Lucifer's Hammer Part 10
Posted: Jan 11, 2007 11:33 PM
Lucifer's Hammer Series Index: Part 1 • Part 2 • Part 3 • Part 4 • Part 5 • Part 6 • Part 7 • Part 8 • Part 9 • Part 10 • Part 11 • Part 12 • Part 13 • Part 14 • Photos
Part 10.
In which we attempt to break the motor, find that this churn can use a lot more huff-and -puff, and get some results that are a bit out of hand.
Justin and I have gone over our collective notes, driving impressions and the first collection of dyno pull results. The fuel delivery system is going to get some scrutiny; the dyno pulls hint that All May Not Be Right with the turbo. As in Part 9, the "LH page and image numbers" refer to graphite's website: http://www.doggunracing.com/mye28/LucifersHammer
10/5/05 The brake bomb has been replaced. Relatively easy to access with the intake plenum pulled. Dave has lent me a known-to-be-good ABS control unit. The TBs have been tapped for individual vacuum fittings The secondary fuel rail is in. Test fitting looks like everything will fit. Snug, but they will fit. LH, page 6, image #148.
The existing fuel lines get pulled, and are replaced with runs made from .625" Type K soft copper tubing. This provides a non-restricted -10 size feed from the fuel pumps all the way thru the rails. The only restriction is the injectors themselves.
The post-regulator return is replaced with a .750" copper line. This connects to a "dump" pipe inside the tank, emptying fuel at the rearward end of the tank. By keeping the return separated from the fuel pickup points, any aeration that might have occurred is minimized before fuel is fed to the pumps. It also allows the fuel in the tank to act as a heat sink for any BTUs gained during delivery. LH, page 2, images #42 thru 44; page 3, images #52 thru 56.
As outlined in a previous chapter, the fuel delivery will be via a pair of Aeromotive A-1000 pumps, one feeding each fuel rail, and drawing from separate pickups. The pump's operating speeds are managed from separate controllers. LH, page 3, image #57. These allow the pump speeds to adjust based on fuel demand, while keeping pressure sufficient to match demand under boost or rpm. This approach avoids overfueling at low engine load levels or when boost has not risen appreciably, e.g., at cruise when boost levels are low, but rpm is up there (typically zero boost or in vacuum).
10/14/05 The oil sensor had been located in the "out" feed from the oil cooler. My error; my bad. This explains it's non-action. The sensor and its mounting block get relocated to the feed line from the engine block which solves the problem.
10/15/05 The upsized fuel lines mean using a larger volume fuel regulator, so the present brand new one gets swapped. The regulator lives in the upper RH corner of the engine bay, above the ABS and turbo control units. Getting to the ABS unit will mean disconnecting fuel lines, but nothing terribly involved. The turbo controller is still accessible, but crowded.
The IAC unit can go back to its original location on the left hand side of the block, the MAP sensor having been moved. Having test-fitted the TBs with the secondary rail in place, there is sufficient room.
10/18/05 Another software upgrade/patch from Electromotive to run the dual injector sets at peak-and-hold, full sequential. FWIW, this upgrade has been included in the more recent WinTec software releases. You're welcome.
Justin has been on the phone virtually daily with Electromotive, dealing with various bugs, receiving patches and so on, depending on what's turned up on the dyno pulls.
GRATUITIOUS BITCH AND GRIPE:
I gotta think I'm functioning as Electromotive's beta test site. This tends to not increase my comfort level, but I do appreciate the efforts Electromotive is putting forth to get this bad boy working correctly.
IMHO, all of the aftermarket ignition/engine management systems suffer from incompletely developed software and less-than-sufficient or less-than-lucid documentation. The only questions are to what extent, and the manufacturer's willingness to deal with the problems. Not a whole lot different from financial institution systems software in this regard. BTDT. Please excuse my cynicism, but I can speak to the latter with considerable authority.
11/22/05 The next set of dyno pulls are to prepare for emissions inspection. 5-gas analyzer sez Close, But No Cigar. HC is still high. The cats may not be getting hot enough to light off. The problem is 5 feet + of distance between the exhaust ports and the cats allows for a fair amount of heat loss, even with the tubing being ceramic-coated for heat retention going into the turbine. The next step may be to do sufficient highway driving to get the exhaust temps post-turbine up ovewr 800 deg. F, then do the test without letting the cats cool.
Justin talks with Paul about the problem. Paul suggests advancing the exhaust cam 4-6 degrees via the adjustable sprocket as a way to get some additional exhaust gas into the pipes and bring the cats up higher on temp. This will serve to retard the timing and allow more heat into the exhaust gas mass. Finally, the answer may be to pull the exhaust section after the downpipe and install a pair of cats located as far forward as possible to take advantage of the hotter exhaust gas. LH, page 7, image #151.
12/1/05 Revised exhaust section is substituted. Emissions is passed. WOOOOOHOOOO!
Numbers: HC 3.0000 grams/mile standard. Actual: 2.6663. Passed
CO 20.0000 grams/mile standard. Actual: 4.6390. Passed
NOx 6.0000 grams/mile standard. Actual 2.6274. Passed
The actual numbers are based on averages taken over the 240-second run on the emissions dyno. Our thinking is fuel delivery gets a bit of a "burp" as it comes off idle and gets a momentary enrichment. This pushes the HC number up, but net-net, the readings pass.
Justin is still a tad concerned about the relatively high HC value, but tells me that Electromotive is going to have updated software in about six months which should help bring the HC numbers down considerably. But for the present, no matter; a major worry has passed.
12/12/05 OK. with everything back in place, on to more dyno time with the "normal" BIN file. Findings aren't necessarily wonderful. No problems with the reworked fuel system; delivery is not an issue now. Run the boost up to 8 psi and things are OK. A/F is around 12.5 or so. Outputs: 535.5 rwhp @ 6100 rpm, 500.5 rwtq @ 5200 rpm. Note: these are rear wheel values delivered to the rollers; flywheel would be around 637 and 595, respectively, when drivetrain losses of 16% are figured in.
So we bring the boost on the GT-35R up towards desgn levels -- 15 psi. Things turn into shit. Boost will reach 14.7 psi around 32-3300 rpm, then begins to decay beyond 4200 rpm. A couple more pulls. 621.8 rwhp @ 5800, 627.8 rwtq @ 4900. Boost drops down to around 12 psi or so. EGT temps head for 1400 deg. F +. Pulls suspended forthwith. I know 1400 degrees doesn't sound out of line, but keep in mind there is five feet of header and crossover piping between the exhaust ports and the entry to the turbine; the EGT sensor is located at the feed into the turbine, so exhaust gas port temps are waaaay iup there.
