How you cheat fate out of $20,000

[Genesis]

Seriously.

I don't know if my deal is with the guy in the white robes or the red suit, but I'll take whichever it is.

After our aborted trip to the Angelina last week (a wreck we intended to dive which was called due to a line of boomers bearing down on us) I decided to pull down the coolers on the starboard engine and have them cleaned, as the port (which I did) was running nice and cool and the other a bit warmer.

No big deal, other than waiting for Saunders to get me the new seals and gaskets that I needed to finish the job (remind me to shoot them sometime)

Anyway, I put it all back together and lit it off the other night. Flawless.

So this morning I tool over to the boat to clean up the mess. See, when you pull the blower on a Detroit Diesel, oil goes EVERYWHERE. Its unavoidable. This, of course, make a hellish mess in the engine room, which you then need to clean up. So to the boat with the steam cleaner and rags I go, along with a gallon of simple green and a slop bucket to wetvac the mess into.

I get there and happen to glance down into the OTHER (port) engine's drip pan. There's a bit of coolant in there. Now, this is an engine that was NOT touched. There shouldn't be ANYTHING in that pan. At all.

Coolant leaks are not a good thing. So I start feeling around looking for what's wet. I find it - its a fitting on the bottom of the main elbow coming from the heat exchanger back to the water pump. Ok, so its gotten a bit loose I think to myself. I reach for a 9/16", touch the fitting, and it breaks off in my hand.

Oh shit.

I now have a finger in the dike (literally) keeping the green firehose from spraying all over. I managed to find a bolt and a piece of oil rag that would stanch the flow to a dribble while I grabbed two 5 gallon buckets, hoses, and funnels (which I had on board as I just finished draining the OTHER side to do the work on it!), and opened the other drain valve to get the coolant out of the engine.

I then examine the fitting that I have half of in my hand, and discover that there is a crack that has been there for quite some time (probably since installation about a year ago) which apparently had propagated all the way around! It looks like a casting defect to me, as there's no evidence of it shearing due to torque or other loads. (A year or so back I replaced the drain cocks on the elbows and oil coolers with fittings and ball valves with a hose barb on the end, to make draining the cooling system a non-eventful thing. I remove the handles when not in use so I don't accidentally drain the contents of the system while underway!)

Anyway, the rest of the fitting is broke off flush (of course) and the fitting IT is screwed into won't come off the elbow (cast iron .vs. brass, not a good combination after 20 years.) So I pull the elbow, get my Easyouts, hammer in the appropriate one, and extract the piece of the nipple that was still in there. Then I get to make a new gasket out of gasket paper as I don't happen to have one of those laying around, change the hose to which it was connected (just for good measure), and stuff it all back together with a new 97 cent nipple.

There went four hours of work, including the cleanup of THAT mess.

However - if that had failed offshore 20nm+ out, dumping the entire cooling system into the bilge, that engine would have almost certainly been destroyed. Once the temperature sensor is uncovered you get no warning of the overheat-in-progress until you either get a fire or a seizure 90+% of the time. Since the breach was directly in front of the waterpump, it would have been catastrophic and dumped the entire system inside of 30 seconds, as it would have been pumping under full system pressure.

I estimate that another couple of hours of engine runtime and it would have cracked right off. No doubt about it.

I dodged a major bullet here... best case is I would have had to come in on one engine. Worst case is that I would have trashed the other engine at the same time, and Gig would have been out of service for at least 2-3 weeks for a rebuild, with a parts cost alone (if I did all the work) exceeding $10k.

As it was, the damage was 97 cents.

I'll take it... now I know the call I made Friday was the right one, even if not for an obvious reason ......

Lesson? Always look in the engine room and if something doesn't look right, it isn't. Find out what's up before you get an expensive surprise!

 

[Traveler 45C]

Oh brother…

Sorry to hear of the failure but glad it happened at home and you caught it before it caught you.

Was it one of your fittings that failed?

Coolant level monitors are in your future, I’m sure. I want them.

I must have made a similar deal as you: While doing a pre-flight before pulling out, I notice a wet spot on an exhaust hose. It’s spitting from a bubble ready to let go. No high-water alarms on that boat. That trip got cancelled.

 

[Genesis]

Yep.

The fitting in question was about 18 mionths old. Other than the leak there was NO WARNING that it was about to split, and when I did the cold checks on the previous run the coolant was not there.

In other words, the total engine runtime from first ability to notice the failure to complete failure was about 5 engine hours.

That's pretty scary.

The others (there are four total valves, two on each engine) LOOK ok. However, so did this one up until the point that it came apart in my fingers!

I've gone over to Swagelok and ordered up $50 worth of stainless steel HYDRAULIC RATED (5000 psi) nipples and street "Ells" to replace the brass ones. This is not a place I can afford to have a fitting drop off, and the $50 is cheap insurance. The royal bitch will be draining the systems (AGAIN) to put in the new ones, but the day they show up its happening.

