How To Waterproof A Metal Box?

[Mastermime]

Hey everyone,

I am building a robot that is going to have to withstand things such as heat and water and I need a way to wetherproof a metal box that will hold all of my electronics.  Would rubber seals work?  Any ideas?

[Soeren]

Hi,

I am building a robot that is going to have to withstand things such as heat and water

How much heat and how does it reach the box (radiated, flames, what)?
How much water and how does it reach the box (immersion, rain, what - fresh- or saltwater)?

and I need a way to wetherproof a metal box that will hold all of my electronics.  Would rubber seals work?

The material needed depends on the above and how well you can make the mating surfaces - eg. silicone rubber can be found for pretty high temperatures.

[Mastermime]

How much heat and how does it reach the box (radiated, flames, what)?

It will be flames that would be in something such as a burning house.

How much water and how does it reach the box (immersion, rain, what - fresh- or saltwater)?

It is not intended to be fully submersible.  So yes it should be capable of withstanding rain.  It will always be freshwater.  I don't live near any bodies of salt water.

[Gertlex]

For rain, the easiest approach is going to be just having a roof structure that removes water away from the insides...

For fire... well there's active flames, and then there's pure heat.  My best recommendation is to not surround this box with burnable stuff... e.g. clear the shrubbery around it (how far, I've no idea).  But again, house burning, but exposed to rain, it's not clear what we're after.  From what I've heard, a brick house can combust from the inside if there's enough heat on the outside...  So it then becomes a question of how much heat for how long, which you would need to answer in order to determine how much insulation or whatever to use.

[Mastermime]

Well my idea is to make my robot impervious to high heat temperatures (fire) and water splashed on it every now and then.  It should be able to go through 2 1/2 in puddles and what not.  It should be able be withstand excessive heat (flames) for 90 minutes.

[Soeren]

 Hi,

Well my idea is to make my robot impervious to high heat temperatures (fire) and water splashed on it every now and then. 

Splashing water on a very hot box cause steam and steam penetrates far better than rain from the above.

It should be able to go through 2 1/2 in puddles and what not.  It should be able be withstand excessive heat (flames) for 90 minutes.

2½" puddles isn't really an issue, but direct flames for 90 minutes may be something of a challenge, as your electronics will behave erratically when it reaches somewhere between 80°C and 120°C (if not sooner), but htere's a huge difference between a welding torch directly against the box (burning through in a short time) and gentle flames nearby.

Best would probably be an aluminum box, with aluminum shields mounted 10..20mm out on aluminum rivets/tubes/whatever, to keep some of the heat away from the box itself.

A mist of water and a fan mounted in a sensible way may help as well (as long as it doesn't feed the flames.

CO2 might be used as well, but there's an inherent risk in bringing pressurized containers into a fire

What kind of wheels are you planning for going into a fire?


Is this for some school competition (or similar), or is it something you are planning for serious use?

[Mastermime]

2½" puddles isn't really an issue, but direct flames for 90 minutes may be something of a challenge, as your electronics will behave erratically when it reaches somewhere between 80°C and 120°C (if not sooner), but htere's a huge difference between a welding torch directly against the box (burning through in a short time) and gentle flames nearby.

Best would probably be an aluminum box, with aluminum shields mounted 10..20mm out on aluminum rivets/tubes/whatever, to keep some of the heat away from the box itself.

A mist of water and a fan mounted in a sensible way may help as well (as long as it doesn't feed the flames.

These are some great ideas!  Thanks!

What kind of wheels are you planning for going into a fire?

I am actually using treads that will be slightly angled upward (somewhere around 120 degrees, haven't got that far yet) in the front for climbing over things.

Is this for some school competition (or similar), or is it something you are planning for serious use?

This is serious use

[Soeren]

Hi,

This is serious use

Then perhaps it would be viable to make a hollow wall box that could be packed with dry ice - that would both cool and quench flames (but it can be hard o find an instant supply of dry ice).
The electronics shouldn't get too cold either, but I don't think that would happen in a burning building.

Assuming it would be for remote monitoring, you should factor in extra time to get a cam to work reliably in the heat - talk to companies that specialize in that. Perhaps they can tell you exactly what optics will work best and whatever else you need to address.

A good place to start would be a simple ROV made as cheap as possible, to just monitor the range of temperature vs. time that are realistic - no problem if it crash and burn doing that, as long as it has sent the data to you up to that time.
If you have that kind of data, it will be much easier to get the proper one to last, without going overboard, as you'll then know rather than guess.

