve reference. Seemed to help a little, but not a ton. Easier to break the tires loose in 1st. One blip of the throttle did it at one point. Then today I rolled into it, and when the turbos lit, it knocked them loose. Still not seeing the full 15psi on the gauge I want, but it does hit pretty hard.
I'm going to switch the wastegate references. Right now they're set at 5psi and relying on the blower's 1.56 pressure ratio to get to the full 15. I thought it would better keep me from overboosting, but it might be holding me back. I checked with another guy running a similar compound setup and he's referencing the manifold. It looks like I'm going to try that route and see if it gets me the full 15 I want.
But here's the juicy stuff, even if it was only 3/4 throttle and not making full boost. Still sounds pretty decent, even if it doesn't capture how it felt.
|July 28th, 2019 08:48 PM|
I need to do a better job of uploading pictures and video.
Went to a car show last week. Due to the heat there was about 30 cars...and more than 30 trophies. Thought I had a good shot. There were 4 cars in my class, and they gave trophies for 1st, 2nd and 3rd. Guess who went home empty handed, losing out to a stock 1985 Silverado? Even got shit talked by the judges about the exterior and broken trunk struts. At least I got 3 points for having a fire extinguisher, 1 point for being a daily driver, 1 point for interior, and 1 point for the engine for a grand total of 6 points, including a 0 for the exterior. I didn't see any other cars get a 0 for anything. So...par for the course on the car I guess. They don't call it a dumpsterfire for nothing.
On to the more fun stuff. Romped on it some more. She pulls hard, but not as hard as I'd like. This time giving it 75% throttle I made 10psi at 4000 RPM, still short of what it should be making. Airflow and load data match up to what I saw; still short of what it should be making, maybe around 425hp, 1.45 load vs the 500hp and 1.9X load it should see. I also did a quick 1st-2nd pull. It pulled hard, enough to lock up the seat belts and pop some of the gauges out, but didn't even register boost. I doubt the 300hp NA version would do that.
I know the turbos are making their 5psi by 2500 RPM at 50% throttle from the previous test. Boost climbs with RPM. It should spike up to 8psi from the blower right away and then climb like a rocket as the turbos spool. The question is why. Could the blower have an issue with achieving its pressure ratio? Is it just not behaving like an ideal compressor? I don't think it's completely off since the turbos light way faster than they should on their own, and it makes more than 5psi total. Is the boost gauge off? You know I calibrated it before.
I've got plenty of levers to crank the boost up if need be. Going to a smaller pulley will do it, but will create more heat. I could switch the boost reference for the wastegates to the manifold and command them to 15psi regardless of what the blower does. Just a matter of turning her up if need be.
I'll try to get some pictures and video up this week, even if it isn't at full power.
|July 1st, 2019 10:11 AM|
|June 23rd, 2019 11:42 PM|
|porterhousestk||Very cool! Glad to see you've got it working!|
|June 23rd, 2019 08:32 PM|
Car update stuff:
Over the moon with it. Smooth as silk. After taking the no load rev data and a little extrapolation I updated the MAF transfer function. It drives like a dream. It's incredible. No hesitation, no bucking, no breaking up, no stalling. I made more progress in a week with this hardware than I did in two years. I've got it up about as fast as I can get around town and it's all on the money. +/-3% everywhere, +/-1% in a lot of areas. For what it's worth most tuners say being within 5% is about all you can ask for, and you'll be chasing your tail to get it perfect. If I wasn't polling all the data all the time you wouldn't even know. I can actually get even more aggressive with the dashpot and DFSO functions now if I wanted, rather than using them as a bandaid to make it behave better.
Wicked's tuning corner:
This whole setup is pretty trick. I've got TunerPro RT and my wideband function loaded. Now everything is working as it should. You can run it and run through whatever duty cycle you need. You start off with the no load revs. Just sitting in neutral, bring it up from idle little by little, let it stabilize at each step for a bit, and keep creeping it up and up in 500 or 1000 RPM increments.
Now once you do that, you can see how much deviation there is using a histogram function. It's really the same as LiveLink in that regard too. I find it works best as a "running average". In the simplest way, you can plot your STFTs versus your MAF signal. With the more sophisticated approach of fuel error (as built in to the efidynotuning adx file, but you can do the same with an equation in LiveLink), you find a total delta combining both the STFT and the wideband reading.
I'm sure I lost a lot of you there, so here's an example.
Let's say at 130 counts you have a STFT of 0.75 and a wideband reading of 16 (1.136 lambda for E10 gasoline, or AFR actual/AFR stoichiometric, in this case 16/14.08). That means the ECU is adding 25% fuel, and it's still 13.6% lean, so your total error is 25+13.6=38.6%. You multiply that airflow point by 1.386, and it should be right on the money.