BTW there is no good coolant level alarm fix that helps you in a failure like this. With a SLOW leak, yes, but not in a situation where a fitting blows off the dumps the entire contents of the system in seconds, which this would have. The problem is that the internals of the engine are so hot that if the coolant disappears even an instant shutdown probably doesn't save you from a complete rebuild, as the odds of the engine surviving without cooked liner seals and/or a cracked head are pretty close to zero.

Pay close attention to your hose condition; there are two hoses on each exhaust manifold that are in the back of the manifolds and often neglected (they're NOT fun to get to for replacement), there is the connector hose from the aftercooler to the cooling system (if you have an aftercooler) which is similarly buried, and then there are the hoses on the thermostat housings and water pump that require partial disassembly or removal of components to get to them.

A blown hose will screw you just as fast as will a failed fitting, and some of those almost NEVER get changed by most people...

 

[Traveler 45C]

Yea, that is scary!

When I get to it, and it is on “The List”, both mains get a cooling system cleaning and new coolant. And while I’m at it, fittings such as yours and new hoses.

I don’t know when the last time the cooling system was cleaned or the hoses replaced. So I’ll just do it.

 

[SEAWISE]

Your right Genesis once that coolant level drops away from the temperture sending unit, it doesn't have anything to sense. I was wondering if you could rig up a low pressure gauge, one that went only to maybe 10 or 15 pounds. Then if the coolant, along with the pressure fell, well you could rig up a low pressure alarm. Or maybe an oil pressure switch like the one in your car, that hooks up to a idiot light would work with coolant. I dunno, something to think about! If it would save thousands, its worth some thought.

 

[Genesis]

Nope - won't work.

The normal system pressure is extremely low, and the full "cap pressure" is only 7 psi on these engines.

A liquid level sensor SOUNDS like a good idea, but in the event of a catastrophic failure all it does is tell you that you need to limber up the checkbook, because once any part of the block goes dry you're immediately screwed. I don't think there's enough room for warning on a liquid level alarm, and I'm not sure you could add that anyway given the design of these tanks and systems.

 

[MikeP]

Wow, you did cheat fate. Usually, the easy-out breaks! Nice one !

Did you put in new antifreeze as per the old saying that any liquid removed from the engine never goes back?

 

[Genesis]

Not in this case, but only because I didn't have any handy and it was out for about an hour in clean 5 gallon pails.

If I had 5 gallons of EG + SCAs handy I would have

 

[DaveH]

Genesis:

Hi there. I'm newly registered to the forum but have lurked here for a few years.

With regard to your recent stroke of good fortune, I thought I'd try to problem solve along with you. I have a technical background (engineering degree) and have sold industrial equipment for about 15 years now.

Here are a few choices:

Pressure switches: There are pressure switches available that will have low dead band (less than 1 psi) and set points from tenths of a psi to about 20 psi. The problem is setting an initial alarm delay until the coolant system gets up to pressure. Immediate loss of pressure would show catastrophic failure.

Differential pressure switches: This is one that I need your help on. What is the pressure range that the water pump puts out from idle to WOT. A DP switch could be set up between the water pump high and low pressure side to alarm if the differential does not meet the minimum.

Proximity switches: A capacitance probe (detects the presence of liquid) could be mounted to show the loss of fluid in the high point in the system Unfortunately, a baffled coolant reservoir tank would be needed due to wave action sloshing the level around quite a bit.

The other alternative is to mount a turbine flow meter inline to measure the positive displacement of fluid through the system.

What are your thoughts?
-Dave

 

[Traveler 45C]

Welcome Dave.

Could a proximity switch be mounted on one of the coolant pipes to detect liquid?

Ya’ know, a turbine flow meter would let you know of a coolant loss instantly, and you just might, might have time to get down there and start pouring in fresh water from your tank until the engine cools.

But where would it be mounted? It would have to be high up in the system.

 

[Genesis]

None of the above help.

The problem is that once coolant loss happens its already too late. The fire in the hole is well north of 1000 degrees celcius. Differential cooling will cause immediate cracks, overheating of the liners, even transiently, will cook the seals (they're an elastomer compound which CANNOT take high temperatures), etc.

If you uncover any part of the liner or head's coolant passages and then reintroduce coolant (or water) you get shock cooling which will crack the head. This is one of the reasons that coolant system integrity is so important, and so is proper coolant composition - localized cavitation or boiling can cause the same sort of damage by the exact same mechanism, even if it APPEARS all is ok.

Basically, by the time the alarm - whatever it is - goes off, you're screwed if it happens under significant power demand. The alarm would save you at idle in some cases, but not at cruise power.

Even an immediate shutdown doesn't help you - you'd have to be able to shut down BEFORE any part of the metal in the block is exposed. This would require both a capacitative pickup and a near-instantaneous response.

The capacitative pickup won't work without drilling holes, since the tank is cast iron (can't sense through it), and it would have to be baffled, as you note.

