My Favorite Example of a Gold-Standard “Good Kluge”
A “kluge” (pronounced “klooj”) is typically defined as a cost- or time-saving work-around that’s inherently of lower quality than the original configuration. Sort of like “hack,” which, for obvious reasons, I take personally. I mean even I would never do something like hold a battery cable in place with Vise Grips. (Well, not a positive battery cable :^).
Over the years I’ve developed the idea of the “good kluge.” This is a work-around that A, is safe; B, saves money and/or time, and C, is easily reversible. I’m going to set the way-back machine to 2008 and describe what, to me, still sets the standard for a good kluge.
I still had my 1982 911 SC—the only Porsche I ever owned, the car whose selling was my biggest automotive regret. I adored that Porsche, except for one thing: People joke that British cars leak fluid to mark their territory, and as a vintage Lotus owner, I can attest to the truth of that statement. But nothing I’ve ever owned leaked oil like that SC. I eventually understood why.
The engines on vintage Porsches and Volkswagens are often referred to as air-cooled. That’s really a bit of a misnomer, because they’re air- and oil-cooled. The air-cooled part is obvious, as you can’t miss that big belt-driven mechanical fan when you open up the deck lid and look at the engine. The oil cooling isn’t obvious until you look upward from the drips on the driveway.
There’s a primary oil cooler mounted directly on top of the crankcase. It has the engine’s fan blowing through it, just like on an air-cooled Volkswagen Beetle or bus. But on the 911 SC, there’s another oil cooler mounted in the nose of the car. On a water-cooled engine, oil leakage occurs when some component on the block or head, like a main seal or a front timing gasket, starts to leak. On an air-and-oil-cooled car, the engine is designed to have oil flow to components that are outside the block and head.
Thus, there’s inherently more opportunity for oil leakage. One fact that highlights how crucial oil was to cooling the SC’s 3.0-liter six-cylinder engine is that it takes 11 quarts of oil. I always thought that one of the massively cool things about owning the 911 SC was that you could hear all that oil gurgle as it got sent around the car.
![porsche engine bay](https://hagerty-media-prod.imgix.net/2024/07/Hack-Mechanic-Kluge-img_1860.jpg?auto=format%2Ccompress&fit=crop&h=480&ixlib=php-3.3.0&w=640)
So, 16 years ago, I was in yet another round of work to squash the car’s pernicious oil leaks. In my first go-round, I replaced one of the seals to the primary oil cooler—the one that was reachable with the engine still in the car. In round two, I dropped the engine to replace the other unreachable seals. But in this third round, I found that there were leaks in the oil return tubes, and another one in a line that plumbs the nose-mounted oil cooler.
The oil return tubes are funny relics from an earlier design that still perform an essential function. If you’ve ever had an air-cooled VW Beetle or bus, you’re likely familiar with the pushrod tubes. Since air-cooled VW engines had the camshaft inside the block ( I mean crankcase), pushrods were needed to mechanically couple the camshaft lobes to the valvetrain in the heads, as is the case with any pushrod engine.
However, the combination of this engine’s horizontally-opposed design and its air-cooled cylinders meant that the pushrods had to essentially run outside the engine, so they were encased in tubes to protect them. In addition to shielding the pushrods, the tubes, fitted with thick rubber seals at both ends, allowed for the return flow of oil from the head back into the crankcase. On post-356 air-cooled Porsches, since the cams are in the heads, the pushrods are gone, but the tubes are still there to perform the oil return function. And they still eventually leak through the seals at the ends, just like on an old VW.
![engine jack block support](https://hagerty-media-prod.imgix.net/2024/07/Hack-Mechanic-Kluge-oil-return-tubes.jpg?auto=format%2Ccompress&fit=crop&h=481&ixlib=php-3.3.0&w=640)
When air-cooled VW engines were originally assembled, the tubes were rigid and sandwiched between the crankcase and head, so you couldn’t replace the oil seals without sliding the heads away from the case along the studs. Fortunately, some bright soul eventually came up with the idea of collapsible oil return tubes. Since on an air-cooled 911 there are no pushrods inside the tubes, you can simply cut the tubes in half or crush them, yank them out, and replace them with an expand-it-like-a-curtain-rod tube with a lockable clip in the middle and a new rubber seal on each end.
Like everything in life, of course, it wasn’t quite that easy, as I found that the engine’s heater boxes and exhaust manifold obscured several of the tubes, and removing them all would’ve necessitated dealing with rusty nuts on pencil-thin studs that would’ve snapped if I even looked at them the wrong way. But fortunately, the ones that were leaking were the ones that were accessible without removing exhaust components. I replaced them, and that set of leaks was stanched. I could barely believe my good luck.