12/16/05 Looking at the graphs from these pulls, we think the HP decline above 5200 rpm is showing that the GT-35R is not capable of delivering the needed air volumes at 14.7 psi boost. Justin estimates we will need to see corrected air volumes of at least 70 to 72 lbs/minute, with demands in the 80+ pound range under maximum load. Analysis using the not2fast model tends to support this. Running the turbo as presently configured is pushing the loading very deep into choke. This creates a stall condition in the compressor, and the intake air mass temperatures go through the roof. This leads to heat being transferred into everything related to the combustion chamber. Detonation, turbine failure, burnt valves and melted pistons are prompt consequences. So we stopped right now.
First solution considered was reworking the GT-35R compressor housing to accept a GT-4088 wheel. The T3 turbine can be replaced with a T4 divided housing with a .81 A/R. This means replacing the present mounting bracket platform, and reworking the downpipe to accept the V-band exducer configuration of the T4. These changes will give adequate airflow and move the 15 psi max boost point very close to the choke line, staying on the compressor's "efficiency island."
The .81 A/R turbine, combined with the larger compressor wheel should minimize turbo lag and still deliver the needed impetus to spin the 4088 impeller. We have considered going to a complete 4088 compressor housing, but there are space limitations, as the GT-40 "snail" will not fit between the block, wheel well and the plenum. Why the f*&k did i let myself get talked into FI?
12/20/05 A lot of research over the past few days. Analysis points to the GT-35R being about 33% too small in air mass delivery for the motor's capabilities. Justin thinks the answer is using a T66 compressor, with a ball bearing rotating assembly, a T4 turbine with a P trim exducer and the .81 A/R. The compressor maps indicate an air mass delivery of about 72 #. With a 2.2 Pressure Ratio, this puts the "point" just off the edge of the island, only slightly into choke. This is not ideal, but going with a large-frame compressor, has both space limitations, as well as bringing lag problems into the picture. The GT-40 and -42, on paper, would seem to be a closer match; the T66 is about as physically large as we can cram into the space available. Given that the amount of time at 15 psi boost and full throttle (6300 rpm) is going to be very small, the objectives of minimizing lag time and sharp mid-range (2800-4500 rpm) response point toward a smaller compressor. Plugging the T66 capabilities into the not2fast, we think the midrange flows can be effectively delivered and stay closer to the high efficiency percentage islands.
So we order the T66/T4 with water cooling in the CHRA and ceramic bearings. Turbo should be here just after Christmas. Looking at the information and compressor maps in Garrett's website, http://www.turbobygarrett.com, and in the sketches inTurbonetics, Inc's website, http://www.turboneticsinc.com the T66/T4 unit should physically fit, though an actual trial is going to be needed. The amount of work on the bracket mounting plate won't be excessive; clearances -- block, wheel well, plenum -- will also be "try it and see."
An observation: as we have researched various turbo configurations, the vendors and manufacturers do a really piss-poor job of making useful information available. One of the worst shortcomings is the near-absence of dimensioned drawings showing sizes, clearances or anything to assist making the engine builder's job any easier. When asked, the answer from the vendor is, "buy the (particular model) and see if it fits." Helluva way develop customer service to boost one's sales. Awful pun, that . . .
Assumingf the pieces are a 'go', the dowwnpipe will need reworking for a V-band flange. This is a dissapointment, as the stainless steel 4-bolt flange is now superfluous . . . the time and dollars, both considerable, are now lost.
12/29/05 The non-working ABS appears to be the result of a dead control unit. This was an original part on the car, and worked fine when it went to EB. I asked Dave why an ABS control box would fail. He tells me the units need to be removed, or at a minimum, electrically isolated or disconnected if any welding is to be done on the car. EB did some welding on the chassis to fabricate the IC mounting brackets. So who knows?
12/30/05 T66 is here from Turbonetics; the -35 is pulled. Definitely some fitting circumstances to be addressed. Additional clearance items are the cruise control servo and the oil line fittings at the block. Getting close-angle oil line fittings is probably going to mean a visit to Combs Aviation and paying thru the keester for trick aircraft-certified Aeroquip fittings. About ten times the price of race-grade stuff at Peterson Fluid Dynamics (the local Aeroquip thief).
The intake plenum is going to need a LOT of metal being removed. Happy New Year. 
1/5/06 The T4 exhaust flange is welded onto the crossover pipe feeding into the turbine. The mounting platform has been reconfigured as well. We have some worries about the welding and reshaping of the crossover pipe. The end of the pipe has had a lot of welding and bending done on it. There may be a problem down the road with the amount of heat stressing applied to the relatively thin 304 stainless tubing. This may make the area prone to cracking from vibration and heat cycling. Part of this is due to the large variation in thicknesses of the flange (.500") vs. the tubing wall (.060").
The answer may be to replace the entire crossover tube, with an appropriate circle-to-rectangle flare being fabbed out of an intermediate thickness strip of 304. Before jumping into this, I want to see how the present item holds up. In any case, the crossover and turbine heatshield are going out to Premier Coatings the first of the week for the zirconium porcelain heat barrier coating.
1/13/06 The T66 has been mock assembled and the bottom of the plenum hogged out for clearance. Plenum is ready to go back to Paul for shaping and fabrication of the concave insert and redoing the powdercoating. LH, page 7, images # 153, 154, 155, 164.
1/22/06 Pieces back from Premier and are installed. Paul advises the insert is in. He is 95% certain the clearances over the turbo will be sufficient and will clear the heat shield.
2/3/06 Plenum is here. Rework will allow clearance, but is very close to the turbine heatshield. Justin thinks there is an outside chance there may be enough heat radiation to discolor the powdercoating, but the risk is small. The alternative is to send it back to Paul for further work. This would mean 2-3 weeks delay. KH decides to proceed as-is; the issue is cosmetic, not functional and the porcelain coating on the heatshield should keep temps down.
2/21/06 Justin has the motor running, no issues at the lower boost settings. With the recent below-zero weather, work on the dyno has been backed up at Mile High. I'm #5 or #6 in line; possibly onto the rollers late this week or into next (more likely).
3/1/06 Onto the dyno. Initial pulls at 7 psi. ~ 570 rwhp @ 6000 rpm. Boost tried at 12 psi. Indicating 608 rwhp @ 13.2 psi and 5250 rpm. Indications the head gasket is leaking, so enough for one day.
3/3/06 We pull the head. Bad Shit, man. Gasket has definitely begin to fail. No evidence of detonation or other component damage or failure. The apparent reason for the gasket failing is the head lifting against the bolts. Why? it's a good question. New factory bolts. Properly torqued to spec. Cylinder pressures well within reason; certainly nothing other FI motors can support without incident . . .reference what TCD has been running on his M30's. We have a bunch of very upset people at the moment.
Rather than try for a repeat with the stock S38B38 gasket, or go through the drill with SCE, the choice is to contact Cometic and use one of their MLS gaskets. TCD has spoken highly of Cometic; they can provide S38 pattern units. So calls get made. Yes, they have S38B35 gaskets in stock. Yes, they can laser-cut a bore to almost any size we need. Yes, they can supply thicknesses between .071 and .140. We'll take one in .080 and with a hole sized to fit a 94.36 mm bore, TYVM. At $220, cheap at half the price.
will be here in a week. 
