The Project Pad

Lol @ supervising the supervisor. Race gas is better than crack. I really wish there was an elegant way to do a dual fuel setup where it transitioned to 100% race or e85 at high boost. I would ditch the meth and put a 5 gallon fuel cell in the trunk.

It would be easy to do with software and 10 injector drivers. I think you could so it with 034 scaling and trim for each injector, I just don't like the idea of batch firing 5 injectors with a single injector driver. Now, of we could re-purpose unused ign drivers for fuel injectors..

audifreakjim wrote:Lol @ supervising the supervisor. Race gas is better than crack. I really wish there was an elegant way to do a dual fuel setup where it transitioned to 100% race or e85 at high boost. I would ditch the meth and put a 5 gallon fuel cell in the trunk.

It would be easy to do with software and 10 injector drivers. I think you could so it with 034 scaling and trim for each injector, I just don't like the idea of batch firing 5 injectors with a single injector driver. Now, of we could re-purpose unused ign drivers for fuel injectors..

Ben's 90 is running 10 injectors on 8 drivers; the 5 big injectors are batched off of three channels. 2 and 2 and 1 fired evenly 240deg out IIRC. I put a thread on Motorgeek about setting that up, it works really nicely. It would work even better if all of your injectors were the same size/characteristics, it would make dealing with accel enrichments and such much simpler to tune.

Sam

Oh yeah, don't do clutch drops with your rear torsen.



After work on Friday I was doing a little tuning. A bunch of coworkers were standing in the lot and one of them is all 'do a burnout'. I have a super fast car that could smoke all four tires with a little prompting, so why not, right? Well, I haven't launched a quattro car since I was 17 or 18, so it didn't go well. First attempt was foot-to-the-floor and dumping the clutch when the tach hit about 5k. Nothing much happened; engine bogged and car crawled away. Let's try again. Foot-to-the-floor, two hits off the 7800rpm limiter, LET'S RIP IT!!!!

The car moved a bit, then horrible noises, and not much movement. Sweet, killed an 01A, what a pain. Nope, tried second and it was the same deal, then people yelling that I nuked the rear diff. This was when I remembered people always say that Torsens shit the bed with high power launches. I get out of the car and it looks like it had the hershey squirts for about 40 feet. The drain plug and a variety of quarter-sized broken pieces of metal were littered all over the ground leading to the car.

Bmay and I pushed the car into the shop and installed a stock diff and axles in about 90 minutes. The old diff was WRECKED. It looks like a/several spider gears broke, wedged themselves against the diff and housing, and destroyed everything. There was a huge chunk of the housing broken out, and everything inside is heavily damaged in some way or another. Coolest failure I've had to date.



The stock diff sucks. The torsen made a huge difference in traction and handling, I can't wait to find another one. The Wavetrac would be even better, but not in the budget in the short term.

Sam

Oh yeah, a video of post-diff-replacement testing:

[youtube]http://www.youtube.com/watch?v=Or9aaFKnxpc[/youtube]

Windows Movie Maker says 5.0 sec 60-100 including a lazy .8sec shift. If I want to screw around, I might try powerbraking in 4th and see how fast it does it then. If I can cut .5sec from the time, that puts me in some pretty serious supercar territory.

Sam

diff don't like that

At first I was like

Then I was like

Nice carnage sam! Way to let the Monkey out!

So, as long as you respect the torsen, Is it a worthy mod?

Yes, one of the best mods I've done if you like going fast around corners. If you just dink around on the street it isn't a huge deal (except powerslides), but if you track the car it is a must-have.

Sam

awesome, Im totally doing it. is it pretty much bolt in, or is there machining involved? if you explained in your thread previous, just point me to the right page. no need to be redundant for my benefit.

thank you sir

The diff itself basically drops in to the stock housing. Check preload-backlash to be safe but of the few I've talked to that have done it there's usually no need to reshim anything.

The trick comes with the output shafts/axles. The torsen uses longer splines to engage the middle gears, so it has unique drive cups that only come in the 108mm size. I used the stock V8Q cups along with B3 90 FW 5spd axles, flipped left-to-right. OEM S2 rear axles would work as well The passenger side is the exact right length, the driver's side ends up being about a half inch longer than stock, so you have to be careful with your alignment and ride height so you don't bind anything. I also know people who have welded a big bead on the V8 cups then machined in a 100mm flange. You might like that option given your capabilities. Also note that you'll need a new driver's side output shaft seal, from the V8Q, with the V8 cups.

