27 Feb 2018

It sometimes seems to take women forever to get really ready to leave. Reminds me of a comedian who said that when your wife says she’s ready to go, it means you can watch the second half of the game and not hold her up.

Anyway, with the “storage area” of the garage freed up, work continues on pulling the engine. Interesting what you find when you take apart a car when you think there’s nothing wrong:

Leaves, dead bees, and small rocks in the intercooler – surprising given how little time the ducting has been in-place.

A braided-stainless oil line that passes right next to the alternator bracket, where the two are wearing into each other.

Found the mystery oil leak that’s been messing with me every since the engine was installed. I’ll get a better look once it’s pulled out but there’s strong evidence of a casting flaw that’s allowing oil to leak right through the pan. Will either weld it or use something like JB Weld or even gasket sealer.

Will probably pull the drivetrain tomorrow, then immediately drive the transmission to WaveTrac to have the axle CV housing removed and get the differential fixed. They don’t want me hanging around, waiting for them to remove the axle part, so it means delaying getting the fully running. The reality though is that the delay probably won’t matter because it takes a lot longer connecting everything back together than removing it, and then there’s the side project of correcting the rearr engine mount.

25 Feb 2018

Started in on the transmission swap, just ancillary parts for now since there’s no room to store the big stuff yet (wife’s car’s in the way!). During removal and inspection, found the likely source of the recent vibration that’s been noticing during acceleration. The rear engine mount plastic insert has deformed (squeezed out) enough to allow the two halves of the mount to contact metal-to-metal, hence the extra vibration. With more than 400 ft-lbs of torque, I guess I should expect this sort of thing. Going to have to redesign the mount but that can be treated as a separate project.

20 Feb 2018

Synchrotech Transmissions provided pictures of the transmission build. I’m familiar with the insides of engines but transmissions have always been a little mysterious; I would have liked to have been there but the pictures are almost as good at showing what’s involved. Dealing with Synchrotech went smoothly; once they received the parts, the build and delivery happening promptly, can’t ask for more than that.

As part of the transmission change was the thought of also changing the twin-disc clutch; could a replacement be found with lighter foot pressure and possibly quieter (twin-disc clutches have a characteristic rattle when in neutral). The twin-disc also has a feature that for a race car is a non-issue, but for the street is a little annoying; there’s always some slight coupling between the input and output even with the clutch fully disengaged – meaning the gears are still spinning slightly. Shifting into gear from neutral, it can always be felt that the gears are moving, so it has to be done deliberately.

I contacted Competition Clutch, maker of the clutch, and explained the situation. They said that one of their “Stage 4” clutches would do well, even after I explained the 430 ft-lbs of torque, chewing-gum-soft tires, and being mid-engine, traction increases under power rather than decreasing like a FWD. They also said that pedal pressure would be about the same – not sure how I expected it to be less. Part way through the discussion though, I realized that swapping in any other clutch meant also changing the flywheel, since it’s unique to the twin-disc. Separately from that was the realization that because I’m switching back to a synchro transmission, it’s highly likely that any slight gear rotation in neutral will be dealt with by the synchros and no longer be noticed.

I was a little skeptical how a single disc clutch could hold the same torque as a twin-disc, so to get a second opinion, K Series Parts (formally Club RSX) was called. The surprising advice was “stick with what you have, it’s the right solution.” They said that the engine’s torque is too much for any streetable single-disc clutch and it either won’t last or will be a real bear to operate (high pedal pressure). It really impresses me when a business gives an answer that prevents a sale – that right there demonstrates that they’re  more interested in the right solution. They weren’t left empty-handed though, as I bought the Gear-X gear set through them.

Lastly, with yet another engine pull looming, I wondered if there was any way to drop the engine out the bottom rather than pulling it out the top. Looked at the engine bay today and was surprised to see nothing preventing it – other than the cross-bracing of course. It would have to be cut out and then either be converted to a bolt-in assembly, or welding back in after the fact.

18 Feb 2018

In a couple weeks, the Honda K-series “DC5” transmission in Midlana will be up for sale. The specs:

Straight-cut dog-engagement 1-4 PPG gear set, rated for 1000 hp:
1 – 2.615
2 – 1.611
3 – 1.15
4 – 0.909

5th and 6th stock helical Honda (even though it’s a DC5, 6th was adding for freeway cruising:
5 – 0.825
6 – 0.659

NOTE: Because 5th and 6th are present, this transmission does not contain the PPG “center brace attachment”.

Final drive: stock Honda 4.389

WaveTrac LSD.

