The Vexing Case of a Perplexing Panda

[jeremyshaw]

It's two weeks now since I posted that a friend of mine in the Caribbean is having trouble with the cooling system on his Fisher Panda generator. He still is and despite the help of many seasoned boaters and local mechanics, they can't diagnose the cause of the problem.

The problem is overheat. Overheat of coolant arising from the introduction of air. Pump, pump seals, hoses etc. have all been checked and or renewed, and the whole system pressure tested, but it's still happening.

My belief is that it must either be a problem with the cylinder head gasket (but there's no misrunning or contamination of coolant) or a problem with the expansion tank - the Kubota engine uses a rather unusual pressurised expansion tank.

Any thoughts on what this must/might be and or ways to get a definite diagnosis? He doesn't want to rip the cylinder head off unless he's sure..

A rundown of the position follows:

The 'mystery' air begins to displace coolant as soon as the engine is started - like 2 or 3 seconds into it. If the cap on the expansion bottle is put on (it is supposed to be off during bleeding) the air displaces at a slower rate. With it off, the air displaces the coolant and sends it to the bottle. Air ends up filling the cavity where the CFW pump is whirling and (since I'm using a clear hose now) it is obvious the flow of coolant simply quits when enough air is present. My thought was to try and test the air and see if it had a diesel aroma or burnt diesel aroma, but I've not been able to detect anything.

Since with the engine off I do not lose pressure, it would seem all hoses and gaskets in the coolant side of things are intact. So that leaves areas of greater pressure than my tests to diagnose. The only possible area where greater pressure is present in this area that contains air would be inside the piston chambers when the engine is running. So. I'm thinking a bizarre gasket leak blow by at the top of a valve. I've been told that head gaskets or warped heads would usually cause an engine to run very poorly. I've not noticed this. Also I've been told I'd lose coolant when doing a pressure test, and I don't notice this either. So as you see this is very strange indeed.

 

[duncanmack]

Hi Jeremy,

Been thinking about this problem - I've a garage and we do lots of breakdown work. I am the "diagnostician" and I love stripping things to check my diagnoses. Fortunately I have a team to put the things back together again!! /forums/images/graemlins/grin.gif

The 'mystery' air begins to displace coolant as soon as the engine is started - like 2 or 3 seconds into it

Ok. That's easy. The "air" is combustion gases. No possible doubt about that. As you rightly suspect from the cylinder. I don't know the engine in the unit, but that doesn't really affect things.

The difficult bit is how are they getting out of the cylinder and into the cooling system.

1 Cylinder head gasket. Possibly. It's not uncommon for there to be a very slight leak across the gasket which will cause this ( sometimes with the system ever so slightly overpressuring and losing small amounts of coolant - the cap on the expansion tank copes with the XS pressure) BUT a pressure test of the cooling system will usually, but not always, get some coolant back into the cylinder. I've seen loads of vehicles with this problem. Even owned a few.

2. Cracked head.
Hairline crack into the water jacket. Or a slightly porous head. The gas will get out of the crack (300+psi on cranking, god knows what on firing) but the coolant won't necessarily get back in (20 psi).
Difficult to prove, but a head off job and a pessure test by an engineering shop should show a problem. Sometimes repairable. Had a few of these too.

3. Hairline crack in a cylinder wall.

Same scenario as 2. above, but a swine to prove. Sometimes the crack is "easily" visible, sometimes not. It is possible to test it but that means a complete strip down. Seen some of these too. Best tested by elimination of easier testable components.

So, the way forward?

Head off and get it pressure tested and tested for flatness. When OK then refit with new gasket.
If problem persists I'm afraid you need the engine out, new liner(s) if possible or new block if not.

On a lighter note, does your friend have new for old insurance covering the accidental dropping of generators into deep water? If not, can he get it? /forums/images/graemlins/grin.gif

Could you keep us posted about the progress and outcome please?

 

[jeremyshaw]

Duncan
Thanks for your very helpful response. Sounds 'on the money' to me! I will report back, though as my friend is in Bonaire I fear it will take him a few weeks to sort it, given it sounds from your diagnosis, like he needs professional help.

Thanks again

 

[ccscott49]

I concur with the above, air /combustion gases the dead give away is the speed at which it enters the system. I've also seen this and sometimes it is a head gasket, that's split as has been said high combustion pressure, weak test pressure. Looks like a genny out jobbie.

