# Should ice take longer to melt when used to run a heat engine?

• Tom Booth
In summary: From what I've seen of Stirling engine models, they are just demonstrators of the concept without any real energy production.So to better test the issue you should run the experiment with a load applied to the engine.
Tom Booth
If a heat engine converts heat into "work", will ice used to run a Stirling heat engine last longer than ice allowed to melt by itself?

To try and answer this, I obtained a Stirling engine and ran this experiment:

With the engine running:

And not running:

Without the engine running the ice melted in 28 hours. With the engine running, it took five more hours (33 hours) for the ice to melt.

Conditions could not be held completely constant, the wether got a little cooler during the time when the engine was not running.

Only one engine was used, running and then not running consecutively rather than concurrent.

One person on a Stirling engine forum predicted that the ice would melt faster when used to run a heat engine because the engine would be actively transferring heat from one side of the engine to the other, taking heat from the warm side and transferring it to the ice, but the results of this test were the opposite.

Someone else on another physics forum recently predicted that the ice would melt faster without the engine running due to "reversability". I don't entirely understand exactly what that means or if it is a valid explanation of the results.

28 hours is just about 15% less than 33 hours, but still seems significant to me.

Is this result what should be expected?

If possible I would like to see others perform the experiment to see if they get similar results.

I plan on running additional experiments also, but only using single ice cubes, as this took over 60 hours using full cups of solid ice. It took the ice much longer to melt than I had anticipated for both instances.

For clarification, for the non-running phase of the test, a fresh cup of ice from the freezer was used. Very occasionally throughout the test the engine was given a spin manually to see if it would start or not, as an indication as to wether the ice had melted yet or not.

If the engine started and ran It would be stopped manually and left for a few more hours.

The second video is a recording made after it was found that the engine would no longer start. Many attempts were made to start the engine at that time. When it became apparent that the engine could no longer be started, it was removed so as to check the actual condition; had the ice melted or not. That is why the engine is seen turning in the second non-running video. Trying to start the engine by turning it manually from time to time was the method used to determine if the ice had melted.

Finally, when an effort to start the engine failed repeatedly, a visual inspection was made to confirm the diagnosis.

Tom Booth said:
Someone else on another physics forum recently predicted that the ice would melt faster without the engine running due to "reversability". I don't entirely understand exactly what that means or if it is a valid explanation of the results.
If the engine is truly reversible with 100% efficiency, it follows that a stopped engine transfers zero heat.

Of course, that says precisely nothing about the real world.

russ_watters
jbriggs444 said:
If the engine is truly reversible with 100% efficiency, it follows that a stopped engine transfers zero heat.

Of course, that says precisely nothing about the real world.

Even a stopped engine transfers heat, by conduction. The thing that begs an explanation, is: this experiment indicates that a running heat engine transfers less heat than an idle heat engine.

Tom Booth said:
Even a stopped engine transfers heat, by conduction. The thing that begs an explanation, is: this experiment indicates that a running heat engine transfers less heat than an idle heat engine.
A 100% efficient engine does not conduct heat when it is stopped.

Tom Booth said:
The thing that begs an explanation, is: this experiment indicates that a running heat engine transfers less heat than an idle heat engine.
That is a good experiment, but I would have to ask.
Do you see in this experiment where the times for each trial could be compromised?
Such as lifting the heat engine and allowing hot air access to the ice, for example.
Also, does the engine run only on ice, or it could be that it will also run on cold water for some time after the ice has melted.

~30 hours is a really long time for a cup of ice to stay frozen, even with really good insulation. It tells me the engine is taking very little energy from the ice.

From what I've seen of Stirling engine models, they are just demonstrators of the concept without any real energy production. So to better test the issue you should run the experiment with a load applied to the engine.

jbriggs444 said:
A 100% efficient engine does not conduct heat when it is stopped.

Sure, but as you said, then there is the real world. It's less than very unlikely this toy engine is anywhere near 100% efficient. Being made mostly of aluminum, it certainly conducts heat when idle, if there is a temperature difference.