If you have a Quarterhorse, you can do that on the fly! Though honestly I like to run my points, finish, look at the data, adjust things, and reload the tune. But that's still a 5 minute or less process rather than a 30-45 minute process with the SCT software. I highly recommend it.
Now you can do that throughout the whole range. The histogram tells you which cells are off and by how much. You multiply accordingly. If it's in between MAF points, you may have to interpolate. For the higher end ones, you'll have to extrapolate. For instance if your free rev only goes up to 200 counts, and you multiply the 200 count point by 1.3, you also multiply everything above 200 by 1.3.
Once you have all that done, you can repeat the process on the road. It can get a little tricky here. Lots of jumping around with shifting or DFSO can throw your data off, so the more of it that's just a nice steady drive and lugging up, the better your data will be. Then you run it through the same process and update accordingly.
The only exception is where commanded lambda is <1, normally as you approach full load and/or boost. In this case you're usually in open loop so your STFTs are 1 anyway, and you just need to make sure the AFR matches the one in your stabilized fuel map based on RPM and load. It can get to be a bit of a pain if your map varies a lot based on RPM and load. But once you do a good WOT pull, that will get most of the curve populated, as long as you don't zip through it too quick.
|June 16th, 2019 07:32 PM|
Well I know I wanted to get some video, but mother nature had other plans. Too busy not getting blown away to Oz. This motivational poster seems all too fitting with the weather and this project.
The current tune is pretty awesome under power, or anything above 45mph. But puttering around town at 30mph? Still bugging me. I'm attributing it largely to the non-linear part of the MAF function. Hot idle is waaaaay low. Higher load does ok, and is a lot more air. At very light load, with all the variation, it can cross over between the two regions. I'm thinking that is part of the issue.
So me being me, I have to fix things until they're broken. I took it out in the one hour of rain free time, took the car out, got a log, and warmed it up some. Then ripped apart the intake and put the MAF screen in again.
Everyone right now, reading this (ok, maybe the two people that have followed this thread and are still reading it):
But hear my dumb ass out now. My issue is largely the signal variation. I'm seeing the MAF signal fluctuate by 60+ counts at a pretty high frequency. That coupled with the non linear step area of the MAF curve is making things really difficult. I need a way to even things out.
"But you already tried the screen, you fucking moron!"
Yes, yes I did. And it ran rich and surged like a motherfucker. But I only tried it once, and that was back when I was using SCT, where revisions took 30 minutes. Now that I can change on the fly with TunerPro, it's another matter.
I threw it in and fired the car up. Pig rich again, a faint surge. But with the ability to change on the fly, I took some rough chops at the airflow. And I had to make some big adjustments. I'm talking reducing airflow by over 50% in the transfer function. And then I'm in the ballpark! It can even smoothly rev with no load without breaking up which it couldn't before. MAF count variation is down to 6. That's a 90% reduction. The whole transfer function needs to be rescaled, but I've got a good feeling this shitbox might get some much, much better street manners.
Now we just need the weather to cooperate.
|June 12th, 2019 01:55 PM|
Originally Posted by 01Bullitt82 View Post
1.) Read book on tuning. Everyone tells you that you only have to tune MAF curve, borderline knock spark, injector hi/lo slope and breakpoint, and maybe your fuel table. Easy peasy, right?
2.) Tune those parameters.
3.) Car runs like shit.
4.) Discover there's another table that no one will tell you about.
5.) Tune that table.
6.) Try again, feeling optimistic you fixed it.
7.) Car still runs like shit.
8.) Repeat steps 4-7 ad nauseum.
10.) Throw money at tuner/buy standalone/sell car
Although if you go with throwing money at a tuner in step 10, you might get worse results. After seeing some of the stuff my previous tuner did...No thank you.
I'm far from the most knowledgeable when it comes to this stuff. I know just enough to be dangerous. There's some guys out there I'm still in awe of. If you want a remote tune, efidynotuning is a whiz at this stuff. I haven't caved and handed it off to him, but his write ups are invaluable.
|June 12th, 2019 10:25 AM|
|Eagle2000GT||I don't understand tuning either. I had my supercharger installed when handheld tuners were just beginning to have more functionality. My tune is on a chip. I couldn't change it if I wanted to.|
|June 12th, 2019 08:59 AM|
|01Bullitt82||Great wright up, I just wish I better understood how this whole tuning thing works. I really haven't gotten into tuning myself yet. I'm very interested in it and love reading about it. Just not sure if I will ever trust myself to do it.|
|June 12th, 2019 08:43 AM|
Haven't been updating this since no one goes on forums anymore. Do I have to move to the gram like all the cool kids did 3 years ago?