Full disclosure, the WaveTrac LSD is apparently one of their early units that has a known issue where it wasn’t machined correctly and traps the driver’s side CV housing in place – mine’s stuck good. The good news is that WaveTrac offers free repairs and the transmission will be delivered to them to be corrected before being sold. It’s a win-win because the buyer gets a known-good LSD, their axle won’t get stuck, and I get my CV housing back. That last part matters because the axle shop wants $250 just for that part…

The transmission works perfectly. As for price, it was $7000 to build and given that it only has 5000 miles on it, it’s worth $4000. If it doesn’t sell via this site, the Midlana forum, or K20.org, it’ll head to Ebay.

Because of the cost, I made a demo video, just some casual driving showing that the transmission is operating perfectly and it also demonstrates what a straight-cut dog-engagement gearbox is like. There would have been more video but the bug-ridden GoPro struck again, stopping the video mid-drive. Oddly, the red light continued blinking, indicating that it was recording, and when turned off, acted normally. I’m This Close to switching to the Sony like my brother has…

17 Feb 2018

With the finished transmission on the way, it’s looking like it’ll go into the car at the end of the month during a marathon garage session. While I’m doing that I’ll also be figuring out how to attach the rear wing assembly. On the Midlana forum I received a good tip about making the mount hinge so that the assembly can be rotated down out of the way for engine access. Thanks, Bill!

10 Feb 2018

Back when the oil and smoke incident occurred during the dyno tuning session, the theory at the time was that maybe the turbo oil return line got sucked flat by the dry sump scavenge pumps. My buddy Dave asked what crankcase pressure was; how did I know that it wasn’t just the opposite, that under heavy boost, maybe blow-by (past the rings) was pressurizing the crankcase to a positive pressure and actually pushed oil back up the return line? I said that can’t happen because it’s a dry sump; the crankcase is always at a vacuum. He countered with, “do you know that for sure; have you measured it, or are you just guessing?” Ugh, he was right, it’s bad science to just decide something without knowing for sure, and even worse to make decisions based upon it.

A test setup was assembled consisting of a 200kpa (+/-15) psi gauge, hose, and valve cover adaptor. I prefer kpa because it’s clear what vacuum is – 0, and ambient is 101. With English units, “zero” can be confusing because it depends upon context; it can either be ambient pressure or a perfect vacuum.  Ironically, the gauge manufacturer scaled the meter wrong; there’s no such thing as negative kpa, zero is zero, a perfect vacuum.

Warmed up the car and crankcase pressure settled out at 55 kpa (-13 in. Hg). The picture below was a couple minutes after starting it and before oil and coolant came up to temperature. Took it out for a test drive and during cruise, crankcase pressure fell (meaning vacuum increased) to about 40 kpa (-18 in. Hg). Found a deserted stretches of road, cranked up the boost to 15 psi and did a few 4th gear pulls*. Vacuum dropped to 55 kpa (-13 in. Hg), the same as at idle. This was a relief because that’s what’s supposed to happen; the dry sump pump maintaining a negative pressure even when producing maximum power. So for now at least, the collapsed turbo oil return hose still seems the most likely cause of the engine spitting out all the smoke and oil. Being immersed in hot oil for extended periods of time under vacuum very likely softened the rubber hose enough to allow atmospheric pressure to squash it flat. Once that happened, the oil couldn’t leave the turbocharger and filled up the center section and pushed past the seals into the inlet and exhaust sides. The anomaly hasn’t happened since, but than again it hadn’t happened before, and the rubber hose was since replaced with a Teflon part.

After the test and after shutting off the ignition, it took about 30 seconds for crankcase pressure to slowly rise back to ambient pressure – sort of a poor man’s leak-down test and a reassuring sign that there aren’t any major crankcase leaks.

*Ever since the retune, the car is a serious handful at full boost. Even a very slight bump in the road causes some wheel spin at even triple digit speeds. It’s one reason why there’s a knob on the dash for boost and it’s normally kept turned down to keep both me and the car out of trouble. That said, flooring it in fourth on the freeway at full boost is – frankly – effing awesome, as it’s as if everyone else put their brakes on 🙂

8 Feb 2018

All the transmission parts are on the way to the builder, who’s also sourcing the core. This avoids the concern about shipping him a used tranny of unknown history,  bought from a stranger from out of state, and having the builder possibly say, “this core is trash.” After checking what used transmissions go for and what the shop’s charging, their price is very fair so we both benefit.