[Mastermime]

Then perhaps it would be viable to make a hollow wall box that could be packed with dry ice - that would both cool and quench flames (but it can be hard o find an instant supply of dry ice).
The electronics shouldn't get too cold either, but I don't think that would happen in a burning building.

I'll definitely look into that.  Is it possible to use a large cooling fan to direct the heat away from the electronics just like you said above?

Assuming it would be for remote monitoring, you should factor in extra time to get a cam to work reliably in the heat - talk to companies that specialize in that. Perhaps they can tell you exactly what optics will work best and whatever else you need to address.

Ok I just contacted the company to see the maximum temperatures the camera can run at.  I'll post once I get a reply.  I am using a weather proof security camera with a wireless transmitter and receiver so we'll see what they say.

[Soeren]

Hi,

Is it possible to use a large cooling fan to direct the heat away from the electronics just like you said above?

In itself, no.
A fan doesn't cool a thing, it just moves air and if the air is just as hot at the intake, there
s no change in temperature (hence the added water mist - evaporation takes heat and will have to take it from somewhere - remember how cold your forehead can become when you're sweating like a pig running a marathon?)
If you use dry ice outside the box (in the mentioned "double wall"), all power components can use the sides (or bottom) of the box for heat sinking and the entire inside should be cold enough without any fan, especially if the top side is "iced" too.

Ok I just contacted the company to see the maximum temperatures the camera can run at.  I'll post once I get a reply.  I am using a weather proof security camera with a wireless transmitter and receiver so we'll see what they say.

You could make a larger tube to stuff it into and pack it with dry ice as well, but you have to keep an eye out for condensation issues (perhaps add some siccative inside the inner box).

[Mastermime]

Ok the reason why I am steering away from dry ice is because I don't want to constantly be buying dry ice.  However, if there aren't any other solutions I will definitely be going with your great idea.  I am more interested in the mist idea right now though.  Would it be similar to how a humidifier would work for your home? 

Edit: Even though the costs may be a bit more expensive, I think the solution involving dry ice will work best.  I'm not sure how much dry ice I will need though.  If I remember correctly doesn't dry ice sublimate 1 pound every 24 hours, but for my case this time will probably be sped up.  Thank you so much!  Would a reusable pack like this work? http://www.cameronpackaging.com/Dry_Gel_Ice_Sheets.html#cold

[Soeren]

Hi,

Ok the reason why I am steering away from dry ice is because I don't want to constantly be buying dry ice. 

I hear you!

I am more interested in the mist idea right now though.  Would it be similar to how a humidifier would work for your home? 

I'm not really familiar with humidifiers - here (DK) we do our best to keep humidity down most of the year, but I assume it's some kind of fine mist sprinklers, perhaps directed towards a heating element(?)

If you make/buy a very fine nozzle and have a means of getting water to it under pressure (experiment) to create "dew" on a surface, it should be about right. The fan will then direct hot air towards the dew, which will evaporate, sucking heat out of the surface - how much, how fast and all such questions takes experiments to answer for your exact construction.

Edit: Even though the costs may be a bit more expensive, I think the solution involving dry ice will work best.  I'm not sure how much dry ice I will need though. 

Neither am I.
I think the time is ripe for you to try out some different ideas in situ. One can speculate, but there's no such thing as hands on experience 

If I remember correctly doesn't dry ice sublimate 1 pound every 24 hours, but for my case this time will probably be sped up. 

I'm sure you could calculate some approximations, but there are so many variables and you don't even know what temperatures you'll see yet, so I'd personally just make a test 'bot with telemetry temperature monitors and send it into a fire to get data realistic to your setup - perhaps start with a small fire 

Would a reusable pack like this work? http://www.cameronpackaging.com/Dry_Gel_Ice_Sheets.html#cold

Hey, good find - although I cannot see an immediate use for it, I have saved the link for later.
How it will behave in a fire... I honestly have no idea, but it cannot be colder than your freezer and even with the freezers that goes down to -100°C, I wonder if the chemicals will actually play well. I'm doubtful as to their use in your project - at least it won't quench flames as the dry ice will, but again... Experiment.

If the 'bot is for pro use, only the best solutions should do (and I don't think there's many non-pro applications for what you make), as there may be lives and at least property at stake - but it's time to gather a little of each and make physical experiments.