256bits said:
That is a good experiment, but I would have to ask.
Do you see in this experiment where the times for each trial could be compromised?
Such as lifting the heat engine and allowing hot air access to the ice, for example.
Also, does the engine run only on ice, or it could be that it will also run on cold water for some time after the ice has melted.

Yes, there are a gazillion and one ways the experiment could be flawed. Drafts, ambient temperature, activity in the room, the temperature of the freezer the ice came from. (Ice could be colder than 32 F.) Was the refrigerator door kept closed before ice was removed? etc.

And yes, the engine can run for some time on just cold water. At current ambient temperatures (85F or so) with this setup, that seems to be limited to no more than a few hours.

It is a pretty shoddy set up, to be sure. Loads of room for improvement, additional controls, eliminate possible hidden variables, rule out cheating (which describes many of the videos posted to YouTube).

I need to buy some additional equipment, like a simple thermometer would help.

I also don't know with any certainty or exactitude how long the engine will run on just Ice water alone. But I do know it will, for about an hour as I checked the ice at 31 hours because the engine started running very slow. there was still a little ice left. The engine ran for two more hours. At 33 hours it was obvious the ice had melted much earlier, as the melt water was not very cold.

Last edited:
russ_watters said:
~30 hours is a really long time for a cup of ice to stay frozen, even with really good insulation. It tells me the engine is taking very little energy from the ice.

From what I've seen of Stirling engine models, they are just demonstrators of the concept without any real energy production. So to better test the issue you should run the experiment with a load applied to the engine.

I completely agree, and intend to do so as soon as I'm able to rig something up.

BTW, in case I didn't mention it, the ice was frozen by filling one of those new vacuum insulated mugs (I've not seen them till recently anyway) with water and putting the whole thing in the freezer. It took a long time to freeze ice in a vacuum flask. They only cost about \$5. I would have used a Dewar if I could afford it.

33 hours is too long though. Right now I'm repeating the experiment with just ice water with 1 ice cube for quicker results.

BTW, what do you think the effect of adding a load will be?

Tom Booth said:
BTW, in case I didn't mention it, the ice was frozen by filling one of those new vacuum insulated mugs (I've not seen them till recently anyway) with water and putting the whole thing in the freezer. It took a long time to freeze ice in a vacuum flask.
Why did you freeze the water in an insulated container?

You should be aware of the Mpemba effect.
https://en.wikipedia.org/wiki/Mpemba_effect

Tom Booth said:
It is a pretty shoddy set up
Well, you have to start somewhere to see where you are at, so refining the experiment and pin pointing areas of improvement come with the territory. I think it has a lot of potential for the inquisitive mind,

The thermometer seems like a good idea to record a log of change in temperature of the ice, or an ice bath which might be a better choice since that way the cold temperature would fairly constant at 0 C, until all the ice melts. Also, that way you could just make ice (cubes ) and drop them in the flask with some water and stir to get an even temperature.

If you want a colder temperature you could use a salt bath, like that for making ice cream
You would have to investigate that further, or just try it out to see how the engine runs.

Your little engine use very little energy to run.
Enthalpy of fusion for water ( ice ) is 335.55 kJ/kg .
While the heat capacity of ice is 2.04 kJ/kg/ K - around 150 times less energy needed for ice raising the temperature as per melting. For water liquid, the heat capacity is 4.184kJ/kg/K, 80 times less, so your correct, the engine shouldn't run too too long after the ice has melted.

Neglecting the raise in temperature of the ice,
30 hours = 108000 seconds.
Assuming a kg of ice ( looks less than that but ), and the enthalpy of fusion
1 kg of ice x 335.55 kJ/kg / 108000 seconds = 3.1 W of power

Baluncore said:
Why did you freeze the water in an insulated container?

You should be aware of the Mpemba effect.
https://en.wikipedia.org/wiki/Mpemba_effect

Interesting, though I don't see how that could be any influence. The cups of ice were both drawn from the same tap (from a tank connected to a reverse osmosis filter) at the same time and frozen in the same freezer at the same temperature. Both flasks were in the freezer together for weeks, while I was waiting for the engines to get here.

