Thread Tools
Nov 02, 2012, 05:38 AM
Blessed by anvils from heaven
Chophop's Avatar
You might want to search for electronic heat sink paste that can take the heat range. I have home construction going on, if I remember I'll find time.
Sign up now
to remove ads between posts
Nov 02, 2012, 09:55 AM
Registered User
Well, for now, the flow seems to be the problem, as well as the radiator.
The coil should be able to take the heat provided sufficient flow, but I am not getting that flow I need...

Most likely too long, too narrow lines

At idle it is cooling though, holds a steady temperature and a steady flow, so the concept is sound, but when I fly, the liquid starts boiling in the coil.
It might be possible, that I will get slighty better results with a different location of the radiator, as the return temperature is still a bit high (50 degrees at idle

It is not a succes (yet) but also not a total faillure, because I did a few test hoovers of approx 30 seconds, and I could hear the engine go a little hotter and leaner but it kept going, and took a normal idle immediately after touchdown (usually when the engine is hot it wll only hesitantly go into idle).
And than we're talking about 8 kilo of helicopter, with an engine without fan and with the fins covered by the "cooling coil".

And while I am writing this, it is idling allready for 15 minutes like a happy little engine (unfortunaltely, you cannot fly at idle.... ).
I am convinced I can get this working....

Brgds, Bert
Nov 02, 2012, 11:03 AM
Registered User
Originally Posted by I-FELN View Post
Thanks for the valuable information on this interesting project by the radiator. I hope you provide data on cooling capacity of the system. Do you think that 3 rounds of 5.5 mm pipe is sufficient to subtract enough heat from the cylinder?. My idea: a copper pipe 2 mm with more rounds would not increase the contact surface with the cylinder?
Best regards
Flow is the biggest factor. Water can extract amazing quantities of heat, but it needs flow to prevent boiling. If the tube is smaller, flow will be restricted, and the forces that create the flow are very small.

As it was now, I had 5.5 mm piping with a total length of roughly 60 cm.
I am going to relocate the radiator to directly above the engine, and try short, big pipes (for now, 8 mm) but the head will still have the coil, because I will not be able to fabricate a good cooling head until after the weekend....

Brgds, Bert
Nov 02, 2012, 01:04 PM
I-FELN's Avatar
Ok understood, a smaller diameter tube is an issue to the flow, then I think if the convective circulation is insufficient the solution may be a forced circulation with a pump!
Nov 02, 2012, 01:15 PM
Registered User
I think, it has to do with the coil: the water spends too long time in the coil with the low flow.
Tomorrow I will fabricate a head with a cooling water jacket, and big inlet and outlet connections

That will allow for bigger piping, and most likely, my radiator still was too small (I calculated the surface based on heat-transfer formulaes in an online tech vademecum) but maybe I have misjudged the amount of air movement under the rotor close to the mainshaft.

But I am definitely convinced, this system can work. After all, that Grossi Engine has similar HP rating as my helicopter engine, and does with a lot smaller radiator, and a lot less height difference.

I am reasoning as follows: normally the engine looses all its heat via the heat sink. The heatsink typically has an operating temperature of approx 80 to 100 degrees.
That means, if I fit a waterjacket instead of a heat sink, and I manage to keep the waterjacket at 80 degrees, it will absorb sufficient heat to effectively cool the engine, because the metal to metal contact temperature the engine "sees" is the same temperature....

But if all else fails, a pump will not be too much trouble to make.