LH, page 7, image #166.
Why I didn't go to Cometic in the first place a year ago, I dunno. I didn't do sufficient homework and look at every possible parts source. As a result, it cost us at least a couple of months in delays and a bunch of $$. Live and learn.
Given what I've seen with the rest of Paul's workmanship, I doubt that he overlooked or mishandled anything to do with securing the head. But with the unknowns about how well the head was secured, the next item on the menu is to see about head studs, rather than bolts. VAC is advertising ARP head studs for the S38. This is news. $336 plus shipping. Should be here by Monday 3/13.
With the head off, we have a hard look at things. During the running-in and the dyno pulls, there has been a small amount of oil seepage on the header slip fitting for #3 cylinder. Nothing excessive, just a small dribble. Hard to tell how new it is; dosen't look like there was anything fresh after every drive or pull. Additionally, with the motor stone-cold, there has been a very brief puff of white smoke on initial turn-over. This goes away within a few seconds. During the run-in and on the dyno there has been zero oil consumption as referenced on the dipstick. No indications of any coolant loss by volume or any oil-coolant mixing. More on this in a bit.
Looking at the combustion chambers and valves, there aren't any indications of oil leaks, i.e., wet streaks in the ports or residue accumulation. Nothing wet in the chambers; plugs are dry. Exhaust tracts are sooty and there is some soft sooty deposit on the backs of the valves. This may be from the fact we have been running fairly rich on the A/F.
The valve guides are new, as are the seals. Paul doesn't like to run guide clearances real tight. He has concerns over different expansion coefficients between the stems and the guides. Given that in a turbo application the exhaust gas temperatures on leaving the combusation chambers can hit 1500-1600 deg F., he feels more comfortable leaving a bit of clearance, typically around .004" of radial displacement as measured on the edge of the valve face. This measurement is taken with the valve extended (top of the stem flush with the top of the guide). FWIW, the BMW TIS specifications manual shows that the tolerance on this measurement is .0315", so we are well within acceptable limits. So we are puzzled by where the oil seepage is coming from. The fact it goes away on warmup and there is no indication of blowby pretty much eliminates the rings.
As part of the detailed once-over, we find the valve stem seals on # 2 and #3 exhaust have leaked and are replaced. The head is checked for warpage and is dead flat.
3/6/06 while waiting on the studs, Justin goes about dealing with the hair-trigger clutch problem. This begins with reboring the clutch slave cylinder to 25.4 mm, from the stock 20 mm, machining a clone of the piston sized to the nerw diameter and including a second O-ring seal. The larger diameter piston should reduce the felt pedal effort and sensitivity at engagement. This won't lower the clamping force, but the idea is to make clutch engagement a less touchy proposition.
We aren't taking bets on how long the QuarterMaster is going to live. This has nothing to do with the quality of the unit, but rather its ability to deal with the power going through it. Alternatives may include a multi-plate Tilton, but just about anything that can survive the power delivery has street tractability problems.
Given the present horsepower deliveries, I think a good case can be made for bringing the upper boost limit down to around 12 psi.
The reasons are as follows:
1. The available power is more than adequate for any conceivable over-the-road application.
Higher numbers may be fine for bragging rights, but the dyno pulls to date raise significant reliability questions at the higher boost numbers.
2. Given the limitations of gaskets, head bolts, etc., things seemed to be pretty much OK at 12 psi.
To be noted is the psi dropoff in the pulls done in December with the GT-35R. This may have been when we originally damaged the gasket in attempting to run 15 psi boost.
3. There may be some question as to how much power the clutch and the rest of the drivetrain can tolerate. Justin thinks its limit is around 700 ft-lbs. That being stated, backing off on the raw power is simply prudent.
4. One of the primary considerations for the motor is longevity. Giving up the last 15% or so may well be the price. Not worth breaking the motor over the issue.
3/7/06 Paul has been kept apprised of what's happened. He is not happy; wants to know what we did to one of his motors. Justin sent him a bunch of phone camera pics. These I found in a file after I had sent the first bunch off to graphite, so not included here. Saaaarrreeee.
After looking at the bad news, Paul has a few thoughts.
1. When we assembled the 3-into-1 exhaust collectors, the slip joints were very tight., and received a healthy dressing of assembly/anti-seize lube to aid the process. The female part of the joint is on top. Paul thinks the weep stains may be from the assembly lube melting out and working it's way thru the slip joint and onto the pipe exterior. Next step here is to thoroughly clean the collectors and see if anything else reappears.
2. Paul was very surprised at the amount of gasket stretch and the fact that the #3 fire ring had moved as much as it had . . .nearly 8 mm towards the exhaust side of the block. He noted that the degree of deformation has pushed the gasket into the oil galleries in the block, most notably at #3, but the process is well under way on the other cylinders.
This may have permitted some leaking between the gallery back into the combustion chamber, thus creating much of the exhaust smoke we've seen and affecting how the thing was running.
The extent of the deforming points to more than one would expect from the time on the dyno. In other words, the problem may go back to the first 4-10 hours of operation. At that time, the car was terribly out of tune -- running pig rich, and ignition advance was not where it should have been. This could have allowed detonation; the gasket yielding as the weak point.
The gasket being pushed out into the galleries very well may have allowed some coolant leakage as well. We never saw any indications of gases getting into the coolant--smell of fuel, bubbles, oil slick in the expansion tank. Certainly no "milkshake." But early on during some of the initial operation, and during the over-the-road testing, there definitely was white exhaust smoke. Not a lot, and invariably at startup or on hard acceleration after a coast-down.
He thinks cylinders 4-5-6 are seeing pretty uniform cylinder pressures, based on appearance and his practiced eye. #3 is obviously 'way weak, with #2 not as bad, and #1 kind of in the middle. One thing we will need to do once back together and on the dyno is to have a look-see at the A/F ratios by cylinder. The TEC-3r can do this, and I am running a PLX wide-band sensor anyway. Paul suspects there are hole-to-hole mixture imbalances and things definitely will benefit from being leaned out.
3. Another possible answer on the oil seepage issue. Looking at the photo of the block with the gasket in place on the deck. On #3 it has pushed out far enough to where it intersects the rectanguilar opening of the oil gallery. The gasket has a very small tear in it next to the steel of the fire ring -- more or less at 6 o'clock on the fire ring -- looks like a dark brown vertical streak. This is sufficient (we think) to allow some degree of oil seepage into the cylinder on startup when the heat hasn't expanded the block/head/gasket sufficiently to offer a better seal. So for the first few seconds cylinder vacuum draws lubricant into the combustion chamber, thence out the exhaust. This passage also allows combustion byproducts -- soot, etc., to be forced into the lubricants in the course of operation. This may offer an explanation as to why the oil is (a) very dirty, almost grey-black after a couple of hundered miles, and (b) why the oil has a strong odor of unburnt fuel. Running 'way rich, remember? We have no indication of any weakness in the rings and the valve guides appear snug, save for the two weak seals.