I used vacuum insulated cups to exclude infiltration of ambient heat into the ice as much as possible, so that as far as possible, heat would be forced to go through the engine rather than the heat in the air being able to melt the ice directly.

I also modified the engine by using nylon bolts in place of the steel bolts that came with it. The plastic being much less prone to transfer heat. As far as possible the heat should only be allowed to pass into the air inside the engine, which is doing the actual "work".

Running the experiment again, using just one regular ice cube in ice water, (rather than an entire mug of solid ice so as to shorten the time frame), results were similar.

The ice and cold water stayed cold longer when running the engine: 4 hours and 17 minutes vs about 4 hours.

I was not checking every 5 minutes, so the time for the non-running engine is not exact, but after 4 hours the engine did not respond to efforts to get it started and it was found that at that time all the ice was already melted.

The difference in time frame seems, percentage wise, about the same. That is, in both trials, the ice apparently lasted about 12 to 15% longer when being used to run the engine.

I have some plans for further modifications to the engine to increase it's efficiency so, if there really is a difference, the effect might be more pronounced

At this point the engine has so little power, I'm not sure if it could operate under any kind of actual load, but that is something I would like to compare also: is there a difference in how long the ice will last while being used to run an engine, with a load vs. no load.

An interesting observation during the experiments was that the engine running on ice tends to condense a lot of water. The not running, controls, not much if any at all.

Another phenomenon:

I had the non operating "control" covered by a 1/4 once of styrofoam for some time, which I did also with both runs. But after about 10 hours, the non-running engine would not start. Thinking the ice had already melted, I opened up the setup, only to find that the ice had hardly melted at all.

Without running the engine became so uniformly cold under the insulation that it completely lost its temperature differential.

So, I left the styrofoam off for a while to allow some heat in, to make sure the engine wasn't actually broken. Maybe the displacer fell off.

There was a little condensation around the cylinder when I took off the styrofoam cover.

A little while later I went to try to start the engine but the piston and cylinder had apparently frozen.

I assume this must have been due to cooling caused by the condensed water on the cylinder evaporating when the insulation was removed, if that makes any sense.

Before the piston froze up in the cylinder, it had been 10 hours the control engine sat idle, but just a few hours with the extra insulation.

As can be seen, the engine was functional prior to the freeze up, but would not start for anything, though there was still plenty of ice and it was now exposed to 85 F air on top. I assumed it had just become thoroughly and uniformly cold.

After making this video, I let the engine sit to warm up for an hour or possibly two (I didn't record how long), without extra insulation on the top plate.

As I went to try and see if the engine had warmed up enough to run, that was when I discovered that it had instead, apparently frozen from evaporative cooling, though I'm open to considering any alternative explanation.

I thought that possibly my vigorous but failed efforts to start the engine may have produced some cooling, (mechanically driving a Stirling engine can cause it to behave as a cooler), but I don't think that is likely. Not by itself anyway, as the engine had at least an hour exposed to 85F ambient heat after that before I attempted to start it again. I did however, spend about the same amount of time trying to start the engine before making the video, but I had fully expected that the video was going to unveil the fact that the engine could not run because all the ice had already melted in the control engine after just 10 hours

Imagine my surprise to find the ice had hardly melted!

This brief video shows the aluminized styrofoam, which was kept on the engines while running (or not running) off and on, about half the time.

The running engine, on ice, did not seem influenced by this extra insulation much. It continued to run the same with it on or off, but sometimes slowed down very little it it were left on for several hours.

The intent of the experiment initially, was just to see how long the engine would run on ice.

Anyway, I'm showing this so it is understood what I mean by the styrofoam insulation on top of the engine.

The disk of insulation was originally made to see if a Stirling engine really needed a "sink" for heat to flow through to.

This engine is running on a vacuum flask (mug from Walmart) 1/2 filled with hot water from the tea kettle.

Insulating the top (sink) had no noticeable effect. The engine continued to run at the same speed as far as I could tell.