Regarding the LSD, I found some good information in an unexpected place, Porsche 911 forums. 911s have an higher rear weight bias than Midlana, and being Porsches, many owners track their cars, so the topic of limited slip differentials often comes up (“what’s ‘best’ “) . There are many brands that fit 911s and the Giken holds its own. It was informative to read how the Giken helps stabilize the tail-heavy 911 under braking, during partial mid-turn throttle, and during full-throttle corner exit, all things I’m interested in. Of course, the Internet being what it is, you have to stand back to get an overall view of impressions instead of fixating on only desired posts.

Another good tidbit was that you shouldn’t install just any LSD into a mid- or rear-engine car because the chassis dynamics are different from front engine cars, be they FWD or RWD. Because of the differences, a proper clutch-type LSD for a mid-engine car is a “reverse 1.5-way”, and you can tell from the name that it works opposite what a front-engine configuration needs. Very glad I found that out before having it installed! As an aside, I also found that OS Giken is owned by Toyota, which seems like a good thing.

Speaking of 911s, I found this post which I can relate to:

“One of the things I had to force myself to accept when I first was learning to race [rear-engine Porsches] was that “more throttle equals more rear grip”. It was very counter-intuitive.

Luckily I was driving my instructor’s car and he kept pushing me to “Get on the gas!” at corner apex when I “knew” I was at the limit of adhesion.

So I remember thinking “Well I’ll show him, I’ll do what he asks and then he’ll see that the car is going to spin”. I mashed the throttle and the car hooked up and flew out of the corner. “Wait, what?!”

I’m hoping that the transmission shows up before the end of the month because with the wife out of town, there’s an uninterrupted week to do the laborious gearbox swap. It’s going to be a ton of work but it’ll be worth it. Another reason to do it while she’s out of town is so I don’t have to hear about her car sitting outside for the duration.

In my spare time I’ve been (re)reading my aerodynamics books. Until now I’ve skipped the wings chapters because all I cared about at the time was cooling system design and whole-vehicle airflow.

So yeah, I’m a little excited, looking forward to a better track experience, but I also realize that the above changes aren’t a magic bullet and don’t correct bad driving habits.

4 Feb 2018

The wing material is on the way and until it arrives, it gives time to figure out something that’s been bouncing around a long time – my transmission. The current unit is a base-model RSX 5-speed converted to a 6-speed, gears 1-4 being straight-cut with dog engagement and 5-6 being OEM, and a stock Honda 4.39 final drive running a WaveTrac LSD. In short, another transmission’s going to be built. I can already hear it “why do you keep messing with stuff instead of driving it?”, and, “why didn’t you think this through the first time?” Easy answer to the latter: the gear sets weren’t available then. As for the former, well…

Anyway, the existing transmission works great on the street but isn’t optimum for the track on several points. As previously mentioned, 5th and 6th are OEM so they can’t be “used in anger” on the track with a turbo engine (I turn boost down to protect them). This isn’t a problem on the street because presumably you aren’t going >150 mph. The OEM gears can deal with ~160 ft-lbs from the stock engine but users report bad things happen when pushing high torque. Additionally, first gear also isn’t that useful on-track because it’s numerically than optimum, better suited for the street or drag racing. Lastly, the differential works perfectly on the street but like the gears, isn’t the best for on-track. All these issues came together into the idea of building a new transmission and selling the current one while it’s working perfectly. It’s worth decent money, versus practically nothing if I break it. The idea is that it can help pay for the new one, which will consist of:

1. Full Gear-X gear set, with lower numerical ratios and stronger 5th-6th
2. OS Giken LSD (reverse 1.5-way clutch type)
3. Carbon synchros

The final drive ratio will remain the same, 4.389.

Gear-X offers two gearset ratios, identical other than 5-6 which vary slightly depending upon application. Assuming a self-imposed redline of 8000 rpm, the more long-legged set tops out at a theoretical 170 mph, while the very slightly shorter set tops out at 162 mph. The fastest Midlana might ever see is about 160 mph at AutoClub Speedway, but wings are planned, so expected top speed will drop to something less, so the slightly shorter gearset appears best. Back in the practical world, with this gearset, 70 mph on the freeway results in  an engine speed of around 3400 rpm, so that works.

I asked an Arial Atom owner what gear ratios he uses because this particular Atom has as much or more power than I do. The reason I asked is because the Gear-X first gear is really low numerically (2.313) compared to 2.615 of the current PPG first, which is lower than the OEM ratio – it’s a pretty big difference. Because of this, there’s a lot of complaints that such a low first gear ratio makes it all but impossible to take off fast and is “obviously just for road racing, not the street.” The Atom owner correctly pointed out though, that when those ratios are put in a car weighing half as much as OEM, it changes everything and is downright perfect.