The other leak, though, was where I had to do something that most would call an ugly kluge, but I thought of it as risk management and cost containment. Flow to the 911 SC’s front-mounted oil cooler is regulated by an oil thermostat mounted under the right rear fender. When the oil rises to a certain temperature, the gate in the thermostat opens, allowing oil to pass to the cooler in the nose and then back to the crankcase. This means that that there are oil lines running from the engine in the rear, to the thermostat under the fender, to the cooler in the front, then back to the engine.
![porsche engine plumbing closeup](https://hagerty-media-prod.imgix.net/2024/07/Hack-Mechanic-Kluge-thermostat_cropped.jpg?auto=format%2Ccompress&fit=crop&h=480&ixlib=php-3.3.0&w=640)
The leak was in the return line that ran from the thermostat to the big metal pipe that wraps around the back of the engine, shown in the photo of the oil return tubes above. The section of the line that goes into the thermostat is metal, but the engine end of the line has a rubber section crimped to it to allow for the engine’s vibration. The leak was coming from where the fitting was crimped to that rubber section. I was able to loosen the fitting on the rubber end that attached to the big wrap-around pipe, but the fitting on the metal end leading to the thermostat wouldn’t budge.
![porsche engine fitment nut](https://hagerty-media-prod.imgix.net/2024/07/Hack-Mechanic-Kluge-oil-line-fitting-pelican.jpg?auto=format%2Ccompress&fit=crop&h=480&ixlib=php-3.3.0&w=640)
I read up on the problem in the 911 forum on the Pelican Parts website, and learned that this fitting is steel but the oil thermostat itself is aluminum, so you have decades of corrosion of dissimilar metals essentially welding them together. In addition, because the thermostat is up under the rear fender, it’s difficult to get leverage on it, and it’s in a constrained area that discourages trying to apply heat with a torch. I ponied up for the correct 36mm wrench for the fitting, but after several hours of trying, all that got me was a bent wrench.
![porsche engine wrench](https://hagerty-media-prod.imgix.net/2024/07/Hack-Mechanic-Kluge-IMG_9467.jpg?auto=format%2Ccompress&fit=crop&h=480&ixlib=php-3.3.0&w=640)
The great thing about forums is that you’re usually not the first person to encounter a problem. I learned that the preferred method for dealing with frozen oil thermostat fittings was to disconnect all of the cooling plumbing from the engine to the nose-mounted oil cooler, and drop it like a giant squid with rigor mortis. With everything on the floor of the garage, you’re freer to do the heat-and-beat thing or to try cutting the nuts with a Dremel tool, but still, there were plenty of tales of broken thermostats.
Like any thermostat that sends liquid somewhere above a certain temperature and bypasses it below it, this one has four connections—the input from the engine and the thermostat-closed output back to the engine, and the thermostat-open, send-and-return connections to the cooler in the nose. If you crack the thermostat, you need to undo all four of these lines, which risks damaging them as well. And there’s the nose-mounted oil cooler to worry about, too. There were stories of people who tried to replace a single line, went screaming down the slippery slope, and regretted touching anything at all. Yeesh.
Part of the backdrop of this repair was that I was leaning toward selling the car (I felt like Matthew McConaughey in Interstellar yelling at his past self, “No, Rob, don’t do it!”), so the idea of this morphing into a major time-and-money sink that I’d get little enjoyment from doing, and that the car might not even be around for me to bear the fruits of my labor, made me stop and consider a different path.
I wondered if I could simply repair the existing line. It was leaking at the place where the rubber section was crimped to the metal fitting. First, I identified a source for a new barbed fitting to go on the end of the rubber section that connected to the engine. Next, I took a Dremel tool and carefully cut away the crimped collar. Passing the point of no return, I sliced off the rubber part of the hose, revealing an intact barbed end on the remaining metal section. While I don’t doubt that a crimped-on fitting is the best way to secure the hose, it’s only as good as the hose that it’s grasping, and this hose was leaking.
![porsche engine fitting before clamp](https://hagerty-media-prod.imgix.net/2024/07/Hack-Mechanic-Kluge-before_clamp.jpg?auto=format%2Ccompress&fit=crop&h=480&ixlib=php-3.3.0&w=640)
I’d recently had some hydraulic hoses made up for me by a trusted local shop for an unrelated application, so I took them the cut-off section of rubber hose and a new fitting and had them match it to a piece of high-pressure hydraulic hose that was massive overkill for this 100-psi oil scavenging line. They verified that the new hose was the correct interference fit for the barbed fitting, cut the hose a bit longer than I needed it to be, and crimped a band clamp onto the fitting that went to the wrap-around pipe on the engine.