Last edited:
One thing that seemed to indicate these engines could run without much, if any real "sink" is some solar heated models have acrylic top and bottom plates.

Acrylic is not a good conductor of heat at all. So where could the heat be going?

https://en.m.wikipedia.org/wiki/List_of_thermal_conductivities#Sortable_list

If heat is converted to motion by the engine, is it really necessary for the heat to be able to pass THROUGH the engine to a "sink"? Does the engine, in the contrary, prevent heat from passing through? Is that why the ice lasts longer when used to run such an engine?

Tom Booth said:
BTW, what do you think the effect of adding a load will be?
? That's an odd question. A "load" is extracted energy. The point is to extract energy from the system faster, to melt the ice faster.

Tom Booth said:
One thing that seemed to indicate these engines could run without much, if any real "sink" is some solar heated models have acrylic top and bottom plates.

Acrylic is not a good conductor of heat at all. So where could the heat be going?

If heat is converted to motion by the engine, is it really necessary for the heat to be able to pass THROUGH the engine to a "sink"? Does the engine, in the contrary, prevent heat from passing through? Is that why the ice lasts longer when used to run such an engine?
It's made of acrylic so you can see through it. It's not intended to actually be an efficient engine, it's supposed to simply be a demonstration.

russ_watters said:
That would not apply here --- it isn't clear to me why you think it might.
You quoted without the context.
It was with regard to the time taken to freeze water in the insulated containers.
Obviously it is not relevant to the running of the engine, or to the melting of the ice.

russ_watters said:
? That's an odd question. A "load" is extracted energy. The point is to extract energy from the system faster, to melt the ice faster.

For gasses, though, energy = temperature. More energy = higher temperature.
Extract energy from the system = less energy in the system = lower temperature.

A lower temperature in the system due to faster energy extraction from the system should (by my own not necessarily intuitive or sensible logic), result in the ice remaining cold longer.

Running the engine on ice, the engine is powered by the incoming ambient heat. Correct? So if energy is being extracted from that incoming heat faster, less incoming heat is left to enter the system for melting ice.

If apples are taken out of the basket faster the result is fewer apples in the basket.

If apples = heat, and basket = system, personally my assumption is that with an additional load the ice would last longer.

It makes no sense really, and I'm probably wrong, but the idea has kept nagging at me, which is why I started this line of experimentation. To settle the question one way or the other.

BTW, in a previous post, the autocorrect on my phone turned 1/4 inch of insulation into 1/4 ounce of insulation. I didn't notice until now and it is too late to edit. The insulating disk used on the top of the engine is common 1/4 inch foil faced "house wrap" left over from a construction job.

Baluncore said:
You quoted without the context.
It was with regard to the time taken to freeze water in the insulated containers.
Obviously it is not relevant to the running of the engine, or to the melting of the ice.

I'm sorry for the confusion, I thought it could have an influence. For example, if I froze warm room temperature water for phase 1 but bottled water from the fridge for phase 2 and kept them in the freezer the same amount of time, thinking the colder water would freeze more deeply to give some advantage or something, the result could be the opposite.

I appreciate the information, it is, or could be a potential variable under some circumstances. Something to be aware of anyway.

russ_watters said:
It's made of acrylic so you can see through it. It's not intended to actually be an efficient engine, it's supposed to simply be a demonstration.

True, but to demonstrate anything, the engine has to at least be able to run.

The engines with an acrylic "sink", from what I've seen, run just as well as those with an aluminum sink.

Thermal conductivity for aluminum = 237

Acrylic = 0.2 basically a thermal insulator.

But, for such small engines using an infinitesimal amount of power, it doesn't make a difference? Could be I suppose.

So styrofoam is 0.04 and that still makes no difference?

I thought it warranted some investigation.

Tom Booth said:
True, but to demonstrate anything, the engine has to at least be able to run.

The engines with an acrylic "sink", from what I've seen, run just as well as those with an aluminum sink.

Thermal conductivity for aluminum = 237

Acrylic = 0.2 basically a thermal insulator.