Regarding the differential, everything I read indicates that the OS Giken LSD is good for somewhere between 0.4-3 seconds a lap. Granted the numbers are anecdotal with little basis in hard fact, but what was telling was how virtually everyone who switched to it said they went faster. It’s supposedly also able to make the car easier to control in turns. Again, what “easier” means, who knows, but it’s promising that all the comments are positive. BTW, the “reverse 1.5” configuration of the LSD was recommended by the manufacturer specifically for the mid-engine Midlana and would not be a good choice for a FWD engine placement.

Because the above gear set is rated for 500 hp and is also helical, I’m somewhat taking a step backwards. The reason is that for street driving, it’s the right choice for me. On the other hand, for someone who’s built a turbo car for drag racing and wants to be a badass on the street, the dog-box will serve them well.

Lastly, adding the carbon synchros should prevent the dreaded Honda 2nd-gear grind that tends to happen if shifted too energetically for too long. There’s still some logistics to work out but it looks like it’ll happen. When it’ll be done and when it gets installed, who knows.

Okay, there might be one more perk of the above gear ratios. Between the new first gear ratio, the existing final drive, tire diameter, and the engine’s red-line, I can reach 60 mph in first gear – I’ve always wanted at least once to own a car that can do that. The reality is that it’s totally pointless, good only for setting a rather awesome 0-60 time :)

I’d love to try out a sequential gearbox but can’t justify the ~$10,000+ entry fee. There’s the cost, but there’s also the suspicion that it might be a pain to live with in traffic, given how shifts aren’t buffered by clutch engagement, going “bang” every single time, up or down, unless you perfectly match gear speeds. Also, users report that the sequential unit should be considered a wear item (caused by imperfect shifting) that requires periodic teardowns. I think I’ll pass.

28 Jan 2018

Had to load a previous tune. One of the many things done to achieve a smooth idle was to run open-loop idle (no AFR adjustment). Once the smooth idle was reached, I wondered if maybe closed-loop lambda control could be reenabled – nope. Car regressed back to falling flat on its face leaving a stop. Fixed.

In other news, my 2013 Ford F150 has worked fine until a couple weeks ago, when the blower fan started making the truck feel like one of those cheesy vibrating beds in motels – something had thrown off the fan balance.

A few online videos made it look doable (anything under the dash of a newer vehicle is always a dicey proposition). All the videos talked about how hard it was to get the screws out but didn’t mention actually getting the unit out from under the dash as an issue, yet both took about the same amount of time. I was guessing that leaves, seeds, maybe a dead mouse, or paper from the floor had got sucked in. What it turned out to be was… I have no idea. Whatever this is was likely always in the ducting somewhere and took 5 years to fall into the fan.

27 Jan 2018

As I was putting the new tires on I saw something that made me think “that’s strike 2 for Discount Tire.” The first time was when they bent all four of my lightweight racing wheels and then denied it. This time, a lazy employee mounted one of the tires with a label on or very near the tire bead. What you see is after I tried tearing it off, hoping that it just looked like it went across the bead. Apparently it does and it remains to be seen if it’s going to leak.

Went for a test drive and being new and very sticky, the tires throw all kinds of grit into the air; I wouldn’t want to be driving behind me on a less-than-clean road! Noticed pretty quick that the ride was really firm and realized they were inflated to support a 3500-lb car. Sure enough, they were all 35 psi and were backed down to 25, but they’ll likely need to drop to 20 psi, if not 18. One of the big perks of a light car is the low cost of consumables like tires and brake pads. Hellcat, Corvette, Mustang, and Camaro drivers have substantial tire budgets.

In other news, there’s how to mount the wings. The 1-meter wing sections are designed to accept a 1″ tube through the center of pressure. Because of the width and where the support points will be, the overall assembly will consist of a 36″- 1 meter inner piece with the support points at each end. To the same supports will be additional 18″ outboard sections attached to the same mounting points, making the entire assembly as wide as the outside of the tires. It would be nice if the assembly breaks to for storage, as it’s going to spend most of its time up in the rafters. I don’t want a huge and clunky 3D assembly, so some sketching is in order.

26 Jan 2018

I know it’s been quiet but plenty has been going on:

During initial cold-start, the car always stalls during the first attempt, but after that it’s fine. Though not a big deal, I feel the car needs to demonstrate that it has a tune better than just “okay”. To fix that means backing up a bit to explain.