Brgds, Bert
Last edited by Brutus1967; Nov 02, 2012 at 01:37 PM.
Nov 02, 2012, 07:00 PM
Registered User
With the boiling in the pipe, might be able to get a little more out of it with a pressurized system. That is if your running it open.
Nov 02, 2012, 09:14 PM
Blessed by anvils from heaven
Chophop's Avatar
You'll probably need to insure even cooling, you might cause squeeze rings or spots on the cylinder where cooler areas contract.
Nov 03, 2012, 03:35 AM
Registered User
Well, I have been thinking about pressurizing the system, but that only raises the coolant temperature. it does not increase cooling capacity.
Boiling is not even that bad of a cooling system, as the heat absorbed to evaporate a given amount of water is approximately equal to the heat needed to heat that same amount of water by 20 degrees. To an extent, it works as a very effective temperature limiter.
But it does have a negative effect on waterflow and it creates 'hotspots" in the coil or jacket, so if for example the jacket would be in direct contact with liner or combustion chamber, boiling is very bad. As it is, cooling is effected by heat conduction through the metal of the combustion chamber, to the metal of the cylinder head, to the metal of either heatsink or cooling coil, so I guess that evens things out a bit.
I remember old fairground attractions driven by stationary engines that had an open bowl of boiling water as only means of cooling, and as long as there was still water in the bowl, the engine would be OK.
I know, at least one F3C pilot here in the netherlands experimented with exactly that principle, having a tank of 100 cc of water fitted to the cylinder head, boiling off during the flight. That amount of water should effectively be able to absorb 1 kW of heat for 4 minutes on evaporation alone, at a guaranteed constant temperature of 100 degrees, which is definitely in the normal operating temperature range of a helicopter engine. I prefer a closed loop and no evaporation however....

I have been looking at the original air cooling arrangement, and it is pretty clear, only the heat sink was in the airduct, the rest of the engine was not encased, and not subject to the forced airflow.
The only part of the engine cooled by the original airflow was the cylinder head+heatsink, and that mainly via the contact surface between heatsink and head.

The cylinder does not even have a finned area to speak of (there are fins, but it looks like they are there for cosmetic reasons, the fins are maybe 1 mm wide and there are 5 of them.

So I guess, the coil has too little water volume to absorb the heat, and too little contact surface to the head.
As all was just an experimental set-up, it is of course still very possible that I have not located the radiator in the optimal part.

But I am out of piping to rearrange, so I want to get out to get some flex hose that has sufficient heat range, like silicone hose or similar so I can experiment a little bit easier

One of the problems I encountered, was to figure out how much flow there actually was, so I have added some particles to the liquid, to make things visible. The flow (before the water starts boiling) is something like 5~7 cm/sec (estimated, very hard to measure accurately, which in my piping arrangement comes down to approx 1 cc/sec.
That is unfortunately only 1/10 of what I think I need, based on a temperature difference of 20 degrees between supply and return temperature.
As far as I know, flow velocity (cm/sec) will stay approximately the same, give or take a bit, regardless of piping diameter, so it should follow that increasing the piping increases the volume flow, but I still have to confirm that.
Increasing the temperature difference will of course also increase flow (a bit) and reduces the required flow, but in order to do so, the radiator size has to increase many times. So I have to try and figure out the optimal balance between radiator size, temperature difference and flow.

I have just boiled out Tygon hose for an hour, and it seems it is not affected by 100 degrees (which should be the maximum system temperature in an atmospheric cooling system) it makes for a lot more easy experimenting with radiator locations.

I am still convinced, it can work, because the system worked perfect (constant flow, constant temperature difference, no surging in the expansion vessel) when the engine was idling, so it is just a matter of increasing cooling capacity.
The power gain from not having a fan is definitely very noticeable: The helicopter always lifted off a little bit above midstick, and during the experiment it lifted off clearly below midstick, with a slightly higher RPM.

Whether the experiment will succeed or not, it is very educational stuff, this...

Brgds, Bert
Last edited by Brutus1967; Nov 03, 2012 at 03:59 AM.
Nov 03, 2012, 05:51 AM
Registered Misuser
Vince77's Avatar
Awesome experiment. Reading through it seems you could increase flow and decrease the potential for the water overheating by splitting the flow among more pipes. Instead of a coil with water getting hotter each turn, a few loops each entering with cool water and leaving after a single pass. This also means the flow goes up one time with each pipe added without the need for a pump.

