Author Topic: MCT Speculation and Discussion Thread 4  (Read 570899 times)

Offline lamontagne

  • Full Member
  • ****
  • Posts: 1082
  • Liked: 1541
  • Likes Given: 241
Re: MCT Speculation and Discussion Thread 4
« Reply #300 on: 06/29/2015 03:22 PM »
With the size of a large depot the boiloff problem in LEO may go away. The square cube law helps. Plus constantly arriving sub cooled propellant. What's left of boiloff may justbe accepted for the sake of simplicity of operations.

True but your tank size is going to be constrained by the size of the launch vehicle payload, unless you weld it together in orbit (also a possibility). A 12m diameter tank should be fine though - and hold plenty of propellant.

Plus you have to make ullage burns to get the stuff flowing, or else use something clever like low-temperature bladders. Rotate for (very weak) artificial gravity? That'll probably cause more problems than it solves.
With two sets of tanks, it should be possible to arrange for flow of the liquid in the tanks using low velocity mixers.  If the mixers were contra rotating, the angular acceleration would tend to cancel out.  It would be important to keep velocities low, and to be able to vary the velocity of the mixers for operation with mixed phase fluids.  Rotating the whole tank seems complex, and might require rotating joints for fluid transfer, which are weak points.
As a negative point the mixers would add energy and possible increase boil off because there would be constant friction, but this could be offset by using active mechanical cooling of the fuel.  Careful operation should create laminar flow; that would create the lowest friction.

Offline wes_wilson

  • Armchair Rocketeer
  • Full Member
  • *
  • Posts: 120
  • Florida
  • Liked: 54
  • Likes Given: 96
Re: MCT Speculation and Discussion Thread 4
« Reply #301 on: 06/29/2015 03:52 PM »
With the size of a large depot the boiloff problem in LEO may go away. The square cube law helps. Plus constantly arriving sub cooled propellant. What's left of boiloff may justbe accepted for the sake of simplicity of operations.

True but your tank size is going to be constrained by the size of the launch vehicle payload, unless you weld it together in orbit (also a possibility). A 12m diameter tank should be fine though - and hold plenty of propellant.

Plus you have to make ullage burns to get the stuff flowing, or else use something clever like low-temperature bladders. Rotate for (very weak) artificial gravity? That'll probably cause more problems than it solves.
With two sets of tanks, it should be possible to arrange for flow of the liquid in the tanks using low velocity mixers.  If the mixers were contra rotating, the angular acceleration would tend to cancel out.  It would be important to keep velocities low, and to be able to vary the velocity of the mixers for operation with mixed phase fluids.  Rotating the whole tank seems complex, and might require rotating joints for fluid transfer, which are weak points.
As a negative point the mixers would add energy and possible increase boil off because there would be constant friction, but this could be offset by using active mechanical cooling of the fuel.  Careful operation should create laminar flow; that would create the lowest friction.

Following on to the talk about having the depot behind a sunshade; a less mechanically complex possibility might be to move the O as solid bricks from the depot to the spacecraft fuel tanks; then move the spacecraft into the sun to liquefy the O in the fuel tanks.   

Oxygen solidifies at 54K under 1 atm pressure.

Nasa's site says the JWST sunshield allows passive cooling to 39K. 
http://jwst.nasa.gov/sunshield.html
@SpaceX "When can I buy my ticket to Mars?"

Offline lamontagne

  • Full Member
  • ****
  • Posts: 1082
  • Liked: 1541
  • Likes Given: 241
Re: MCT Speculation and Discussion Thread 4
« Reply #302 on: 06/29/2015 04:30 PM »
Rather than guessing this or that, how about some math?

The solar constant is about 1300 W/m2.  Since the depot is in the Earth's shadow 50% of the time this can be reduced to 650 W/m2 average.  A good reflector foil wrap with a low emissivity of e=0.1 will reduce this to 65 W/m2. Depending on insulation effectiveness, this energy will either be absorbed into the fuel, or radiated back out into space.  If the insulation was 100% effective, the exterior hull temperature would be determined by Q=Be(Ts^4-ta^4), where B is 5.7e-8, e is the emissivity of 0.1 and ta is the average ambient temperature in earth orbit, that I believe NASA usually sets at about 200°K.  Solving this gives a surface temperature of 337°K, or 65°C.
On the other hand, if the thermal resistance is 0, all the energy goes into the fuel and it boils off at the phase change rate of oxygen or methane; using 480 kJ/kg for methane, that's about 0.000135 kg/s/m2 of exposed hull area.
So you want to reduce boil off, and for this you need to insulate the tank.  To know the insulation effectiveness you can use the equation for multilayer insulation, where Q=UAdt, in which U is the insulation value, and is calculated by U=4BT^3 * (1/(N(2/e)-1)+1) where n is the number of layers.  dt is the difference between surface temperature and fuel temperature, T is the average temperature in the insulation.

If you wanted to reduce the heat gain to, let's say, 6,5 W/m2, or a factor of 10, you need an insulation value of 6.5 / (300K - 108 K) = 0,03 W/m^2K.  (300 K is an initial guess, this usually gets solved by iteration)

Knowing that T = 204K

Then 0.03=4*5.7e-8*204*(1/(N(2/e)-1)+1)
You just need to isolate N and you get the number of layers required.
Calculating N is left as an algebra exercise for the reader  ;-) .  It's my lunch break after all.  but it should be about 3 or 4.