4. Taking a second look at the head and #3 combustion chamber, in particular at the defined edge of the quench shelf by the exhaust valves. The edge of the shelf is ever so slightly rounded off. On the other CCs, this edge is crisp and well-defined. No penetration of the ceramic coating, however (good news). Paul thinks this is due to the "hydraulic" erosion caused by gas leakage caused by the gas leakage past the stretched fire ring.
I measured the head for warpage and deformation, paying extra attention to the area between the water jacket openings on #2, 3 and 4 cylinders. Done with a straightedge and feeler gauge, the surfaces appear flat and true; if there is anything out of line, it is well less than .001." This adds to the confidence level that the spate of problems didn't cause (or was the result of) any head warpage.
3/15/06 ARP head studs are here from VAC Motorsports; ditto the Cometic .080 MLS gasket. LH, page 7, image #166. Valve stem seals replaced on #2 and #3 exhaust valves. Exhaust headers and collectors pulled and hot-tanked to remove the hardened oil "dribbles."
3/16/06 The head studs are too short. When installed and properly bottomed in the block, they are .150-.200 too short with the washer in place. LH, page 7, image #168. I call VAC. Their answer: Put a short piece of rod in the bottom of the tapped hole in the block and bottom the stud on that. WTF? The length of the threaded portion will tell you how long or short to make the rod plug. WTF?? Paul advised of the situation. He reaches his contact at ARP, outlines the problem, communicating the needed dimensions. ARP's people do some research and come up with a solution. Seems there is a 12 mm head stud they made for an F1 engine builder (NOT BMW). There are a few pieces left over; not a cataloged item. Price? Youdoan'wannaknow. Done. They are on their way.
That evening, I went home, fixed three fingers of JD, and pulled out Corky Bell's "Maximum Boost. Looked up Corky B's comments on head fixation. CB strongly recommends using studs (duh), but also emphasizes being sure to bottom the studs in the block. This neutralizes any compressive or tension forces in the block or stud before tension is applied to the stud when the head is torqued into place. In any case, the length of the thread chamfer is definitely not the way to establish stud working height.
So why is VAC proposing putting a small piece of rod in the bottom of the block stud/bolt hole so the stud will bottom on that? This would appear to negate part of the attachment /securing design. This advice scares me, as VAC has been developing a pretty good reputation in the BMW aftermarket/engine rebuilding world.
3/22/06 Studs end up getting sent to Paul. He is forwarding them to Justin, so should have them by Friday. Per Paul, the length appears to be correct. Threaded portion will bottom in an M30 block, with about 1 1/2 threads showing. If shank length and head thickness dimensions are correct, there should be about 3/8" threaded portion of the stud protruding above the nut when installed. In other words, correct length. We shall see.
3/24/06 Studs here. Threaded segments as well as overall shank length significantly longer than the VAC offering. Shanks have a slight bulge above the threads which may limit how deep the threaded 12 mm portion can go into holes in the block. The bulge above the threads is a result of the threads being swaged, rather than cut into the body of the stud--a good piece of engineering. LH, page 7, images #169, 170.
We'll begin the swapout and reinstallation on Monday.
I'll be interested to see if the main body of the shank is too long, i.e., if the nut ican be screwed down to the beginning of the threads, even with the washer in place. Paul is trying to reach Justin with the ARP-secified torque settings; apparently the spec sheet was inadvertently not sent with the studs themselves.
3/25/06 Paul tells me the slight bulge above the block threads cleared the block counterbore on his M30 block, allowing the stud to bottom properly. I don't know if that same configuration is present on the S38. He understands my concern about where the nut will end up given the shank lengths. If needed, we may be able to obtain different thickness washers. Paul reminded me to be absolutely certain that a proper Hylomar seal is in place around the front of the gasket to forestall any oil seepage from the chain area when the head goes back on.
3/27/06 One of the new studs is test-installed. The bulge effectively limits screw-in depth, but the threaded shank is 1-2 threads longer, so we are good on purchase, the stud being able to bottom. The overall stud length has interference with the underside of the cam tray. Given the nut height, Justin is blanchard grinding the washers down somewhat thinner. S/b done later this week. This will allow full purchase of the nuts with 1 or 2 turns extending above the top of the nut. LH, page 7, images #171, 172, 173.
Justin has run Hylomar just about everywhere--around the timing chain area, the oil and water gallery openings, between the layers in the MLS gasket. This sucker is going to stay
dry. 8)
3/29/06 Washers ground, turned to fit head recesses for the studs/bolts. Studs installed. Cam tray is relieved on the mill near the intake valve spring portals to allow installation without shortening the studs. Cams installed. Paul had originally set up valves ~.013-.014 on intake, .016 on exhaust. This for break in. "A loose valve is a happy valve." With the motor having been run-in, can probably take up .002" or so.
Unused head stud set returned to VAC. They were not happy with what I described and it took some discussion before they were willing to provide a Visa credit . . .less a fat restocking fee. 
3/31/06 Valves adjusted .011-.012 intake, .014 exhaust. Throttle bodies in place and synched. New plugs.
4/7/06 (Playing the Darth Vader Theme from Star Wars)
Dyno runs completed. Low-boost pulls are done @ 7.25 psi. 565.3 rwhp @ 5700 rpm, 547.4 rwtq @ 4900. Justin does a few more changes to the BIN file (we are now at file #140 ), then leans into it. 707.2 rwhp @ 5700, 725.8 rwtq @ 4900. 
Both HP and TQ curves noticeably flatten out beyond 4900, as the boost drops off from the 14.95 psi peak. "You need a bigger turbo, Ken . . ." "Yah. Pigs fly." 8)
Holy Shat, Bitman.
Since some people have asked, the recorded numbers are delivered at the rollers. This is on a Mustang eddy-current dyno.
Further, the readings have been adjusted for altitude (barometric pressure), temperature and relative humidity. So what is shown here would be the same as a comparable pull done in, say, Silicon Valley, or Connecticut. If one measures the "actual" rwhp, the values are roughly 20 to 22% lower, or in the 550 to 565 rwhp range. Manifold air pressure is taken from a pressurized vessel (the plenum), so the boost level is the same, regardless of the test location's altitude.
What's the power at the flywheel? It depends on whose numbers you want to use for power train losses. If one refers to Puma Racing's website, http://www.pumaracing.co.uk, David Baker suggests adding 10 hp to the roller number, then dividing by .88. This gives fwhp of around 815 hp. I might add that this is a must-read site of you are trying to become eddicated about performance engines, BTW. Plugging in known engine parameter values into not2fast, I can come up with 842 hp. Another method is to assume a 16% drivetrain loss. This yields 707/(1-.16) or about 853 fwhp. In any case, it is more than you want.