The other problem is that in a closed system when the dump happens there's nowhere for air to come in. What happens is that pressure is relieved and the coolant BOILS; the driving force and the "air space" is created by the steam generated. Unfortunately the release of pressure happens system-wide and localized boiling in the block occurs immediately as the pressure is released, meaning you have instant damage - even if you get an immediate alarm and were to wire it to the shutdown solenoid so it would go off instantly it would be unlikely to help.

Any high-output engine is in this same general pickle in the event of a catastrophic cooling system failure.

Basically, low-coolant alarms don't do much. You should pop the caps and dip the H/E tanks as part of your cold checks every morning anyway, so their use in getting rid of that is questionable. As far as protecting you from an under-power loss, it'll work for small leaks (the coolant level sensor) but a catastrophic plumbing failure that leaves you with any significant sized hole is going to screw you.

I replaced all the nipples and elbows in this configuration with hydraulic-rated stainless steel parts. That's the end of that problem $50 for the fittings (they're EXPENSIVE from Swagelok) but heh, how much are the parts for an overhaul?

 

[DaveH]

I agree with all of your statements. There is nothing you can do about a large scale coolant-side failure to prevent engine damage of some kind. However, if the leak is small enough, a differential pressure switch across the water pump may allow an emergency shut down that could prevent the large boat dollars from being spent.

 

[REBrueckner]

I posted about the issue of normal engine vibration causing failure in brass nipples a few years ago....I had a similar issue to Genesis on a 6V53 many years ago....ALWAYS use steel or stainless steel, use the shortest nipples possible, and never hang a heavy valve of any kind on the end of a long nipple....As Genesis noted they will fail within about two years...
It hard to believe, but true.

An alternative solution is a short nipple with a length of hose to an engine room mounted valve....off the engine....This would be analogous to oil pump out hose connections to a engine oil pan.

Also, Murphy gauges for coolant level monitoring are better than nothing...no guarantees, but saying its hopeless to monitor coolant level is like saying what Genesis experienced is never going to happen. So maybe you lose one cylinder instead of three or four.


When making periodic engine room checks, giving a moderate yank on all hoses as a regular part of your inspections is a good practice....I usually forgot to do during the season , but during spring "get ready" activities when I'd get all my hatches pulled up, and could reach everything, I'd do this. That's when I discovered the almost fractured nipple many years ago...it broke off...the hose was fine!!!!
And as noted above, keeping absorbent material under engines and noting any change for immediate inspection is excellent practice.

 

[krush]

Once the temperature sensor is uncovered you get no warning of the overheat-in-progress until you either get a fire or a seizure 90+% of the time.

Since this is a thread from the dead I've never read.....

All depends on the sender. Many temperature senders screw directly into the head/block/engine but poke into the water jacket too. If coolant is lost, the head gets hot and the gauge reflects this--even if no liquid is touching them.

And how long the engine will run without coolant before roaching is a tough one to answer. The mass of the engine allows it to store quite some thermal energy before temps skyrocket---but not very long. I doubt many people are going to experiment though. Regardless, I doubt it is an instantaneous failure.

 

[Genesis]

On mechanical Detroits the temperature sender (and alarm switch) is in the water elbow at the thermostat, which happens to be the high point in the system AND not directly associated with the head.

If you lose coolant on those engines the sender will be uncovered immediately and the overheat alarm will not go off until you're badly cooked, especially on the 92 series which are wet-liner. Drop the coolant on one of those under load and you've got single-digit seconds to figure it out before the liner seals (really just a high-temperature O-ring) are charcoal.

A cylinder-head-temperature switch might not be a bad idea but it won't trip fast enough to save the liner seals. It might trip fast enough to prevent a seizure but it doesn't matter (much) if you have to replace the kits anyway due to the liner seals being compromised.

 

[REBrueckner]

Genesis makes an excellent point in post # 15...92 series wet cylinder liners ARE very susceptible to coolant loss.

Also, operation at moderate RPM of say 1200 to 1500 RPM which many do today to improve fuel efficiency, gives those operators an additional brief recovery/reaction time for shut down before catastrophe ensues.

On the other side of the coin, does anyone know what size connections Murphy gauges typically utilize?? If it's small, the drain down time of the alarm container to an alarm activatio level could be a problem if there is a massive coolant loss elsewhere....

 

[Nonchalant1]

Karl,

Glad you dodged a big bullet. In your original post, you used a lot of plurals:

"I replaced the drain cocks on the elbows and oil coolers with fittings and ball valves".

Sounds like you may have 4 of the suspect fittings on board. I'd replace the other three, particularly if you think it was a casting defect.

Doug

 

[Genesis]

I replaced them all; the boat has of course since been sold.

The offending pieces got replaced with expensive 5kpsi rated stainless fittings on the premise that whatever casting defect was in the brass, it sure as hell wouldn't make it through testing for the stainless versions with an actual pressure rating on them.