Sam

Alright, let's do this.



Two track events in the next 4 weeks. The first one would be pretty optimistic from a car and budget standpoint. The second is more probable. I had some time to kill in the shop yesterday and wasn't feeling up for anything constructive, so I removed everything from the back of the Coupe. I figured it can't get much noisier than it is now, and nothing there was in great shape anyways, so what the hell right? With the rear seat, seat belts, and floor carpetting, I'm guessing somewhere between 50-60lb in all was removed, which would put the car below 3000lb without me and with fuel, IIRC. I kept everything around in case the sound drove me nuts, but so far it isn't bad. I think I got rid of a few squeaks and rattles in the process, lol. Most of the noise is from the fuel pump, plus some glorious exhaust sound at full tilt. Long drives will definitely involve ear muffs.

I'm ordering some caliper brackets shortly and will get that project out of the way. I need to finish the seat project, finally. Then I need to get tight with some sheetmetal and make some ducting to try and address the cooling concerns. If I'm looking ok after that I'll register and worry about making sure I'm safe to run at low boost.

Sam

SEStone wrote:Suspension: Whoever was spouting on the internet that you shouldn't run sway bars on a B3 is full of shit.

Clearly very different (well, AWD, open diff, I5 ) but I was reading about the K-PAX S60 and thought of this thread...

http://www.motoiq.com/MagazineArticles/ ... o-S60.aspx

The K-PAX Racing rear subframe allows for a full spherical bearing conversion and has corrected a lot of the ride and NVH compromises the S60 had when it rolled out of the factory. Moving the rear differential up allows for reduced angles on the axles, which can contribute to roll resistance in a non-linear fashion (not to mention reduce component lifetime). The rear differential housing is fabricated in house and stuffed with an Emco limited slip and either a 3.752 or 3.974 final drive, with the front differential remaining open per the rules. As a result of the front weight bias and open front diff, the S60 uses a huge rear sway bar and no front sway bar as wheelspin was excessive with the front bar installed. Easily accessed connection points for the control arms allow for quick and easy adjustment during practice and test sessions, and the integral jack point is a must with a car with extensive underbody work. The brick located near the differential is there for weight ballast.

I dunno Sam, I think power plays big into the front sway equation. If you lift a front wheel off the ground with an open diff wheels spin causes a lot of time loss. If you have no sway and the rebound/spring rate can keep a wheel on the ground, then you have a better chance of keeping traction despite having more body roll.

One thing is for sure, it sure is spooky driving a car with no front sway. I think i am going back to one the next revision.

KPax probably had to do what they did because of rules--they had to do a band-aid. Band-aids keep blood from falling out but they're not as good as skin.

Given my only rule is that I couldn't spend more than about $500 to fix shitty handling traits, a pair of sway bars seems to be working nicely so far on my issue of regulating suspension travel without impacting balance or traction. I need to find another torsen to make it an apples-to-apples comparison at the next track day, but it seems to be achieving what I needed keeping the struts out of the bumpstops. The other solutions; Koni DAs ala Haydn or buying some KW V3s, has been thought about. I want a car that's a single color first, though.

If I was really resourceful, I'd drag the shop truck and trailer out with all our mobile alignment stuff and corner scales, do a bunch of testing in the current config. Strip off the bars, test again. Raise the car up an inch, and test again with the same two permutations. If my coilovers weren't seized it would be a shitload of work but entirely feasible to carry out.

In other news, I ordered some brake brackets from Hank. Should be here in the next couple of weeks and I'll finally get that box of parts onto the car.

Sam

hey sma if you want to get me the dim's I can plug them into the rollcenter calc program, lmk

I'm with you! Same shade of red and maintenance is above V3s on my list too! Wish I bought them a while back as they keep getting more expensive. I've thought about contacting Fat Cat Motorsports as an alternative but, again, waiting till I have cash in hand.

Make no mistake, I was not trying to dig on your setup just reading about an AWD car with an open front diff and no FSB made me chuckle

Mcstiff wrote: an AWD car with an open front diff and no FSB made me chuckle

chuckle all you want no front sway here

my2000apb wrote:

Mcstiff wrote: an AWD car with an open front diff and no FSB made me chuckle

chuckle all you want no front sway here

I have not had one on in a while

Well, my comeback would be to run stiffer spring rates with no front bar--500-600lb/in or something. Both for stiffness and suspension travel reasons. But the regular Koni sports can't handle rates that stiff without blowing, and are too long anyways.