The dyno pulls don't end the story. There's more to come.
Part 10.
In which we attempt to break the motor, find that this churn can use a lot more huff-and -puff, and get some results that are a bit out of hand.
Justin and I have gone over our collective notes, driving impressions and the first collection of dyno pull results. The fuel delivery system is going to get some scrutiny; the dyno pulls hint that All May Not Be Right with the turbo. As in Part 9, the "LH page and image numbers" refer to graphite's website: http://www.doggunracing.com/mye28/LucifersHammer
10/5/05 The brake bomb has been replaced. Relatively easy to access with the intake plenum pulled. Dave has lent me a known-to-be-good ABS control unit. The TBs have been tapped for individual vacuum fittings The secondary fuel rail is in. Test fitting looks like everything will fit. Snug, but they will fit. LH, page 6, image #148.
The existing fuel lines get pulled, and are replaced with runs made from .625" Type K soft copper tubing. This provides a non-restricted -10 size feed from the fuel pumps all the way thru the rails. The only restriction is the injectors themselves.
The post-regulator return is replaced with a .750" copper line. This connects to a "dump" pipe inside the tank, emptying fuel at the rearward end of the tank. By keeping the return separated from the fuel pickup points, any aeration that might have occurred is minimized before fuel is fed to the pumps. It also allows the fuel in the tank to act as a heat sink for any BTUs gained during delivery. LH, page 2, images #42 thru 44; page 3, images #52 thru 56.
As outlined in a previous chapter, the fuel delivery will be via a pair of Aeromotive A-1000 pumps, one feeding each fuel rail, and drawing from separate pickups. The pump's operating speeds are managed from separate controllers. LH, page 3, image #57. These allow the pump speeds to adjust based on fuel demand, while keeping pressure sufficient to match demand under boost or rpm. This approach avoids overfueling at low engine load levels or when boost has not risen appreciably, e.g., at cruise when boost levels are low, but rpm is up there (typically zero boost or in vacuum).
10/14/05 The oil sensor had been located in the "out" feed from the oil cooler. My error; my bad. This explains it's non-action. The sensor and its mounting block get relocated to the feed line from the engine block which solves the problem.
10/15/05 The upsized fuel lines mean using a larger volume fuel regulator, so the present brand new one gets swapped. The regulator lives in the upper RH corner of the engine bay, above the ABS and turbo control units. Getting to the ABS unit will mean disconnecting fuel lines, but nothing terribly involved. The turbo controller is still accessible, but crowded.
The IAC unit can go back to its original location on the left hand side of the block, the MAP sensor having been moved. Having test-fitted the TBs with the secondary rail in place, there is sufficient room.
10/18/05 Another software upgrade/patch from Electromotive to run the dual injector sets at peak-and-hold, full sequential. FWIW, this upgrade has been included in the more recent WinTec software releases. You're welcome.
Justin has been on the phone virtually daily with Electromotive, dealing with various bugs, receiving patches and so on, depending on what's turned up on the dyno pulls.
GRATUITIOUS BITCH AND GRIPE:
I gotta think I'm functioning as Electromotive's beta test site. This tends to not increase my comfort level, but I do appreciate the efforts Electromotive is putting forth to get this bad boy working correctly.
IMHO, all of the aftermarket ignition/engine management systems suffer from incompletely developed software and less-than-sufficient or less-than-lucid documentation. The only questions are to what extent, and the manufacturer's willingness to deal with the problems. Not a whole lot different from financial institution systems software in this regard. BTDT. Please excuse my cynicism, but I can speak to the latter with considerable authority.
11/22/05 The next set of dyno pulls are to prepare for emissions inspection. 5-gas analyzer sez Close, But No Cigar. HC is still high. The cats may not be getting hot enough to light off. The problem is 5 feet + of distance between the exhaust ports and the cats allows for a fair amount of heat loss, even with the tubing being ceramic-coated for heat retention going into the turbine. The next step may be to do sufficient highway driving to get the exhaust temps post-turbine up ovewr 800 deg. F, then do the test without letting the cats cool.
Justin talks with Paul about the problem. Paul suggests advancing the exhaust cam 4-6 degrees via the adjustable sprocket as a way to get some additional exhaust gas into the pipes and bring the cats up higher on temp. This will serve to retard the timing and allow more heat into the exhaust gas mass. Finally, the answer may be to pull the exhaust section after the downpipe and install a pair of cats located as far forward as possible to take advantage of the hotter exhaust gas. LH, page 7, image #151.
12/1/05 Revised exhaust section is substituted. Emissions is passed. WOOOOOHOOOO!
Numbers: HC 3.0000 grams/mile standard. Actual: 2.6663. Passed
CO 20.0000 grams/mile standard. Actual: 4.6390. Passed
NOx 6.0000 grams/mile standard. Actual 2.6274. Passed
The actual numbers are based on averages taken over the 240-second run on the emissions dyno. Our thinking is fuel delivery gets a bit of a "burp" as it comes off idle and gets a momentary enrichment. This pushes the HC number up, but net-net, the readings pass.
Justin is still a tad concerned about the relatively high HC value, but tells me that Electromotive is going to have updated software in about six months which should help bring the HC numbers down considerably. But for the present, no matter; a major worry has passed.
12/12/05 OK. with everything back in place, on to more dyno time with the "normal" BIN file. Findings aren't necessarily wonderful. No problems with the reworked fuel system; delivery is not an issue now. Run the boost up to 8 psi and things are OK. A/F is around 12.5 or so. Outputs: 535.5 rwhp @ 6100 rpm, 500.5 rwtq @ 5200 rpm. Note: these are rear wheel values delivered to the rollers; flywheel would be around 637 and 595, respectively, when drivetrain losses of 16% are figured in.
So we bring the boost on the GT-35R up towards desgn levels -- 15 psi. Things turn into shit. Boost will reach 14.7 psi around 32-3300 rpm, then begins to decay beyond 4200 rpm. A couple more pulls. 621.8 rwhp @ 5800, 627.8 rwtq @ 4900. Boost drops down to around 12 psi or so. EGT temps head for 1400 deg. F +. Pulls suspended forthwith. I know 1400 degrees doesn't sound out of line, but keep in mind there is five feet of header and crossover piping between the exhaust ports and the entry to the turbine; the EGT sensor is located at the feed into the turbine, so exhaust gas port temps are waaaay iup there.
12/16/05 Looking at the graphs from these pulls, we think the HP decline above 5200 rpm is showing that the GT-35R is not capable of delivering the needed air volumes at 14.7 psi boost. Justin estimates we will need to see corrected air volumes of at least 70 to 72 lbs/minute, with demands in the 80+ pound range under maximum load. Analysis using the not2fast model tends to support this. Running the turbo as presently configured is pushing the loading very deep into choke. This creates a stall condition in the compressor, and the intake air mass temperatures go through the roof. This leads to heat being transferred into everything related to the combustion chamber. Detonation, turbine failure, burnt valves and melted pistons are prompt consequences. So we stopped right now.
