Fun with temperature shutoff gauge

[Jens]

I have an interesting situation happening that I would like the opinion of the membership on.
First, a bit of background. My coolant flow is from the cold side of the heat sink (hot water storage) to the exhaust heat exchanger thrugh a circulation pump to the engine through about 6" of pipe to a thermostat, then it flows past the temperature sender and back to the heat heat sink.The temperature sender is attached to a Murphy switch gauge set at about 220F to trip on overtemperature, The thermostat (195F) has a 1/16 hole in it to allow minimal flow even if it is closed. The system has run for a while so I suspect there are no air bubbles in the cooling system.
On startup, I run for a short time with a half load (4500W) on this 20/2 and then I switch to full load of around 9000W or so. Last year I ran on 4500W load and the issue I am about to explain did not occur.
What happens is that after about 5 to 8 minutes of run time, the engine reliably switches off by tripping the tatteltail that is hooked to the temp gauge. Once the tattletail is reset, the engine runs without further issues until the end of the scheduled run, no more unexpected shutdowns.

Last night I came up with a far out explanation. The thermostat is slightly removed from the engine and is closed at the start of the run. There is cold water in the pipes surrounding the thermostat. There is a small flow of coolant through the 1/16" hole of the thermostat. The engine, running at full power produces a lot of heat fairly fast. It takes a while for this heat to reach the thermostat to open. By the time it opens, the engine is running at let's say 230F for argument sake. The thermostat cracks open, the hot water hits the temperature sender and it rapidly climbs past the 220F set point and trips the temperature shutdown. Once the shutdown is reset, since the conditions are different now, the false trip never happens again. Running at 4500W load the false trip never occurred since the engine did not heat up quick enough to overshoot the thermostat by as much.

Well, this morning I put this theory to test. I started the engine, loaded it at about 9000W and watched the temperature gauge. It slowly crept up until it got to about 190 or so degrees F. The hot water end of the exhaust heat exchanger was getting mighty hot as the water slowly trickled into the engine through the 1/16" hole in the thermostat and on to the heat sink.

At one point, as I was watching the temperature gauge, the needle JUMPED from the approx 190F position, slammed into the over temp set point contact and immediately returned to where it started, the 190F approx. This tripped the engine shutdown as suspected.

The temperature needle did NOT rise rapidly as I would expect.  The event to a fraction of a second. There are two explanations that I can think of. Firstly, I vastly under estimate the response time of this gauge and it is able to register a temperature change almost instantaneously. The second possibility is that somehow the needle hangs as the temperature increases. The sending unit is a bulb type of a device with what I suspect is a capillary tube going to the gauge. I am thinking that the pressure in the capillary tubing gets to a point where it pushes the gauge past the blockage, the needle shoots up, hits the trip point and immediately falls back to where it was.

Since farting with the temp sender is kind of out, I would like a couple of opinions - Is the response of a capillary type temperature gauge fast enough to cause a needle to instantaneously jump like it did, presumably at the point when the thermostat cracked open ? Has anyone ever seen a needle movement hang up like this before - is it something that is possible/reasonable and is it fixable or does it call for a meter replacement ?

My next test run is tonight and I will set up with a 4500W load. I would expect that if the meter movement is impaired somehow, I should get the same reaction as under this mornings' run with the temperature slowly climbing until the meter movement releases and jumps to hit the over temp alarm.

Thanks in advance for any opinions expressed.

Jens

[vdubnut62]

Jens, you COULD have a "sticky" thermostat. Or as you say, the gauge could be bad. The gauge works on a rack and pinion principle, the teeth are thin and tiny, so it wouldn't take much at all to mess them up.
Could you take the gauge out and dunk it in boiling water to see if it jumps?
I have a Dodge 3500 with a Cummins diesel that has always given me fits with the thermostat. It will almost overheat, then the thermostat opens,
and the temp drops to about 120. Got to be hard on the cast iron block. This has continued through countless thermostats, two waterpumps,
and two engines.
Ron

[cognos]

Although I'm not familiar with your gauge, I've seen plenty of mechanical temperature gauges that respond very quickly to temperature change. But I can't say I've ever seen one "jump" up or down - that would lead me to suspect a faullty gauge, or a very unusual operating condition...

I'm sure you know all this stuff.

Mechanical temp gauges in my experience are more like the second hand on a mechanical watch - progressive and smooth - than on a quartz watch - jumping from second to second.

There's a bourdon pressure tube in there that engages a gear, or a series of actuator levers, to position the indicator needle. I have seen the gears wear in pressure gauges where there is a lot of high-frequncy fluctuation. But I can't say I've ever seen it in a temperature gauge.
A failure of the actuator-lever type usually results in a disconnected non-moving indicator.

I'm not sure I understand your flow. If I'm reading this correctly, the coolant water is heated in an exhaust exchanger prior to entering the engine? I must be misunderstanding this... I need to see the flow diagram, I can't hold 3D stuff in my head anymore... too much crap built up over the years...

If this is in fact the case, I can see that on thermostat open, a small slug of very hot water could be flushed from the exhaust heat exchanger and introduced into the engine, causing a momentary high-temp condition, causing the trip.

But that doesn't explain why it does it now, but was OK before. Look for what has changed since then.

[XYZER]

I have found that a thermostat with a 1/16" hole will tend to open and close until it reaches operating temperature. I drilled 3 or 4 1/16" holes in my thermostat to allow a more constant temperature rise but will also open if the demand requires it to. My 6/1 still takes the same time to reach operating temperature but without the thermostat cycling. It might help yours situation and it might not.....

[BruceM]

Thanks,  Dave, for the tip on extra holes in the thermostat reducing fast cycling.  My first thermostat only lasted 2 years of light but daily use. 
Best Wishes, Bruce M