 

[MacW]

Further to the fairly comprehensive diagnosis above, there is something else: There are some engines, due to their design, once having been emptied of coolant, for renewal or whatever, require PRESSURE BLEEDING according to a certain procedure, using special equipment, to get all the air out of the system. Failure to do things this way leads to the symptoms you describe, where trapped air causes a hot spot ( in seconds ), which rapidly expands the trapped air, forcing the water out. If this engine requires this procedure, there is no way round it. Quite a number of car engines are now like this. A handy tool to aid diagnosis is a dye, available from motor factors, which when added to the header tank, changes colour if combustion gases are present, or in this case, perhaps more importantly, doesn`t.

Pete

 

[MacW]

Further to reading your rundown again, you say " air is collecting in the cavity where the pump is running " Normally, air in the system will naturally try to rise to the highest point, unless in the scenario in my previous screed. Where is the pump getting this air ? For air to "collect " that fast and stay in the same place surely it must be locally produced or sucked in. Is the pump mounted on , or in the block? Is it driven by a shaft from the crankcase or by a belt ? If shaft, is ( excessive?) crankcase pressure leaking into the pump past the seals. Are the shaft seals in the right way round? Shaft or belt, is the pump body cracked or pinholed, or the pump gasket face warped, in both cases sucking air when the pump is running but not leaking water stopped or running, both conditions making the pump cavitate, and creating a `bubble` which forces water out ?
Hope all this helps
Pete

 

[jeremyshaw]

Thanks for the additional thoughtful and useful comments.

Re the need to bleed the system, yes the Kubota engine has, unusually, a pressurised expansion tank and thus needs careful bleeding. I know my friend has taken a lot of trouble to do this properly and had assistance.

Re the pump being a potential culprit I agree. The pump was weeping which suggested a failed seal. (though would it let the air in that fast?). However the pump and seals have been replaced which is why we've now ruled that out. Of course they could have been put back in wrong, but it seeks unlikely.

Most likely cause does seem to be a cylinder head problem sadly...

Thanks again.

 

[duncanmack]

Quote

"A handy tool to aid diagnosis is a dye, available from motor factors, which when added to the header tank, changes colour if combustion gases are present, "

The dye detects carbon monoxide (CO). So it's effective in diagnosis of problems in petrol engines.
Diesels produce very little CO. Being much more efficient in combustion of the fuel they produce mostly carbon dioxide (CO2).

As far as I'm aware, there is no chemical available (readily) to check for Co2. We use an exhaust analyser to analyse the "air" in header tanks in diesels and pick up CO2 - pretty effective!. If you can get a hold of a mechanic with a portable gas analyser......?]

I'd considered the problem of a "hot-spot" due to incomplete filling of the cooling system, but ruled it out due to the "2 to 3 seconds" for the problem to manifest. Had it been 20 to 30 seconds it was certainly possible. Had it been a few minutes then a stuck shut thermostat or a failed water pump could have been possible.

I'll stick with the combustion gasses as being the only way for the problem to happen so quickly.

BTW Jeremy, what happens if you fill the expansion tank to the brim and then start the engine with the cap off?

Do you get a "geyser" effect?

 

[fishermantwo]

I have a 3 cylinder Kubota generator in my fishing vessel and over the years I have had some over heating problems. The motor itself is likely to be bullet proof, mine has about 10,000 hours on it and uses some oil now. Generally problems like this would be associated with the marinisation process but I suspect that has been carefully checked.

Has the water pump on the front of the engine been replaced? Impeller may be spinning free. There is a small water pipe on the front of the pump to the thermostat on some model Kubotas. This has a habit of corroding off.

To expel air out of my system I had to connect a fire hose to the bottom hose. Force water though till it ran through with no bubbles and quickly re connect the bottom hose.

In this case however the 2 to 3 seconds is looking like a serious cylinder head leak and bubbles in your clear pipe would have to be visible, plus of course looking down into the expansion tank there would be signs of combustion.

 

[VicS]

Before tearing it apart i would seriously consider the suggestion that Seaboots made about a trapped pocket of air which expands, as the engine warms up, and ejects the coolant. Repeatedly allowing it to cool while returning the coolant then trying again may do the trick. I had the problem with a car engine once. Eventually I could get it all filled with no trapped air then it was OK until next drained.

 

[jhughes]

My 5 series BMW car developed a porous head after 175,000 miles and had similar symptoms, a bottle of Rad sealant once a year sorts it out (I am on 255,000 miles now).