First solution considered was reworking the GT-35R compressor housing to accept a GT-4088 wheel. The T3 turbine can be replaced with a T4 divided housing with a .81 A/R. This means replacing the present mounting bracket platform, and reworking the downpipe to accept the V-band exducer configuration of the T4. These changes will give adequate airflow and move the 15 psi max boost point very close to the choke line, staying on the compressor's "efficiency island."
The .81 A/R turbine, combined with the larger compressor wheel should minimize turbo lag and still deliver the needed impetus to spin the 4088 impeller. We have considered going to a complete 4088 compressor housing, but there are space limitations, as the GT-40 "snail" will not fit between the block, wheel well and the plenum. Why the f*&k did i let myself get talked into FI?
12/20/05 A lot of research over the past few days. Analysis points to the GT-35R being about 33% too small in air mass delivery for the motor's capabilities. Justin thinks the answer is using a T66 compressor, with a ball bearing rotating assembly, a T4 turbine with a P trim exducer and the .81 A/R. The compressor maps indicate an air mass delivery of about 72 #. With a 2.2 Pressure Ratio, this puts the "point" just off the edge of the island, only slightly into choke. This is not ideal, but going with a large-frame compressor, has both space limitations, as well as bringing lag problems into the picture. The GT-40 and -42, on paper, would seem to be a closer match; the T66 is about as physically large as we can cram into the space available. Given that the amount of time at 15 psi boost and full throttle (6300 rpm) is going to be very small, the objectives of minimizing lag time and sharp mid-range (2800-4500 rpm) response point toward a smaller compressor. Plugging the T66 capabilities into the not2fast, we think the midrange flows can be effectively delivered and stay closer to the high efficiency percentage islands.
So we order the T66/T4 with water cooling in the CHRA and ceramic bearings. Turbo should be here just after Christmas. Looking at the information and compressor maps in Garrett's website, http://www.turbobygarrett.com, and in the sketches inTurbonetics, Inc's website, http://www.turboneticsinc.com the T66/T4 unit should physically fit, though an actual trial is going to be needed. The amount of work on the bracket mounting plate won't be excessive; clearances -- block, wheel well, plenum -- will also be "try it and see."
An observation: as we have researched various turbo configurations, the vendors and manufacturers do a really piss-poor job of making useful information available. One of the worst shortcomings is the near-absence of dimensioned drawings showing sizes, clearances or anything to assist making the engine builder's job any easier. When asked, the answer from the vendor is, "buy the (particular model) and see if it fits." Helluva way develop customer service to boost one's sales. Awful pun, that . . .
Assumingf the pieces are a 'go', the dowwnpipe will need reworking for a V-band flange. This is a dissapointment, as the stainless steel 4-bolt flange is now superfluous . . . the time and dollars, both considerable, are now lost.
12/29/05 The non-working ABS appears to be the result of a dead control unit. This was an original part on the car, and worked fine when it went to EB. I asked Dave why an ABS control box would fail. He tells me the units need to be removed, or at a minimum, electrically isolated or disconnected if any welding is to be done on the car. EB did some welding on the chassis to fabricate the IC mounting brackets. So who knows?
12/30/05 T66 is here from Turbonetics; the -35 is pulled. Definitely some fitting circumstances to be addressed. Additional clearance items are the cruise control servo and the oil line fittings at the block. Getting close-angle oil line fittings is probably going to mean a visit to Combs Aviation and paying thru the keester for trick aircraft-certified Aeroquip fittings. About ten times the price of race-grade stuff at Peterson Fluid Dynamics (the local Aeroquip thief).
The intake plenum is going to need a LOT of metal being removed. Happy New Year. 1/5/06 The T4 exhaust flange is welded onto the crossover pipe feeding into the turbine. The mounting platform has been reconfigured as well. We have some worries about the welding and reshaping of the crossover pipe. The end of the pipe has had a lot of welding and bending done on it. There may be a problem down the road with the amount of heat stressing applied to the relatively thin 304 stainless tubing. This may make the area prone to cracking from vibration and heat cycling. Part of this is due to the large variation in thicknesses of the flange (.500") vs. the tubing wall (.060").
The answer may be to replace the entire crossover tube, with an appropriate circle-to-rectangle flare being fabbed out of an intermediate thickness strip of 304. Before jumping into this, I want to see how the present item holds up. In any case, the crossover and turbine heatshield are going out to Premier Coatings the first of the week for the zirconium porcelain heat barrier coating.
1/13/06 The T66 has been mock assembled and the bottom of the plenum hogged out for clearance. Plenum is ready to go back to Paul for shaping and fabrication of the concave insert and redoing the powdercoating. LH, page 7, images # 153, 154, 155, 164.
1/22/06 Pieces back from Premier and are installed. Paul advises the insert is in. He is 95% certain the clearances over the turbo will be sufficient and will clear the heat shield.
2/3/06 Plenum is here. Rework will allow clearance, but is very close to the turbine heatshield. Justin thinks there is an outside chance there may be enough heat radiation to discolor the powdercoating, but the risk is small. The alternative is to send it back to Paul for further work. This would mean 2-3 weeks delay. KH decides to proceed as-is; the issue is cosmetic, not functional and the porcelain coating on the heatshield should keep temps down.
2/21/06 Justin has the motor running, no issues at the lower boost settings. With the recent below-zero weather, work on the dyno has been backed up at Mile High. I'm #5 or #6 in line; possibly onto the rollers late this week or into next (more likely).
3/1/06 Onto the dyno. Initial pulls at 7 psi. ~ 570 rwhp @ 6000 rpm. Boost tried at 12 psi. Indicating 608 rwhp @ 13.2 psi and 5250 rpm. Indications the head gasket is leaking, so enough for one day.
3/3/06 We pull the head. Bad Shit, man. Gasket has definitely begin to fail. No evidence of detonation or other component damage or failure. The apparent reason for the gasket failing is the head lifting against the bolts. Why? it's a good question. New factory bolts. Properly torqued to spec. Cylinder pressures well within reason; certainly nothing other FI motors can support without incident . . .reference what TCD has been running on his M30's. We have a bunch of very upset people at the moment.
Rather than try for a repeat with the stock S38B38 gasket, or go through the drill with SCE, the choice is to contact Cometic and use one of their MLS gaskets. TCD has spoken highly of Cometic; they can provide S38 pattern units. So calls get made. Yes, they have S38B35 gaskets in stock. Yes, they can laser-cut a bore to almost any size we need. Yes, they can supply thicknesses between .071 and .140. We'll take one in .080 and with a hole sized to fit a 94.36 mm bore, TYVM. At $220, cheap at half the price.
will be here in a week. LH, page 7, image #166.Why I didn't go to Cometic in the first place a year ago, I dunno.