As a side note, the belt drive should be all metal and probably of the same kind (say steel) or different rates of thermal expansion may lock up the links (I assume they'd have to go over glowing ember now and then).

[georgeecollins]

There is really good info on this site about waterproofing servos:
http://www.societyofrobots.com/actuators_waterproof_servo.shtml

This shows how you can use two material for waterproofing: liquid plastic and marine grease.  The liquid plastic would be bad for heat, but good for waterproofing. 

I like the box within a box idea, but in that case I don't know why you would want to have an outer box of aluminum, which conducts heat well.  Why not waterproof an internal metal box and then have the outer box be something that is more insulating, like lexan or carbon fiber.   Lexan can melt, but it takes an awful lot of heat to do that.  It is light, easy to work with, and insulating. 

Good luck. 

[Mastermime]

To Soeren:

First off, thank you for all the input you have given me.  I'll just experiment to see what's best.  I'll purchase one of those reusable packets of dry ice and see how it behaves near fire.  If that doesn't work I'll use plain dry ice. 

To Georgecollins:

Ok I'll take a look at that tutorial.  So are you saying I should put lexan or carbon fiber on the outershell of my robot?  That would be a good idea except that this robot has to deal with a lot of abrasion and it has to be really strong because it will have to penetrate and blow through obstacles.  I was thinking I'd use aluminum for the inner box and then steel for the outside box.  I wouldn't need steel for the inner box.  It would just be extra weight.

[Soeren]

Hi,

There is really good info on this site about waterproofing servos:

We already established that nothing will be immersed and just needs to be "splash proof".
Besides, common R/C sercos will be much too weak for locomotion. The electronics should be kept dry though, but there's no need for radical methods like filing it with grease, which will make more damage than good.
All kinds of plastic except a few internal silicone rubber seals should be avoided and all kinds of oil, grease and the like should be kept away. Only dry lubricants will survive in high heat.

I like the box within a box idea, but in that case I don't know why you would want to have an outer box of aluminum, which conducts heat well.  Why not waterproof an internal metal box and then have the outer box be something that is more insulating, like lexan or carbon fiber.   Lexan can melt, but it takes an awful lot of heat to do that.  It is light, easy to work with, and insulating. 

It's not as much a box in a box as a bow with heat shields mounted (like the ceramic tiles on the space shuttles).
There are several reasons for aluminum, like weight and ease of engineering and being at it for 90 minutes each run, using aluminum or stainless steel (which is a poor conductor of heat) won't do much difference - hey, even 2 minutes and there's no difference.
You cannot isolate passively against 90 minutes of heat, unless you make a 200mm shell of mineral wool (or something with a similar K-value and heat resistance) around the box at all sides (with no heat bridges, so the axles and such would need to be isolated in some way as well).

Lexan (polycarbonate) will melt and curl up in a fire. The max. permitted temperature, although high for a plastic, is "only" ~250° (that's why it's so easy to heat bend with a hot air gun) and a fire can reach pretty high temps. If subjected to dry ice,it will be brittle and crack if knocked or vibrated.

Carbon fiber, if heated to a point, looses the resin that holds the fibers together and the fibers without the resin can hold nothing.
I tried to heat bend a carbon fiber rod once... Emphasize on once.

It has to be a mostly all metal construction, perhaps with some ceramics thrown in for street cred (and to get rid of any excess money )


To sum up:
Outer materials have to be able to withstand up to 90 minutes in a fire (which may mean more than 1000°C).
Inside the box, temperatures should be kept under 100°C, preferably much lower (like 50° max).
Motors and wires should be kept low as well, isolation on wires could be Kynar or something similar to help withstand high temperatures, but the magnets in the motors will demagnetize if they get to somewhere between 100°C and 180°C  (depending on magnet material) special high temp motors can go to above 400°C, but here's what one manufacturer says on the subject (my emphasis)

Q - What is the maximum temperature at which standard motors can be operated?

A - The life of the electrical insulation and of the lubricants are adversely affected by high temperatures. A generally accepted "rule of thumb" is that for every 10°C increase in operating temperature, life is halved. NEMA standards normally limit the maximum ambient temperature to 40°C. A probable life of approximately five years under normal operating conditions 8 hours/day, 5 days/week, (ambient temperature below 40°C, and uniform rated load) can usually be expected. Due to the many variables, life can only be estimated before testing under actual operating conditions.