I've got a 2.5 gallon reservoir and the pump mounted in the trunk now. Totally trick. Between that, the battery shutoff, and the quick disconnect bumper, it's getting into racecar territory.
Took me awhile, but I'm up to speed with TunerPro RT. Pretty neat software. Figured out how to change values on the fly, compare tunes, set up a dashboard for it, etc.. Pretty fun.
While collecting data the one day I noticed it was tough getting the boost gauge to move. So what do I do? Strap a pressure gauge to a windshield wiper with the handyman's secret weapon, duct tape, that's tied in to the turbos. I get on it a little and see the boost come up to 4psi out of the 5 that the gates are set at by 3000 RPM. Main (2nd stage) boost gauge still shows 0, and I know that one works. Hmm. Let's go to the tapes. Or rather the log files. Even when I was getting into it, or so I thought, I wasn't even touching half throttle. I did a few sweeps to confirm pedal travel and I can max it out, but there's a whole lot more travel than I thought. I guess that's good news, in that it's functioning, and that the turbos are lit by half throttle, even with that puny 281 feeding them. It's only that pesky throttle blade in the way.
Now going through those log files I found something else that was interesting. Or at least interesting to anyone tuning.
When I was looking at those pulls, it was showing load as 86%. That didn't seem right, with the turbos trying to cram 4psi into it. I hand calculate the actual load at that point, and it was anywhere from 96-110%, but reported load (aka perload) was flat lining. How could this be?
In the software, there's a feature called the perload switch. There's three different ways it can be set.
If it's set to 2, as is the default, it uses alpha-N (or throttle position and RPM) to estimate load for the fueling tables rather than actual load. This is also known as the load with failed MAF (LWFM) table. You know, that table tuners tell you isn't even used on 02 and earlier cars, who needs it? I dive in to that table, and what would you know, it flat lines at 86% just like the values I was seeing. Needless to say, with 2 power adders on a 4V it's not going to look remotely like a 2V stock table. I'll eventually get back to recalibrating it to my engine after a whole lot of data collection, but in the meantime I dumped the 03 Cobra values in there and...
...when you set it to 1, perload = load. Simple as that. That's how it's currently set. Sure it has some drawbacks, less resolution especially at lower RPM where VE is lower, and isn't as accurate with altitude changes, but this will work until I can get enough data to populate the other tables.
Now if you're really ambitious, you can set it to 0. This brings in another table (and isn't supported in all strategies), which basically tells the ECU what your VE curve looks like with MAP = 1 atm and that becomes the reference that your current VE is divided by. For instance, say your engine only hits 80% VE at 1000 RPM when MAP = 1 atm. That's your value for 1000 RPM in that table. With the switch set to 0, if you're at 40% VE at 1000 RPM, your load is 50% (80/40 = 0.5) That gives you better resolution, and a better idea of how to adjust things based on boost.
Now here's the big mindfuck: all those spark tables? Yeah they don't use perload or these tables for calculations. For them load is load. Simple right?
Thus concludes wicked's tuning corner for the week, teaching you things I wish I knew off the bat.
Now with all that nonsense out of the way, I hoped to get a few pulls in last weekend and get some video. Several inches of rain put a damper on that. I did get it out for a cruise night this week. Idle is behaving much better, to the point I can start getting more aggressive with dashpot that was being used as a bandaid before. The drastic shift in load tables is going to require some drastic rescaling of the MAF curve at high load though; for now it's pretty rich. Nothing that can't be fixed.
Another perk of TunerPro is it wouldn't be too hard to run it off a tablet. I might pick up a cheap one and mount it in the car to make tuning even easier, as well as letting me see tons of values and indicators while driving. A whole lot cheaper than the expensive race dashes out there.
And of course, since you kids clamor for media these days, get some video, hopefully beyond half throttle.
|April 7th, 2019 07:35 PM|
One step forward, two steps back.
Put the sight tube and new pressure cap on the intercooler reservoir. That goes well. But then it starts leaking from a seam. Instead of trying to dick around with that and try to epoxy it up, I figure I'll just throw the old reservoir in to make some pulls in the time being until I can get or make a new tank.
I put the old reservoir in, and due to its small size and the flow rate of the new pump, it just aerates the shit out of the coolant. Once it gets aerated and shut down, it air locks the pump the next time you start it. Fuckin' great.