But, for such small engines using an infinitesimal amount of power, it doesn't make a difference? Could be I suppose.

So styrofoam is 0.04 and that still makes no difference?

I thought it warranted some investigation.
I think you are missing @russ_watters's point; the acrylic on the bottom is not the sink any more than the acrylic on the top is the source.

The atmosphere is the sink (and solar energy is the source). Heat energy is transferred to the atmosphere by radiation from the pad on the bottom of the engine. Because the engine is not doing any work the amount of energy that needs to be dissipated is tiny.

This is also the clue to the answers to your experimental 'anomaly'. The only work these engines do is in overcoming the friction in their mechanisms which it tiny therefore the amount of heat energy that needs to be transferred from source to sink is tiny. If you take a large lump of ice and put it in a warm environment then it is going to melt. If you insulate it it is going to melt more slowly, but how much more slowly is going to be governed by a whole load of factors - the initial temperature of the ice, its chemical composition (including dissolved gases), its physical structure, the effectiveness of the insulation, the ambient temperature and airflow, what you stand the container on etc.

I think it is likely that that the variability in these factors is more significant than the introduction of a small transfer of heat from a zero load Stirling engine.

russ_watters
pbuk said:
I think you are missing @russ_watters's point;

I don't

the acrylic on the bottom is not the sink any more than the acrylic on the top is the source.
Nevertheless, it is the path through which the heat must travel to get to the sink. Like a gate. If the gate is locked, there is no access to the garden.

The atmosphere is the sink (and solar energy is the source). Heat energy is transferred to the atmosphere by radiation from the pad on the bottom of the engine. Because the engine is not doing any work the amount of energy that needs to be dissipated is tiny.

This is also the clue to the answers to your experimental 'anomaly'. The only work these engines do is in overcoming the friction in their mechanisms which it tiny therefore the amount of heat energy that needs to be transferred from source to sink is tiny.

That is the Crux of the issue: the misconception that for some reason heat energy must, by some law of the universe, travel from source to sink in order to accomplish work.

If, in a heat engine, energy is converted into work, then that same heat cannot also radiate, conduct or otherwise transfer to the sink it is already accounted for as "work".

If the transfer of heat into the engine is limited, as it is, by the surface area, heat conductivity, ambient temperature etc. Then more work output could mean less necessity for the removal of excess, left over, unconverted waste heat.

If you take a large lump of ice and put it in a warm environment then it is going to melt. If you insulate it it is going to melt more slowly, but how much more slowly is going to be governed by a whole load of factors - the initial temperature of the ice, its chemical composition (including dissolved gases), its physical structure, the effectiveness of the insulation, the ambient temperature and airflow, what you stand the container on etc.

I think it is likely that that the variability in these factors is more significant than the introduction of a small transfer of heat from a zero load Stirling engine.

"Likely", perhaps. But that is the point of an experiment, to eliminate or reduce all of these kind of variables to an absolute minimum as far as possible, so that the effect of the variable under investigation has some opportunity to present itself.

But I absolutely agree, an additional load on the engine, above and beyond overcoming it's own friction, should help to sort out what is actually happening.

Is heat flowing through the engine? As an analogy, like water through a turbine, or water over a water wheel, or is heat simply entering the engine and being converted: one form of kinetic energy (heat) into another form of kinetic energy (mechanical motion and/or additional external "work")?

Personally I'm not satisfied with the current quagmire of assuming it is some combination of these two competing philosophical notions regarding the nature of heat and how a heat engine actually operates.

Because the engine is not doing any work the amount of energy that needs to be dissipated is tiny.

The truth is just the opposite IMO. Because the engine is not doing much work, the amount of energy that needs to be disippated is greater.

If the engine were doing some actual external work, there would be less waste heat requiring disippation to the sink.

Last edited:
My reason for posting to this forum is primarily to find out if it is already well known, or considered "normal" or accepted that by converting heat into work, OF COURSE ice being used to run a heat engine will take longer to melt, EVERYBODY knows THAT!