![porsche engine hose fitment](https://hagerty-media-prod.imgix.net/2024/07/Hack-Mechanic-Kluge-other_end.jpg?auto=format%2Ccompress&fit=crop&h=480&ixlib=php-3.3.0&w=640)
I then ordered a couple of high-quality T-band hose clamps from McMaster-Carr. I installed the end of the hose that went to the wrap-around pipe, cut the other end to length, slid the T-band clamp over it, pressed it over the barbed fitting I’d cut the original rubber section from, tightened the clamp, and verified that it didn’t leak. I planned to drive the car back to the hydraulic shop and have them replace my T-band clamp with a crimp-on clamp, but it didn’t seem necessary. If anything, the T-band clamp was wider than the crimped band clamp.
![porsche engine hose clamp closeup](https://hagerty-media-prod.imgix.net/2024/07/Hack-Mechanic-Kluge-with_clamp.jpg?auto=format%2Ccompress&fit=crop&h=480&ixlib=php-3.3.0&w=640)
I posted all this to the Pelican 911 forum in 2008. Some folks jumped all over me for such an, um, hack, and said that the only question was whether I’d find the 11 quarts of oil in my driveway, or if it would just dump out on some lonely road.
But others agreed that this was a reasonable approach. The work was vetted by the hydraulic shop; it wasn’t like I’d used cheap heater hose and hardware-store hose clamps. But to each his/her/their own. Some would call yanking everything out, letting what would break, break; then replacing it all with new OEM parts, “Doing it once, and doing it right!”
I’d call it an expensive and unnecessary cascading failure. I was glad to avoid it, and make no apology.
![engine hosing porsche](https://hagerty-media-prod.imgix.net/2024/07/Hack-Mechanic-Kluge-IMG_8638.jpg?auto=format%2Ccompress&fit=crop&h=480&ixlib=php-3.3.0&w=640)
It wound up being another three years before I sold the car. It never weeped another drop of oil out that hose.
In writing this story, I went back to the Pelican 911 forum for the first time in at least a decade and searched for other examples of this problem. Sure enough, there were similar posts from people who felt that it was reasonable to arthroscopically repair the leak as I did without risking the whole slippery-slope-cascading-failure thing. Two of the posters used multiple crimp band clamps to replicate the width of the original crimped fitting, something I hadn’t thought of.
Sixteen years later, I’ve yet to come up with a better “good kluge.” Yes, there are times when even I think that “Do it once, do it right” is a reasonable approach to a problem. But I hate it when people intone it like it’s a moral imperative. If you strip a fastener, you don’t need to think that you’re doing something against the laws of Man and God if you don’t follow through, not stopping until you’ve restored the whole car. I’ll freely admit that not all of my hacks and kluges have been successful, but I look back with particular fondness at this one, and still regard it as the gold standard.
***
Rob’s latest book, The Best Of The Hack Mechanic™: 35 years of hacks, kluges, and assorted automotive mayhem, is available on Amazon here. His other seven books are available here on Amazon, or you can order personally inscribed copies from Rob’s website, www.robsiegel.com.
Check out the Hagerty Media homepage so you don’t miss a single story, or better yet, bookmark it. To get our best stories delivered right to your inbox, subscribe to our newsletters.
IMHO, any simple, durable, user-repeatable, improvised, and effective solution, that still allows for an accurate restoration, someday, is an excellent solution, overall.
Dissenters reactions are likely based on a (somewhat) reasonably obsessive compulsion to avoid “violating originality”, as if every significant vehicle is to be restored to a level of accuracy to compete for an award. After all, we all continue to hope, responsible engineers with any pride and integrity, typically analyze and conclude with justifiable reasons, to design things the way they thought best at the time.
However, generally-speaking, everyone should realize, with an arguably few exceptional examples, automobiles are not appreciating assets, but subject to increasing entropy, primarily manufactured to simply generate profit, and (hopefully) enjoyed by their temporary owners, but certainly not to be curated, sacred artifacts.
Excellent post.
I wouldn’t call that a non-standard repair, ie a Kludge. I’ve done that a number of times, both on oil cooler and automatic transmission lines. A band crimp is great because it can come loose like a bolt clamp, particularly useful in an inaccessible location. On the other hand it is also bad because you can’t really tighten it in the future when the inevitable compression of the hose happens.
So combined with the fact that it this application it doesn’t disturb a bunch of other connections or problem points I’d even call it a superior repair.
Kludge – rhymes with “sludge.”
Kluge – rhymes with “huge.”
kludge (plural kludges)
(informal, electronics, engineering) An improvised device, typically crudely constructed to test the validity of a principle before doing a finished design.
(informal) Any construction or practice, typically crude yet effective, designed to solve a problem temporarily or expediently.
(informal, computing) An amalgamated mass of unrelated parts.
(informal, computing) A badly written or makeshift piece of software; a hack.
—
kluge (plural kluges)
Something that should not work, but does.
A device assembled from components intended for disparate purposes.
Perfectly good repair in my book. Plus, the cooling lines run low pressures so no worries there.