The ECU provides many adjustable tables and tuners use different methods and personal preference to achieve a desired tune. Point being, some tables have a choice of several variables to use to achieve the same objective. For example, the idle valve calibration table can be dependent upon engine speed OR coolant temperature. It had been set to engine speed but at some point, some bonehead (cough) switched it to coolant temperature, probably wondering “I wonder what this does.” What I – I mean, he – didn’t realize was that the ECU just shrugs and says “okay”, without altering the table other than the units. This resulted in the rather humorous realization that what had been 700-1500 rpm was now 700-1500 degree C! The only reason it worked at all is a testament of the ECU’s closed-loop idle function – it just made it work. It ran surprisingly well, too, which is one reason why it took weeks to catch the error. It’s also a stern reminder that tuners have to be very careful about ignorantly changing things – setting something wrong can destroy the engine in seconds. I got away lucky, though I do know better than to idly mess with more critical parameters. No pressure at all, this tuning thing…

Anyway, now the car starts first try and idles properly even during initial cold-start in the morning, with “cold-start” being defined (in Southern California) as coolant being less than 20 C.

Another ECU change was to reduce boost in 5th and 6th gear – disappointing but necessary to keep from destroying the OEM gears. With 400+ ft-lbs of torque, the gears, used to seeing ~160 ft-lbs, just won’t last. Unfortunately it makes Midlana uncompetitive at Autoclub Speedway where top speed is 150-170 mph. That said, there’s another change, below, that may make that a moot point, maybe.

The next issue was the persistent engine oil leak. I thought I had fixed it by applying sealant to what seemed to be the leaking fitting – nope. Now it looks like it’s a different fitting that’s leaking – one with another Stat-O Seal – not very impressed with them but it may be due to the machining on the pan. Anyway, the fitting was backed off enough to apply sealant and screwed back in. While in there, sealant was also applied along one edge of the front cover where it’s been weeping since new. The engine was started and warmed up, hopefully drawing sealant into the leak and adding heat to cure it. Now we wait and see, sort of like playing Whack-A-Mole.

Next topic: wings! I’ve gone round and round with putting them on, waffling between thinking they’d be awesome, to being afraid of them for what they allow, letting drivers run much further out onto thin ice and counting on something invisible to save them. For example, a driver gets used to a wing’s benefits of holding the back of the car down and preventing a spin at high speed. Meanwhile, the wing stops working for some reason, like, oh, falling off, and great excitement ensues, like this guy – https://www.youtube.com/watch?v=_LnmD7kxHUw

Another concern is that the benefit depends upon things that the driver has little or no control over: wind speed and direction, and disruptive airflow from other cars. My brother countered that wings aren’t all that different than brakes; if either fail it’s going to be a wild ride and it’s on the designer/builder/maintainer to keep an eye on them. I don’t completely agree with for the above reasons, but curiosity made me decide to give them a go.

Because of my dislike of composite work (read: not having a proper workshop to do things right), I  went with extruded aluminum wind generator blades from  http://www.thebackshed.com/windmill/Trade/AlBladeOrders.asp. With a chord of only 6″ they’re not optimum, but a duel-element assembly should be enough to prove out the idea. As said before, we just do trackday events; there are no points, no championships, and no chance at F1-stardom. Being aluminum extrusions means fabrication will be about 100 times easier and faster than a foam-core layup.  The visual change isn’t going to be subtle – the rear wing and endplates will take up the entire width of the car, 72″ (1.83 meter) wide. The only time they’ll be mounted is for track use; on the street they’re not only pointless, but pretentious and attract all the wrong kinds of attention.

Lastly – tires. They’re ready to be replaced because time has flown by – they’re now 7.5 years old (read: hard)! The rears are really worn as well, bad for a cop to notice. The good news is that they’re fairly evenly worn, meaning the camber gain curve is just right. Nitpicking, the center is slightly more worn, indicating that tire pressure could stand to be a bit lower. The thing is, with such a light car, going with lower pressure makes them more “squirmy” and they currently run around only 18 psi. That said, in general, for track use, more pressure means better traction, at least to a point.

I chose to go with Nitto NT-01s this time. The sizes are the same*, with 205-40/17 in the front and 275-40/17 in the rear, but a different brand. My brother uses the same brand and is very happy with them because they remain sticky a really long time, so I thought I’d give them a go as well. It also removes one more variable of difference between our cars.