Sam

I have 400/600 springs on my 4000 and no sways anywhere... car is very level through the turns and I haven't had any issues. This is w/ torsen center and rear diff as well.

Maybe your driving funny

other than looking like a tonka truck b/c of the bigger wheels/tires...

http://www.eviltwinmotorsports.com/wp/w ... 2011.5.pdf

More interesting food for the no swaybar nerds :p

Sam

great read thx

Giant life changes update.

November 1, my girlfriend and I moved from Minneapolis to Phoenix so I could take a job in sales with Evolution Motorsports (www.evoms.com). Both of us having grown up in Minnesota, this was a pretty big step.

I brought both of my cars with me on the 1900 mile drive down. My dad drove the GTI down, and I spent 26 hours over two days in the Coupe. It performed flawlessly, getting 23-27mpg depending on elevation and wind direction, doing basically 80mph the whole way. I got closed loop EGO working well before I left, and it worked really nicely keeping efficiency up and everything running well even at 7500ft through CO and NM.

Prior to leaving, the Coupe got lots of love to prepare it for the trip and new environment. I went from running 102 octane to 93 octane to 91 in the span of a month. 91 is such garbage fuel. I'm only running 250-260kPa, and bottoming out at 8-10deg of timing. I'm probably down 80-100hp over what I was doing with the 102. Nonetheless, I did a lot of work on the tune to get it safe and running well. Three days before we left MN, I was driving back from my parents' at about 9pm at night. It was in the 40s and the car was running awesome. I floored it from a slow roll in second gear and destroyed another 01A. I drove the rest of the way home, and put out a bat signal on Facebook around 10:30. By midnight I was back home with a sedan trans in the back of the car. My dad and I swapped it in the next day in about four hours. I had also melted one of the rubber-coated washers on my turbo coolant lines, so those got replaced and v-band gaskets were installed on the downpipe and exhaust connections to try and improve the fit and seal.

After the drive to Phoenix, I had to do quite a bit of work redoing my tune at low load. We were at about 600-700' in MN, and we're something like 1100-1300' here. Between the altitude and a difference in fuel used here, all of my super low load stuff went pig rich and had to be redone. I found I'm able to hit much lower manifold pressures under light throttle here, and everything approaching it is a bit richer. I was also under pressure by the emissions laws here, as if I register this car it will have to pass a sniffer test. Having worked so hard over so long to get the car running great in MN, I was not happy to have to go through that process again.

I made some quick adjustments to get the car smooth and driveable, but still struggled to get it as nice as I had it before. I had very large and very small basic map values very closed to each other, which just seemed like the really difficult way to be getting accurate fueling. Gears in my head started spinning. We're ID dealers at EVOMS, and recently started using the ID725s in our 996TT builds instead of the old Siemens 60lb. Once we had characterization data for the new injectors (which ID provides in detail), you plug it into the tune and the car figures it out and runs exactly the same as it did before. Fucking simple. I start looking through this data and some of the fantastic tech articles on the ID blog and website, and start thinking more and more about this...there has to be an easier way. And so began the NYE of discovery. It helped that my girlfriend was out of town and I had nothing else to do for like five days straight.

The first realization comes from Ford. They have a patent that describes an engine management technique. It posits that for a given RPM, air mass consumed is proportional to manifold pressure. The slope and intercept of the curve varies with RPM due to harmonics and such, but that basic relationship is linear.

The second realization was about how that pertains to the 034 fueling model, where you have a maximum MAP value, a main scaler, and an idle adjust. This is fueling based on the same principle, and it's genius. The slope of the air mass (and subsequently fuel mass) vs. manifold pressure relationship is the main scaler, and the intercept is the idle adjust. Dialing those values in sets up your fuelling for any manifold pressure the MAP sensor can see, and the basic map should allow you to tune this based on RPM characteristics in nice, vertical columns. You don't have to plot a VE map cell-by-cell to make sure your engine is appropriately fueled, unlike many standalone systems.

The first problem with this great model is that you have to bake your lambda value into the map as well. Admittedly this is quite easy, but it's something else that is much easier to understand and modify in proper OE engine management when it's in a separate map. The second problem is that you also have to bake quite a lot of injector behavior into it as well, which is totally wrong and makes the whole tuning process more difficult and less successful than it should be--and is the real area, thus far, that I've found standalone will never compare to proper OE management. I've currently got software for four or five <$4000 standalone systems on my PC, and I've found none of them model fueling behavior as nicely as GM/Ford/Motronic does.