I helped a friend setup a cpu heatsink and, in the vacuum of any experience, we decided it would be best to branch the supply pipe into four. The four pipes went into the channels on top of the heatsink then merged again. We used 1/8 in brass tubing and (I think) 3/16 in brass tubing. We widened and flattened the larger tubing over some needle nose pliers then fit the four smaller pipes and brazed it all together. One fat fitting went on each side of the heatsink. Now we had a fat pipes on the (pre-made) radiator and four narrow pipes picking up heat. Only potential problem is it would have to be assembled and brazed while on the engine... tough to remove.

Just a thought.

Vince L.
Last edited by Vince77; Nov 03, 2012 at 05:53 AM. Reason: Android on screen keyboard...
Nov 03, 2012, 06:53 AM
Registered User
Thought about splitting the coils etc, but that brings too many connections in play to be suitable for experimenting. besides, ALL tubing must increase diameter, only multiplying the coils will not really have an effect unless the coil was the limiting factor.
In a final design it might be feasible, but for now I am afraid I will only increase "production time".
But I am going to make a full jacketed head anyway, so that problem should be eliminated...

As things go along, I gather more and more data.
From normal operating temperatures as measured over the years, I can conclude that that would be 80~100 degrees C on average.
Taking into account the transfer coefficient for forced-airflow cooling multiplied by the surface of the original heatsink, it is safe to say that the amount of heat I need to remove, is in the area of a continuous 900 W avarage. Peaks most likely will be buffered by the water mass of the system, like the heatsink does normally in conventional air cooled helicopters.
This would be reasonably in the ballpark, as the helicopter needs approx 1200W for hoover, and a cooling requirement of 75% output, is not at all that far off.

So if I take this amount of heat to be removed,depending on the coolant delta-T, I need 10~11 gram/second of waterflow for delta-T = 20 deg C, or 5 gram/sec for delta-T is 40 degrees C

I hope to get 5 gram/sec with the wider piping.
Then, for a radiator cooled in still air (like, in your living room), I would need a surface of approx 0.5 square metre.
With the slightest form of air movement, that needed surface quickly reduces.
I have redesigned my radiator, and this one will be 0.15 square metre, but with pretty wide fin openings, and it will be located aft of the canopy, so it should pick up some airflow.
Fitting a second radiator will be a piece of cake, but unfortunately, my LHS did not have more sheet-Aluminium

The new radiator of 1500 scuare cm, will be even 10 grammes lighter than the previous (600 sq cm), but unfortunately the Tygon hose is pretty heavy at over 100 grammes. Upside is that as soon as the lay-out of the system is definitive, it will be replaced by Aluminium piping, shedding almost 80 grammes.

Watercontent will be roughly 100 gramm.

All in all I am not getting a weight advantage over air cooling anymore (but the weight penalty is minor and still more than compensated by the elimination of powerloss in the fan).

Brgds, Bert
Last edited by Brutus1967; Nov 03, 2012 at 06:59 AM.
Nov 03, 2012, 06:10 PM
Registered User

Some pictures of works in progress

Here a few pictures of the fruits of labour of today....

The cooling jacket I made myself, but in my hobby-buddy his workshop, as my toolbits are not up for the job.

The only thing I needed help with was drilling the 6 holes for the headbolts and drilling the lamellas, as I am not too familiar with a milling machine.

Parts will be glued together with UHUplus Endfest 300, which is a very strong slow setting epoxy suitable for heat-curing, which almost triples the strength.
It has already been tested in the first experiment and proved to be fully resistant to the high temperatures (I joined the two parts of the coil with it and that held perfectly).

The coolig water jacket has in and outlet pipes of 10 mm inner diameter and the volume is 16 cc (more or less 1 cubic inch.

The radiator will have a cooling surface of 1550 sq. cm (240 sq in) and weigh 110 gramm without water.

Total water content will be less than 100 cc, including expansion vessel

Brgds, Bert
Last edited by Brutus1967; Nov 03, 2012 at 08:58 PM. Reason: Some data.....
Nov 04, 2012, 01:54 AM
Registered Misuser
Vince77's Avatar
Impressive once again. The images certainly help since I was thinking of typical nitro type setups with the cylinder buried between the frames. May I ask, what brand heli are you working with?