Then there's the lubricants - I can only think of either graphite or molybdenum disulfide (both go to ~400..450°C) or boron nitride (1200°C) as likely candidates, as the more esoteric types are expensive and hard to find.

The more I think about it, the firmer my beliefs in dry ice gets.

Designing something that keeps its joy in a fire does take careful selection of materials, wicked cooling techniques and a head full of hair (for regular pulling) - all in all a worthy challenge 

[Mastermime]

Lexan (polycarbonate) will melt and curl up in a fire. The max. permitted temperature, although high for a plastic, is "only" ~250° (that's why it's so easy to heat bend with a hot air gun) and a fire can reach pretty high temps. If subjected to dry ice,it will be brittle and crack if knocked or vibrated.

Carbon fiber, if heated to a point, looses the resin that holds the fibers together and the fibers without the resin can hold nothing.
I tried to heat bend a carbon fiber rod once... Emphasize on once.

It has to be a mostly all metal construction, perhaps with some ceramics thrown in for street cred (and to get rid of any excess money )

Yes I agree with Soeren.  Metals are the way to go on this project; because of both tensile strength and melting point.

Outer materials have to be able to withstand up to 90 minutes in a fire (which may mean more than 1000°C).
Inside the box, temperatures should be kept under 100°C, preferably much lower (like 50° max).
Motors and wires should be kept low as well, isolation on wires could be Kynar or something similar to help withstand high temperatures, but the magnets in the motors will demagnetize if they get to somewhere between 100°C and 180°C  (depending on magnet material) special high temp motors can go to above 400°C, but here's what one manufacturer says on the subject (my emphasis)

Yes! Kynar is the exact type of thing I was thinking of for protecting the wires.  For the motors I am planning to have those inside the cooled area of my robot with the output shafts protruding out of the box.

The more I think about it, the firmer my beliefs in dry ice gets.

Designing something that keeps its joy in a fire does take careful selection of materials, wicked cooling techniques and a head full of hair (for regular pulling) - all in all a worthy challenge 

You make me laugh .  I'll try and post it once I am done.  Thank you for the great feedback and support

[Keithh28]

I might be too late for this suggestion, but have you thought of using a small fire-proof safe which is rated for digital media?  Some even have an integrated USB port that might be useful.  Obviously no option is going to be fire-proof foreever, but insulation in the fire-proof safe may buy you enough time to get the job done, and no dry ice needed.  Obviously, you 'bot would need to be of a decent size to cart one around, but I've seen some that are quite small and only weight a few pounds.

Keith

[Billy]

...have you thought of using a small fire-proof safe...

Definitely studying fireproof safes is where he should start. I think he underestimates the heat in a house fire.
I have a fire proof safe at home that is packed with about 2 inches of what I guess to be Gypsum on all sides. It is rated for only 60 minutes. Gypsum is used in these applications because as it heats to 100C, water that is chemically bonded to the minerals is driven off, acting as a coolant.

If I had to design a robot to spend 90 minutes in a house fire, I would use active water cooling if I was allowed to tether it to a water source outside the fire. If not able to tether it with coolant lines, I would use water jackets that could boil off for cooling and wrap the whole robot in refractory blankets to keep the heat off.

[Tommy]

with the heat shield idea and some type of AC unit could you forgo the dry ice?

http://www.ebay.com/itm/251025142838?hlp=false

Tommy

[Mr. Ninja]

Or, looking at it unconventionally, you could use refridgeration to cool the inside. You would have to make it efficient enough, however, to be able to keep the inside under the maximum temperature (way under). You could then surround it with a layer of insulating material.

Just me thinking. Please bear in mind that I don't have much experience with heat transfer and thermal conductivity.

[Billy]

The OP is gone now so this thread is simply for discussion and learning now.

AC units (or refrigeration, the same thing) move heat from one side of the compressor to the other, but only provided the temperatures on each side of the compressor are reasonably close to each other. So your freezer at home can keep the inside of the box at -15C when out side the box is at 20 - 40 C. So it can support a ~55C difference between inside and outside the box.

In a fire, expect the room to be closer 500 - 600C. Even if we ignore that the refrigerant won't go through a phase change, you'd still have ~450C in the box.

http://en.wikipedia.org/wiki/Refrigerant

[Mr. Ninja]

Hmm, that was something I thought could go wrong, thanks you for telling me what was wrong in a civilized way (unlike some other forums, sheesh).