Armed with that information and the 034 fueling formula (distilled down as pw = main scaler * (map/maxmap) + idle adj by thearchitect on Motorgeek), I came to realize that I could maybe develop a base tune using calculated data. My first step was to try and find ID-style characterization data for my 80lb Siemens injectors. I found my injectors are discontinued, and no public data existed. Without knowing fuel mass at a given pulsewidth, I couldn't know air mass. But since 034 does not use fuel or air mass anywhere, or directly a fuel mass:pulsewidth relationship, and using the air pressure:air mass relationship, I devised another plan.

With a reasonably good tune, with the car holding basically 1 lambda under 100kpa from idle to 4000rpm, I went and drove around logging for about 15 minutes. I took the data and stuck it in a pivot table. I looked at average injector pulsewidth in terms of manifold pressure lambda, and filtered out the data to be only looking at pulsewidths that delivered 1 lambda, since right there is a fixed 14.7:1 air mass:fuel mass ratio. When I plotted manifold pressure vs. pulsewidth, I got this:


Sure looks like an injector flow curve to me (see this for comparison: http://injectordynamics.com/wp-content/ ... ID1600.gif)! To recap, that curve emulates a fuel mass vs. pulsewidth curve like you'd get from ID, only with different units. There are fixed relationships between pulsewidth vs. fuel mass (dictated by the injector behavior at a given fuel pressure and voltage), fuel mass vs. air mass (14.7:1), and air mass vs. air pressure (unknown, but linear relationship) that allows that curve to be used here. The key thing to note is that below about 1.8ms, the fuel flow from the injector changes constantly with reducing pulsewidth, and changes radically below about 1.5ms. This phenomenon is why it is tricky to tune small engines with large injectors at low pulsewidths with oversimplified injector models and limited map resolution.

Armed with that chart, I started postulating a couple of different ideas on how to tune the car. I transposed the axes to be pulsewidth vs. map, and generated a formula that would use the main scaler and idle adjust terms that 034 uses.

Series 1 is pulsewidth as logged, including the low load 1 lambda pulsewidths plus non-lambda-normalized full load pulsewidths. Full load is running .78 lambda.
Series 2 is a calculated fuel curve without tuning based on a user-input main scaler and idle adjust.
The thin black line is a line-of-best-fit generated on the logged injector data by Excel

My first trial was to try tuning the car using the line of best fit. In theory, this should've gotten me the closest and generated the smallest scaler values over the entire map. I believe this gave me a 17-18ms main scaler, and a -.325 idle adjust. It required a lot of fine tuning in all parts of the map, since trying to average the whole map meant nothing was close. Once I had the map scaled in portions, under certain loads the car drove really, really smoothly. Like from 80-150kpa were very consistent and easy to tune...no coincidence as if you look at the chart above, the injector slope parallels the line of best fit in that range. Above and below that, the error got quite high. I still had to run 1.2 map values at full throttle, which was fine because it was linear and there are lots of cells to play with up there. Low load was very difficult, as I ended up cruising in an area where there was a .3 gradient between the 59 and 34kpa cells...not enough resolution to get it right, the car was either at .9 lambda or 1.1 lambda. I canned that theory.

The next theory I rolled with, and ended up pursing fully. This is displayed by the red line on the chart. I essentially set up my main scaler and idle adjust to most closely follow the 'low slope' of the injector curve. This gave me a 12.1ms main scaler and .510ms idle adjust. I have achieved reasonably good results with this. The car cruises great at a consistent 1 lambda up to 140kpa. It idles at a rock solid 1 lambda, varying .04 lambda for slight variations in load or RPM. My basic map is around .8 at idle, varies from .87 to .94 at various loads around cruise, and doesn't go very rich or lean for most spots. There are a few load/RPM bands where moderate 80-100kpa acceleration will result in swings from .93-1.07 lambda momentarily, but with the resolution I have I cannot get rid of them entirely. The car drives smooth and there are no obvious sags in power...trying to really fix those spots are exercises in optimization of efficiency, power, and emissions.