I didn't do sufficient homework and look at every possible parts source. As a result, it cost us at least a couple of months in delays and a bunch of $$. Live and learn. Given what I've seen with the rest of Paul's workmanship, I doubt that he overlooked or mishandled anything to do with securing the head. But with the unknowns about how well the head was secured, the next item on the menu is to see about head studs, rather than bolts. VAC is advertising ARP head studs for the S38. This is news. $336 plus shipping. Should be here by Monday 3/13.
With the head off, we have a hard look at things. During the running-in and the dyno pulls, there has been a small amount of oil seepage on the header slip fitting for #3 cylinder. Nothing excessive, just a small dribble. Hard to tell how new it is; dosen't look like there was anything fresh after every drive or pull. Additionally, with the motor stone-cold, there has been a very brief puff of white smoke on initial turn-over. This goes away within a few seconds. During the run-in and on the dyno there has been zero oil consumption as referenced on the dipstick. No indications of any coolant loss by volume or any oil-coolant mixing. More on this in a bit.
Looking at the combustion chambers and valves, there aren't any indications of oil leaks, i.e., wet streaks in the ports or residue accumulation. Nothing wet in the chambers; plugs are dry. Exhaust tracts are sooty and there is some soft sooty deposit on the backs of the valves. This may be from the fact we have been running fairly rich on the A/F.
The valve guides are new, as are the seals. Paul doesn't like to run guide clearances real tight. He has concerns over different expansion coefficients between the stems and the guides. Given that in a turbo application the exhaust gas temperatures on leaving the combusation chambers can hit 1500-1600 deg F., he feels more comfortable leaving a bit of clearance, typically around .004" of radial displacement as measured on the edge of the valve face. This measurement is taken with the valve extended (top of the stem flush with the top of the guide). FWIW, the BMW TIS specifications manual shows that the tolerance on this measurement is .0315", so we are well within acceptable limits. So we are puzzled by where the oil seepage is coming from. The fact it goes away on warmup and there is no indication of blowby pretty much eliminates the rings.
As part of the detailed once-over, we find the valve stem seals on # 2 and #3 exhaust have leaked and are replaced. The head is checked for warpage and is dead flat.
3/6/06 while waiting on the studs, Justin goes about dealing with the hair-trigger clutch problem. This begins with reboring the clutch slave cylinder to 25.4 mm, from the stock 20 mm, machining a clone of the piston sized to the nerw diameter and including a second O-ring seal. The larger diameter piston should reduce the felt pedal effort and sensitivity at engagement. This won't lower the clamping force, but the idea is to make clutch engagement a less touchy proposition.
We aren't taking bets on how long the QuarterMaster is going to live. This has nothing to do with the quality of the unit, but rather its ability to deal with the power going through it. Alternatives may include a multi-plate Tilton, but just about anything that can survive the power delivery has street tractability problems.
Given the present horsepower deliveries, I think a good case can be made for bringing the upper boost limit down to around 12 psi.
The reasons are as follows:
1. The available power is more than adequate for any conceivable over-the-road application.
Higher numbers may be fine for bragging rights, but the dyno pulls to date raise significant reliability questions at the higher boost numbers.
2. Given the limitations of gaskets, head bolts, etc., things seemed to be pretty much OK at 12 psi.
To be noted is the psi dropoff in the pulls done in December with the GT-35R. This may have been when we originally damaged the gasket in attempting to run 15 psi boost.
3. There may be some question as to how much power the clutch and the rest of the drivetrain can tolerate. Justin thinks its limit is around 700 ft-lbs. That being stated, backing off on the raw power is simply prudent.
4. One of the primary considerations for the motor is longevity. Giving up the last 15% or so may well be the price. Not worth breaking the motor over the issue.
3/7/06 Paul has been kept apprised of what's happened. He is not happy; wants to know what we did to one of his motors. Justin sent him a bunch of phone camera pics. These I found in a file after I had sent the first bunch off to graphite, so not included here. Saaaarrreeee.
After looking at the bad news, Paul has a few thoughts.
1. When we assembled the 3-into-1 exhaust collectors, the slip joints were very tight., and received a healthy dressing of assembly/anti-seize lube to aid the process. The female part of the joint is on top. Paul thinks the weep stains may be from the assembly lube melting out and working it's way thru the slip joint and onto the pipe exterior. Next step here is to thoroughly clean the collectors and see if anything else reappears.
2. Paul was very surprised at the amount of gasket stretch and the fact that the #3 fire ring had moved as much as it had . . .nearly 8 mm towards the exhaust side of the block. He noted that the degree of deformation has pushed the gasket into the oil galleries in the block, most notably at #3, but the process is well under way on the other cylinders.
This may have permitted some leaking between the gallery back into the combustion chamber, thus creating much of the exhaust smoke we've seen and affecting how the thing was running.
The extent of the deforming points to more than one would expect from the time on the dyno. In other words, the problem may go back to the first 4-10 hours of operation. At that time, the car was terribly out of tune -- running pig rich, and ignition advance was not where it should have been. This could have allowed detonation; the gasket yielding as the weak point.
The gasket being pushed out into the galleries very well may have allowed some coolant leakage as well. We never saw any indications of gases getting into the coolant--smell of fuel, bubbles, oil slick in the expansion tank. Certainly no "milkshake." But early on during some of the initial operation, and during the over-the-road testing, there definitely was white exhaust smoke. Not a lot, and invariably at startup or on hard acceleration after a coast-down.
He thinks cylinders 4-5-6 are seeing pretty uniform cylinder pressures, based on appearance and his practiced eye. #3 is obviously 'way weak, with #2 not as bad, and #1 kind of in the middle. One thing we will need to do once back together and on the dyno is to have a look-see at the A/F ratios by cylinder. The TEC-3r can do this, and I am running a PLX wide-band sensor anyway. Paul suspects there are hole-to-hole mixture imbalances and things definitely will benefit from being leaned out.
3. Another possible answer on the oil seepage issue. Looking at the photo of the block with the gasket in place on the deck. On #3 it has pushed out far enough to where it intersects the rectanguilar opening of the oil gallery. The gasket has a very small tear in it next to the steel of the fire ring -- more or less at 6 o'clock on the fire ring -- looks like a dark brown vertical streak. This is sufficient (we think) to allow some degree of oil seepage into the cylinder on startup when the heat hasn't expanded the block/head/gasket sufficiently to offer a better seal. So for the first few seconds cylinder vacuum draws lubricant into the combustion chamber, thence out the exhaust. This passage also allows combustion byproducts -- soot, etc., to be forced into the lubricants in the course of operation. This may offer an explanation as to why the oil is (a) very dirty, almost grey-black after a couple of hundered miles, and (b) why the oil has a strong odor of unburnt fuel. Running 'way rich, remember? We have no indication of any weakness in the rings and the valve guides appear snug, save for the two weak seals.