Also, you may get some more enlightened feedback if you start a thread in the engines forum. While I can read and understand your calculations, I can not validate them. There are some excellent experimenters over there that may have a thought or two to help. But please do let us know if you take your experiment over there.

Vince L.
Nov 04, 2012, 02:09 AM
Registered User
The helicopter is an old Vario X-treme, the engine an even much older Rossi R61 rear-exhaust.

It should be possible to do this on any helicopter, because removal of fan and airduct really tends to free things up a bit

I'll keep this topic here, just because I know heat is the enemy of the people modifying a glow to gas, and although that is not my purpose, for me this is just an experiment to see if I can make it work, they might benefit from it.
My fourstroke helicopter is also very heat-challenged, and a lot less "experiment-friendly".
In a few minutes I will edit in the pics of the finished radiator, and the total set-up.
In about an hour I will be testing....

Brgds, Bert
Last edited by Brutus1967; Nov 04, 2012 at 03:53 AM.
Nov 04, 2012, 03:52 AM
Registered User


I did it!
I have built a functional watercooling system for a helicopter!

I have done some tests of hoovering, gradually increasing timet to over one minute and the engine kept running consistently.

I saw minor vapor forming just after the head, but a good waterflow and a return temperature after the radiator of less than 50 degrees C according to the probe (but I doubt that is an accurate value)

It is still a bit on the limit, and because it is cold at the moment (8 deg C), in summertime, the system maybe is not yet adequate, but hey, it is just a prototype.... There are still improvements (probably fairly large) to be made in optimizing the radiator, radiator location and the piping (for example changing from Tygon tubing to Aluminium alone, increases cooling surface by roughly 20% and save an additional 50 gram...).
And most likely, the system could use a little bit extra cooling surface, just for safety....
At Idle, I can continuously touch the hot branch, meaning (for my numbed fingers) that it will be approximately 60 deg C. Immediately after touchdown, the cold branch is less than 60 degrees, but fairly warm to the touch. I do not know if the temperature measured is accurate as the probe was not isolated.

After the previous system and it running hot, I took the precaution of fitting an (extra?) headshim, as I found there was none. and because the helicopter is now hoovering with a slightly lower throttle setting I fattened up midrange ever so slightly.

And that all before morning coffee!

Here are the last pics, I will try to get a vid as well but that can take some time.

Brgds, Bert
Last edited by Brutus1967; Nov 06, 2012 at 11:16 PM. Reason: Replaced pictures
Nov 04, 2012, 05:35 AM
Registered Misuser
Vince77's Avatar
Very impressive, especially this early in your experiments. My humble suggestion: angle the radiator slightly so the swirling rotor wash is closer to perpendicular/flowing through the radiator. Maybe 10 degrees from vertical could make the radiator more transparent to the breeze from the rotors. This would still keep most of the vertical offset you need.

Looks like your exhaust blows to the left (rotor runs clockwise), so angle the bottom of the radiator maybe 2 cm away from the exhaust pipe.

Also, computer shops may have suitable radiators in their liquid cooling selection. Most have circular routing but I have seen straight ones that will work with a convection setup. They usually list flow rates and the wattage they can handle (with forced flow).

Regardless, with nearly identical weight and no fan losses, it’s an impressive system. Thanks for keeping us in the loop.

Vince L.

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion Ticking noise from tail of Raptor Titan 50, any ideas? Canada Goose Large Electric Helis 3 Aug 13, 2011 05:47 AM
Sold Raptor 30/50 V2 Parts mundo330 Aircraft - Fuel - Helis (FS/W) 2 Jul 25, 2011 09:44 AM
Sold TT Raptor 30/50/E550/620 All metal head doug47 Aircraft - Electric - Helis (FS/W) 0 May 06, 2011 09:26 PM
Wanted Raptor 30 or 50 kit or arf PV P-40 Aircraft - Fuel - Helis (FS/W) 5 Jul 04, 2006 02:25 PM