This reminds me of the time I pulled my E28’s evaporator to fully rebuild the HVAC. Once I had it in my hands, I discovered the big flare nut for the expansion valve had a crack all along one side.
Since there were no apparent leaks, I carefully put it back in the rebuilt box, recharged the A/C and had cold air for years.
Someday I’ll have to get in there and fix it properly, but at the time I was like “Nope, Nope; won’t go down THAT rabbit hole right now” hahaha
That’s it. Exactly.
I wonder if the purists who slammed you on Pelican were also of the view that you should have removed the cylinder heads to fix the return tubes, rather than using the sensible—but, gasp! Non-original—approach of using collapsible drain tubes?
Perfect is the enemy of good. That being said, that is a perfectly cromulent fix and I would’ve likely done similar. That is a good kluge!
Cromulent!
I come here to be entertained, enlightened, and educated on many subjects, mainly centered around cars and driving. Little did I know that I was also going to expand my vocabulary as well. Gonna use some of this stuff at the next Bench Race and see how many of my friends think I’m cussing at them! 😁
Yeah… Colton made me look it up too.
My “great” vocabulary turned out to be — you know.
👍👍
Like the extendable drain tubes, is it really a kluge at all, when it replicates the OEM function and durability (for the most part)? Wrenching on any vehicle, you will find parts or parts of systems that were designed for ease of installation (remember, the vehicle is actually moving while these parts are attached), cost effectiveness or because that’s what the supplier shipped. Look no further than those plastic parts in a BMW cooling system, just waiting to fail. Replacement with better, more appropriate materials is not OEM, but just smarter. So, I submit for your consideration, a kluge or just a genius at work?
The oil leak thing has always appeared on many air cooled engines. Harley Davidsons, Corvairs, VW, Porsche.
Even some air plane engines. Older engines were never sealed up as well as they are today and the use of oil coolers and other oil intensive elements just add to the leaks. Even till Kawasaki got into the tractor engine market almost all Briggs and Techusa engines leaked oil.
As for finding a non factory fix. I have done that all my life. Some this are just not right from the start and you have to get a little creative ands re engineer some things at time.
One I fixed the factory could not figure out. I had a car with a Mass Air sensor from an Alfa that was used on a Chevy. It was a turbo tune kit. The use of the Alfa sensor made it so it would hit the AC pump on hard acceleration. I created a shield to protect it before the MFG even knew where the issue was.
This was a case the sensor needed to point down not back and it would have had clearance. Everything was good till you noted the engine travel on the engine mounts. The engine had 315 FT LBS of torque so it would move much more than stock on the mounts. .
If it fixed the issue it’s not a kludge to me. You were just being innovative.
Zen and the Art of 911 Maintenance
Geordle – Yep. – Rob, while I’m sure you’ll be praised by many for your ingenuity but there’s more to it than that. I think all of us who’ve had older cars have found themselves in the situation of having to jury rig our vehicles just to get them/us home and do a proper fix without causing more damage. It’s good practice and better than waiting for the wrecker that will be there in an hour and shows up three hours later..maybe. ‘I just want to get her in the driveway and then…’ Almost anyone can swap out this and that. Too many mechanics follow the ‘if it doesn’t fit force it..if it breaks it needed to be replaced anyway rule. Of course if your just adding it to the bill regardless why not . It’s the thought process that makes all the difference. If all you’re ever doing is removing A and installing B, sorry, you don’t qualify as a ‘Master Mechanic’ no matter how long you’ve been doing it. How many times have we all seen something that broke and asked – ‘ Why the hell did they do it this way? This was a problem waiting to happen .’ -Sure you could just replace the parts but sooner or later you know it’s going happen again. Have a cup of coffee or two , think your way through it and find the better solution. Your car and moreover your brain will both be the better for it.
Looks like a great repair to me. And let’s get real about originality. It’s nice when you can do a repair that preserves the originality of a vintage vehicle but most driver-quality machines will not be worth one cent less if a nice appearing, dependable, non-standard repair is done instead.
Oh so many OTSES (On The Spot Engineered Solutions; not KLUDGES) for the M38A1, TY36 Elan, Merkur XR4ti, wada wada. One of my best was creating an 1=1/4 inch copper pipe/fitting intake manifold to adapt a 36mm Mikuni Kaw-we carb to my dual port VW airboat engine. After I tapped in a 1/4 inch fuel return line from the float bowl to the tank it ran like a charm.
I am just careful to always park my 73 VW in the street instead of anybody’s driveway. Not exactly a “kludge” but problem solved.
If you are taking the car to Pebble beach, you’ve done a terrible thing. If your purpose is to have a car to drive and enjoy, this is what you do, and how you do it.
Four oil return tubes, though, not six as stated in the picture caption.