Personally I have no idea if that is considered established knowledge . To some degree it seems to be, but on the other hand, the opposite is often promulgated.

The science of what makes a heat engine tick does not seem to be well established. There is no real consensus. It has not been reduced to formula to the extent that something like electricity has for electric motors.

Is there a formula for heat flow resistance through a conductor like Ohm's law?

Amps X Volts = Watts that sort of thing.

I just want to know if this is a "normal" result or not.

Partly because why wast time on it if it is? I have enough to do.

But, I get the impression that this is considered "anomalous". Or something. In that case I need to figure out if there is, or how many flaws there are in my procedure. Or on the other hand, the admittedly less likely scenario that I'm actually on to something.

Tom Booth said:
If, in a heat engine, energy is converted into work, then that same heat cannot also radiate, conduct or otherwise transfer to the sink it is already accounted for as "work".
If you are not tapping that energy away, out of the system as useful work, then surely it will wind up in the cold reservoir?

jbriggs444 said:
If you are not tapping that energy away, out of the system as useful work, then surely it will wind up in the cold reservoir?

That is the prevailing wisdom, is it not:

So logically, without a load (work = 0), all the heat entering the engine ends up at the sink. In this case the ice. The ice melts quickly, I would guess, about the same as if exposed directly to the heat source.

Add a load, (so work > 0) less than 100% of the heat entering the engine ends up at the sink. In this case the ice. (which is otherwise well insulated against heat infiltration.) So with a load the ice should not melt as quickly as with no load

Of course even at no external load, there is friction dissipating heat from the bearings to the air, heat transferred to the air by the flywheel, noise, vibration, etc. Which in this case are beneficial rather than just being inefficiencies, act as a partial load and divert heat away from the ice. So, any load, even the engines own friction results in a slowing down of the melting of the ice.

This all seems like basic thermo to me, and rather obvious but more often than not, I'll end up in a debate about it with someone

Tom Booth said:
Of course even at no external load, there is friction dissipating heat from the bearings to the air, heat transferred to the air by the flywheel, noise, vibration, etc.
And where does this energy go? Is there a third reservoir?

jbriggs444 said:
And where does this energy go? Is there a third reservoir?

It basically remains in or is transferred back to the source, but in a different form. Noise, vibration, motion of the air, are no longer counted as heat but "work" in one shape or form or another.

When a heat engine performs work otherwise, like lifting a weight or driving an electric dynamo, where does the heat go? Do we need another reservoir to account for the "lost" heat?

I don't think so.

Last edited:
Tom Booth said:
It basically remains in or is transferred back to the source, but in a different form. Noise, vibration, motion of the air, are no longer counted as heat but "work" in one shape or form or another.
Why to the source? Why not to the sink? And what makes you think that it counts as "work"? Neither source nor sink as heat reservoirs have any capacity to individually sink "work".

pbuk
jbriggs444 said:
Why to the source? Why not to the sink? And what makes you think that it counts as "work"? Neither source nor sink as heat reservoirs have any capacity to sink "work".

If I hear squeaking from the flywheel at least that portion of the energy is going to my ears, true, some small fraction of sound may reach the sink, under all that insulation.

The air being pushed around the room by the spokes of the flywheel, the piston displacing air, transferring energy to the atmosphere, these are all heat transformed into other forms of energy that don't reach the sink as heat.

They may not be "useful work" but I don't think that the expanding gas in the engine pushing the piston to make all that happen makes any distinction between "useful" or otherwise.

Merlin3189

• Thermodynamics
Replies
6
Views
2K
• Thermodynamics
Replies
3
Views
673
• Thermodynamics
Replies
1
Views
668
• Thermodynamics
Replies
3
Views
2K
• Thermodynamics
Replies
14
Views
1K
• Thermodynamics
Replies
3
Views
11K
• Thermodynamics
Replies
16
Views
12K
• General Engineering
Replies
1
Views
1K
• Thermodynamics
Replies
8
Views
8K
• Mechanics
Replies
6
Views
1K