4. Taking a second look at the head and #3 combustion chamber, in particular at the defined edge of the quench shelf by the exhaust valves. The edge of the shelf is ever so slightly rounded off. On the other CCs, this edge is crisp and well-defined. No penetration of the ceramic coating, however (good news). Paul thinks this is due to the "hydraulic" erosion caused by gas leakage caused by the gas leakage past the stretched fire ring.
I measured the head for warpage and deformation, paying extra attention to the area between the water jacket openings on #2, 3 and 4 cylinders. Done with a straightedge and feeler gauge, the surfaces appear flat and true; if there is anything out of line, it is well less than .001." This adds to the confidence level that the spate of problems didn't cause (or was the result of) any head warpage.
3/15/06 ARP head studs are here from VAC Motorsports; ditto the Cometic .080 MLS gasket. LH, page 7, image #166. Valve stem seals replaced on #2 and #3 exhaust valves. Exhaust headers and collectors pulled and hot-tanked to remove the hardened oil "dribbles."
3/16/06 The head studs are too short. When installed and properly bottomed in the block, they are .150-.200 too short with the washer in place. LH, page 7, image #168. I call VAC. Their answer: Put a short piece of rod in the bottom of the tapped hole in the block and bottom the stud on that. WTF?
The length of the threaded portion will tell you how long or short to make the rod plug. WTF?? Paul advised of the situation. He reaches his contact at ARP, outlines the problem, communicating the needed dimensions. ARP's people do some research and come up with a solution. Seems there is a 12 mm head stud they made for an F1 engine builder (NOT BMW). There are a few pieces left over; not a cataloged item. Price? Youdoan'wannaknow. Done. They are on their way.That evening, I went home, fixed three fingers of JD, and pulled out Corky Bell's "Maximum Boost. Looked up Corky B's comments on head fixation. CB strongly recommends using studs (duh), but also emphasizes being sure to bottom the studs in the block. This neutralizes any compressive or tension forces in the block or stud before tension is applied to the stud when the head is torqued into place. In any case, the length of the thread chamfer is definitely not the way to establish stud working height.
So why is VAC proposing putting a small piece of rod in the bottom of the block stud/bolt hole so the stud will bottom on that? This would appear to negate part of the attachment /securing design. This advice scares me, as VAC has been developing a pretty good reputation in the BMW aftermarket/engine rebuilding world.
3/22/06 Studs end up getting sent to Paul. He is forwarding them to Justin, so should have them by Friday. Per Paul, the length appears to be correct. Threaded portion will bottom in an M30 block, with about 1 1/2 threads showing. If shank length and head thickness dimensions are correct, there should be about 3/8" threaded portion of the stud protruding above the nut when installed. In other words, correct length. We shall see.
3/24/06 Studs here. Threaded segments as well as overall shank length significantly longer than the VAC offering. Shanks have a slight bulge above the threads which may limit how deep the threaded 12 mm portion can go into holes in the block. The bulge above the threads is a result of the threads being swaged, rather than cut into the body of the stud--a good piece of engineering. LH, page 7, images #169, 170.
We'll begin the swapout and reinstallation on Monday.
I'll be interested to see if the main body of the shank is too long, i.e., if the nut ican be screwed down to the beginning of the threads, even with the washer in place. Paul is trying to reach Justin with the ARP-secified torque settings; apparently the spec sheet was inadvertently not sent with the studs themselves.
3/25/06 Paul tells me the slight bulge above the block threads cleared the block counterbore on his M30 block, allowing the stud to bottom properly. I don't know if that same configuration is present on the S38. He understands my concern about where the nut will end up given the shank lengths. If needed, we may be able to obtain different thickness washers. Paul reminded me to be absolutely certain that a proper Hylomar seal is in place around the front of the gasket to forestall any oil seepage from the chain area when the head goes back on.
3/27/06 One of the new studs is test-installed. The bulge effectively limits screw-in depth, but the threaded shank is 1-2 threads longer, so we are good on purchase, the stud being able to bottom. The overall stud length has interference with the underside of the cam tray. Given the nut height, Justin is blanchard grinding the washers down somewhat thinner. S/b done later this week. This will allow full purchase of the nuts with 1 or 2 turns extending above the top of the nut. LH, page 7, images #171, 172, 173.
Justin has run Hylomar just about everywhere--around the timing chain area, the oil and water gallery openings, between the layers in the MLS gasket. This sucker is going to stay
dry. 8)
3/29/06 Washers ground, turned to fit head recesses for the studs/bolts. Studs installed. Cam tray is relieved on the mill near the intake valve spring portals to allow installation without shortening the studs. Cams installed. Paul had originally set up valves ~.013-.014 on intake, .016 on exhaust. This for break in. "A loose valve is a happy valve." With the motor having been run-in, can probably take up .002" or so.
Unused head stud set returned to VAC. They were not happy with what I described and it took some discussion
before they were willing to provide a Visa credit . . .less a fat restocking fee. 3/31/06 Valves adjusted .011-.012 intake, .014 exhaust. Throttle bodies in place and synched. New plugs.
4/7/06 (Playing the Darth Vader Theme from Star Wars)
Dyno runs completed. Low-boost pulls are done @ 7.25 psi. 565.3 rwhp @ 5700 rpm, 547.4 rwtq @ 4900. Justin does a few more changes to the BIN file (we are now at file #140
), then leans into it. 707.2 rwhp @ 5700, 725.8 rwtq @ 4900. Both HP and TQ curves noticeably flatten out beyond 4900, as the boost drops off from the 14.95 psi peak. "You need a bigger turbo, Ken . . ." "Yah. Pigs fly." 8) Holy Shat, Bitman.
Since some people have asked, the recorded numbers are delivered at the rollers. This is on a Mustang eddy-current dyno.
Further, the readings have been adjusted for altitude (barometric pressure), temperature and relative humidity. So what is shown here would be the same as a comparable pull done in, say, Silicon Valley, or Connecticut. If one measures the "actual" rwhp, the values are roughly 20 to 22% lower, or in the 550 to 565 rwhp range. Manifold air pressure is taken from a pressurized vessel (the plenum), so the boost level is the same, regardless of the test location's altitude.
What's the power at the flywheel? It depends on whose numbers you want to use for power train losses. If one refers to Puma Racing's website, http://www.pumaracing.co.uk, David Baker suggests adding 10 hp to the roller number, then dividing by .88. This gives fwhp of around 815 hp. I might add that this is a must-read site of you are trying to become eddicated about performance engines, BTW. Plugging in known engine parameter values into not2fast, I can come up with 842 hp. Another method is to assume a 16% drivetrain loss. This yields 707/(1-.16) or about 853 fwhp. In any case, it is more than you want.
The dyno pulls don't end the story. There's more to come.