Well, no better time to go to a big boy trunk mounted tank because racecar, right?
|April 1st, 2019 09:33 PM|
Changed the oil to some nicer synthetic. Oil wasn't glittery.
Still need to add the sight glass but I have it disassembled and drained for it.
I was looking at the washer reservoir and the chances of modifying it look slim. I might just remote mount it in the trunk. Which begs the question of what to do. Let's hear some input. Do I...
1.) Do things the right way and buy a $13 universal reservoir and pump
2.) In the true spirit of the build go digging at the junkyard for something to stuff in the back.
3.) Keep things sketchy and install a pump into a plastic ammo can.
Forecast is looking good for the weekend. Should be a fine time to get the car out, maybe get a little pictures and video.
|March 28th, 2019 03:41 PM|
Weather cooperated. Mostly. Only got a short run before rain set in, but I got it out. Got the data I was looking for, despite Livelink malfunctioning at every turn it seemed like. Not perfect but it's doing what I expect. It transitions to the "idle" portion of the idle speed mode sooner when coming to a stop, rather than remaining in "dashpot" mode. What I didn't expect is that it still has dashpot control, but with idle feedback. Kind of strange, but that's how Ford programmed it. Basically I just need to crank down the dashpot decay rate in the region I got data for...a lot. I've already reduced it to a fraction of what the base calibration for a 2V is (CVAF1), and it needs to be lower. From what I've heard this is expected, and given how little rotational inertia it has, it isn't a surprise either.
Now once you have a stupid car like this out on the road, everything is going well, and you got the data you need, you just have to play with it, right? I indulged myself and did a few little pulls just to get a feel for it. I got it rolling around 10mph in 1st, and gave it about 3/4 throttle up to 5000 RPM. Lots of turbo noise. Pulls up the nose pretty good, but not violently. AFR seems pretty well in check. It picks up about 30mph pretty quick in that little pull, but doesn't feel that fast. It makes me feel like some ricer in a Honda making all this turbo noise but not feeling like I'm going that fast. It could be deceiving, I don't know. I made a few part throttle pulls last year targeting 5-8psi, getting into the low 400s and they felt slow as hell.
Now I'll be curious to get some data on those hits, see what estimate for power I'm making, and get a look at the boost gauge to see if it's making as much as I expect. Whatever it is I want more. I don't want to put more timing to it. If it's making 15psi I don't want to put more boost to it either...
Weather is going to be a bitch this weekend so I doubt I'll get any more action on it. Might as well get some minor work in like putting in a new sight tube for the IC reservoir, and making a washer fluid reservoir that fits, since it's previous real estate has been claimed by a giant air filter. This is a 100% street car and I won't be giving that up.
|March 27th, 2019 12:21 PM|
Originally Posted by WickedSnake00 View Post
|March 23rd, 2019 10:29 PM|
The dumpsterfire awakens from its winter slumber.
Prepared everything for some datalogging tomorrow. Fresh tune with some tweaks to driveability loaded, new datalogging file to watch dashpot and idle behavior. Checked all the fluids. I expected it to be low on some, especially given I saw a puddle under it earlier. Opened the intercooler filler. Spilled everywhere. Guess it was full. Cracked open the crossover fill. Still held pressure from damn near 6 months ago! Fuel pressure is good. Electrical works.
Tomorrow is for datalogging. If/when I'm happy with the tip out and return to idle behavior, next big step is upping the fuel system to the 25psi capable one. Sumped tank, fuelab external pump, deadhead return system, all new lines, 95 lb/hr injectors. Then we can see what KILL MODE is like.
Minor project: Making a sight glass for the intercooler reservoir during the week.
Let's hope weather cooperates.
|February 17th, 2019 06:11 PM|
Shit's still too cold to get much done.
Got some minor work done. Put on a set of quik-latches on the side of the bumper so I can remove it in a matter of seconds, because god knows that I'll take it off and on enough times. Shoutout to y2k02 for that tip. I retained the original studs on the bumper for alignment, and drilled an additional set of holes through them and the bumper to hold it in place.
Put on a set of hoodpins. All black hardware. Now if you're like me, you'll quickly realize that the standard 4" hood pin kit isn't long enough for your fuckin hood. Just great, right? The only option longer than it is a chrome cotter pin kit from Moroso, and I'm not a fan of the look. But my dumb ass stumbled upon something great. Allstar makes a set of hood pin extensions, P/N ALL18524 for the 1 3/4" ones that work pretty nicely.
Anyway, here's wonderwall (with black hardware).
|This thread has more than 20 replies. Click here to review the whole thread.|