Sort of tire related, Miata front hubs are one-piece units that aren’t intended to have the bearings removed (maybe they are but I’m not bothering).  People who race Miatas complain about short bearing life, but since Midlana has about half the weight riding on them, I think they’ll be fine. That said, some play has developed in the front tires that at first I though was due to rod-end play, nope, it’s the bearings. Given that they’re used parts of unknown lineage, they may well just be worn out, so to “reset the clock”, they’re being replaced. This will better serve to see how long they’ll last on Midlana from new.

Oh, one more thing. Because the rear tires were so worn and about to be replaced anyway, I intended to record a burnout for people who enjoy that sort of thing. When the moment came though, several things happened. One was that the location wasn’t quite as deserted as I would like, so there was some concern about undue attention. Another was that I discovered that I have a lot of sympathy for mechanical things and don’t enjoy subjecting the car to such antics. That said, I did do two, only they didn’t turn out as expected. Flooring it in first did indeed spin the tires, but the car sped up fast enough such that the tires never got hot enough to smoke or leave big shameless marks on the street. I’d also tried it in second, thinking that maybe I needed to increase how fast the tires were spinning, but again, the car just shot forward, a likely consequence of a lot of weight on the driven tires and being so light – a good thing. Yeah, I know F1 cars can smoke their tires on demand, so maybe I just suck at abusing the car 😛

*When magazines do tire tests, they typically pick some common size and do back-to-back track tests, then post the times. Whatever tire has the lowest time is deemed best, at least until the next tire comparison. There’s one thing many of the magazines leave out though, how a given tire “size” is more of a suggestion than fact. That is, a “255-40/17” tire from factory X is indeed 255 mm wide. Factory Y decides to one-up their competition, coming out with their own “255-40/17” tire which has a tread width of 265 mm. Factory Z, not to be outdone, released yet another “255-40/17”, this one with 270 mm tread width. Guess which one will most likely win? Tire comparisons based on a given size won’t necessarily find the best tire, exactly, yet the irony is that it doesn’t seem to matter. Class rules usually go by what’s printed on the side of the tire, not by what they really are. Anyway, these new “275mm” tires are about 6mm wider than my old “275mm” tires, making them 281 mm. Okay.


12 Jan 2018

Copied from the Midlana forum (link above):

So as big things get solved (intercooler ducting, ECU retune, drivability, etc.), issues previously less-important have bubbled to the top of the list.

After installing Engine V2.0, an oil leak was found under the engine, unfortunately traced to one edge of the front cover. I say “unfortunately” because on a Honda K-series engine, the front cover is trapped by the oil pan and cylinder head. Even if I got it out without destroying the head or pan gasket, there’s no way I’d get it back in, and I’m loath to remove the pan and head! Neither gasket in those areas are critical but it would most certainly introduce new oil leaks.

What’s interesting is how the dripping only happens with the engine is not running. Due to running a dry sump oil system, a side perk of it is that it pulls a vacuum in the crankcase which prevents any oil from leaving! The puzzling thing is that just sitting there it drips a fair amount of oil. Over the last six months or so, it’s filled the drip pan on the floor about half way. With the leak along one edge of the front cover, it’s fairly high up on the engine, so where the heck is all the oil coming from? I’d expect that as soon as the engine’s shut down, the hot oil in the head and front cover would quickly find its way back to the pan, and yet there it is on the floor.

So the challenge is how to fix it in-place and what *might* work is to take advantage of the dry sump’s vacuum. That is, apply gasket sealer along the suspect area, then drive for an hour or so. The idea is that the vacuum will draw the sealer into the leak, and the heat will cure it relatively fast. The trick is finding a low viscosity gasket sealer that gets drawn easily into cracks, which seems unlikely, so I’ll probably be stuck with whatever’s available. That said, if it’s too low a viscosity, it’ll get entirely sucked in and not seal anything; water-thin is too far in the other direction. I guess the first thing is to double-check that the leak’s coming from where I think it is.


Thanks to reader Bill who said “maybe it’s time for some UV dye and a UV lamp.” Well, huh, I had no idea such a product existed, and ordered AC Delco 4-In-1 Fluorescent Dye, PN 10-5045, and it’s affordable. During our discussion, it occurred to me that the leak at the front cover might be a red herring, that maybe there’s a second larger leak lower down. The plan is to add the dye, run the engine until warm, then shut it off and watch and wait. The instructions say that the dye can be left in the system without harm, but for peace of mind, the oil will be changed after the leak source is found and fixed.