Ultimately, it's still a poor substitute for a real injector model. The Ford characterization would be the simplest to build-in: http://injectordynamics.com/articles/fo ... erization/. I believe I've determined the terms and order-of-operations that would have to be used to get the twin slopes and breakpoint working. Not sure if it will ever happen, but it would make this all so much simpler.

Hope this was helpful and provides guidance for others down the road.

Sam

Jeez, 6 months since last update.

I threw it on the dyno at work sometime this year. On an untouched 100oct tune with light ignition misfires and a forced 6500rpm redline due to a trigger issue, it did 425whp and 420wtq. I was very low on fuel that run, and unable to diagnose why the car cut out at 6500, I went with 'out of fuel'. Filled up with 91 octane while strapped down and went to work on that tune. First run was 290whp and the car was off-the-scale rich. It hit an IAT cell that was unnecessarily high. 30min of work fixing that, leaning things out, and adding timing got the car to 375whp and 375wtq before I pulled the plug due to the cut and increasing frequency of ignition misses.

After looking around and talking with Ben K (600whp+ Holset 90TQ), I went to Autozone and picked up a set of $25 build yourself Chevy 454 plug wires. I formed them up and installed them and fixed the ignition issue. Trigger problem still exists, and I need to take apart the front of the car to get that sorted as I believe it may be related to a crusty trigger wheel.

While all of this was going on, I made some seriously awesome local Audi car friends in Val and Savannah Ivanitski and Jason Trumpio. I got to know the Ivanitskis from a roundabout exchange involving brake pads for the Coupe, which evolved into 'my CIS UrQ runs funny after I put it back together' and stopping by to take a look with that. Val has a Mars Red '83 UrQ that had seem some better days after years of winter driving and ice racing. He decided he was going to do a bunch of bodywork on it including rust repair and respray, which I took keen interest in as owner of a three-color car. I went over and helped him out a handfull of nights, as he cut out previous patch jobs, rusty panels, and uncovered what was underneath several layers of paint and standard UrQ bodywork. Finally, we bondo'd the car up, sanded it a bit, then painted it in his garage-turned-paint booth. It was a very enlightening process, as I'd never participated in bodywork like that before. It instilled the idea about what that is like--before it gets better, it must get worse; bodywork is very messy; and the results largely reflect how much time you have to put into it. Even in somewhat of a rush as Savannah was like 8.9 months pregnant at the time he painted it, the results were fantastic for his $ investment and will get even better after a little bit of time wet sanding and buffing the car.



Anyways, between the prodding of my new friends and the obscenely long and expensive list of mechanical upgrades and fixes I feel like the car 'needs', I decided to pick my battles and focus on bodywork as the next major focus for this car. Gotta look cool for BBQ5, right?

My first step was to install the straight driver's side fender and door skins I have, replacing the mangled and creased pieces that have been on the car since I got it. The fender was easy enough. Replacing the door was not particularly fun, and a lot of work compared to the newer, easier-to-assemble cars I've worked on. You have to take the whoooollleeee thing apart on the car in order to strip the door skin down and finally remove it, then reverse the process. Everything is bolted and screwed in using multi-piece assemblies, and it's all adjustable. Adjusting fit at the hinges and latch is easy enough once you understand how to do it and what changes affect what dimension. Just be aware that due to the length, heft, and flexibility of the door, it's pointless to adjust everything without at least the window regulator installed, as the door will sag more and the regulator has a multitude of fitment requirements on its own. After a ton of screwing around, I was eventually able to get the door lined up perfectly, opening and closing with a finger weight on the door handle. I put the fender on along with it, which lined up perfectly with the door and headlight straight away (still need to readjust the hood), and suddenly the car looked about a million times better, even with one more color added.



I went over to Jason's house and we buffed out the oxidized factory tornado red paint. I was hoping for an exact match with the resprayed unibody, but the door is brighter in all but direct sunlight.

With that, I'm at a juncture of deciding how deep I want to get into this--whole body strip and respray, or just paint the front clip red and call it good? There are enough defects in the rear of the car that going all the way would give me appreciably better results overall, but that is much more work as well. Following my experience with Val's car--it's not the painting I'm worried about, it's all the prep work. I'm hoping to learn some things about how the side windows and trims come apart (and if they can be put back togehter) in the next couple of weeks that will be important in deciding which direction I go.

Sam

In for a penny, in for a pound... my vote is to do it all and be done with it.

At any rate, if you do remove the rear windows there's no better time to install the euro-spec rear opening windows (basically vents). Would come in handy with that Arizona heat.