In a sense, a shadow shield is simply a extra layer of reflective insulation, put further away in space. 
Eventually, residual heat gain can be removed by mechanical compression. This in a very inefficient process, specially at low temperatures, where Qcold = Qhot*(Tcold/Thot).
For a residual load of 6.5 W tcold at 108K and thot at 324 K, Qhot =3*6.5 = 19.5W, and the total energy radiated out by the cooling system will be 26 Watts, using 19,5 watts of electrical energy.  This is an ideal Carnot process, in the real world it should be about twice that, or about 40 W per m2 of hull area.  Since solar cells can deliver 1300 *.2*.5=325 watts on average, they can provide the required energy.

Although this is full  of outrageous simplifications, I hope this helps.
Michel Lamontagne


« Last Edit: 06/29/2015 05:18 PM by lamontagne »

Offline Lobo

  • Senior Member
  • *****
  • Posts: 6739
  • Spokane, WA
  • Liked: 543
  • Likes Given: 335
Re: MCT Speculation and Discussion Thread 4
« Reply #303 on: 06/30/2015 09:45 PM »
That's pretty good mathery.

I'll chip in that the ULA ACES depot architecture wasn't trying to achieve zero boiloff becuase it had station keeping needs.  So they just needed to mitigate the boiloff until it was about equal to what was needed for station keeping.  Methalox RCS thrusters would burn those boiloff gasses to keep the depot in proper orbit and position.

I'd assume that'd be the case here, but easier to do with LCH4 instead of LH2.


Offline Burninate

  • Full Member
  • ****
  • Posts: 1130
  • Liked: 346
  • Likes Given: 72
Re: MCT Speculation and Discussion Thread 4
« Reply #304 on: 06/30/2015 10:05 PM »
Rather than guessing this or that, how about some math?

The solar constant is about 1300 W/m2.  Since the depot is in the Earth's shadow 50% of the time this can be reduced to 650 W/m2 average.  A good reflector foil wrap with a low emissivity of e=0.1 will reduce this to 65 W/m2. Depending on insulation effectiveness, this energy will either be absorbed into the fuel, or radiated back out into space.  If the insulation was 100% effective, the exterior hull temperature would be determined by Q=Be(Ts^4-ta^4), where B is 5.7e-8, e is the emissivity of 0.1 and ta is the average ambient temperature in earth orbit, that I believe NASA usually sets at about 200°K.  Solving this gives a surface temperature of 337°K, or 65°C.

I don't quite understand this.  The effective temperature of a blackbody at 1AU should be 254K.  You can set the emissivity fairly high (for rocky Earth, emissivity approaches 1.0), but albedo is the variable you can tweak heavily, and with higher albedo should come lower hull temperatures.

In Low Earth Orbit, you have the added factor of a hemisphere radiating at ~288K (may be subject to some corrections).  But nothing, as far as I understand it, should raise the hull temperature up to 337K.

Offline lamontagne

  • Full Member
  • ****
  • Posts: 1082
  • Liked: 1541
  • Likes Given: 241
Re: MCT Speculation and Discussion Thread 4
« Reply #305 on: 07/01/2015 12:36 AM »
Rather than guessing this or that, how about some math?

The solar constant is about 1300 W/m2.  Since the depot is in the Earth's shadow 50% of the time this can be reduced to 650 W/m2 average.  A good reflector foil wrap with a low emissivity of e=0.1 will reduce this to 65 W/m2. Depending on insulation effectiveness, this energy will either be absorbed into the fuel, or radiated back out into space.  If the insulation was 100% effective, the exterior hull temperature would be determined by Q=Be(Ts^4-ta^4), where B is 5.7e-8, e is the emissivity of 0.1 and ta is the average ambient temperature in earth orbit, that I believe NASA usually sets at about 200°K.  Solving this gives a surface temperature of 337°K, or 65°C.

I don't quite understand this.  The effective temperature of a blackbody at 1AU should be 254K.  You can set the emissivity fairly high (for rocky Earth, emissivity approaches 1.0), but albedo is the variable you can tweak heavily, and with higher albedo should come lower hull temperatures.

In Low Earth Orbit, you have the added factor of a hemisphere radiating at ~288K (may be subject to some corrections).  But nothing, as far as I understand it, should raise the hull temperature up to 337K.

The fuel depot has low emissivity and likely fairly high albedo, these are usually more or less inversely proportional.  So it reflects a lot of radiation (high albedo), and therefore absorbs very little , but for the radiation it doesn't reflect, it had a lot of difficulty getting rid of, so it has to heat up quite a bit.  That is why you can burn yourself on a piece of aluminium left in the sun.  The depot is very far from a black body, in that sense.  My use of e only in the calculations was just a quick simplification.  But the surface temperature can get quite hot.

Here is a little spreadsheet, see the cooling tab.
https://docs.google.com/spreadsheets/d/11edSaSqnDQeWBPgz1XMa4E3X0R7hPWcTS5EYxe853U8/edit#gid=2038396500

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 6571
  • Germany
  • Liked: 1687
  • Likes Given: 1599
Re: MCT Speculation and Discussion Thread 4
« Reply #306 on: 07/01/2015 06:46 AM »
The fuel depot has low emissivity and likely fairly high albedo, these are usually more or less inversely proportional.  So it reflects a lot of radiation (high albedo), and therefore absorbs very little , but for the radiation it doesn't reflect, it had a lot of difficulty getting rid of, so it has to heat up quite a bit. 

In another discussion it was mentioned that there is something as simple as a paint that combines both properties. It is high albedo - white - in the visible spectrum where the sun emits most of its energy and at the same time low albedo in infrared where a depot needs to get rid of excess energy.

Offline lamontagne

  • Full Member
  • ****
  • Posts: 1082
  • Liked: 1541
  • Likes Given: 241
Re: MCT Speculation and Discussion Thread 4
« Reply #307 on: 07/02/2015 12:15 AM »
The fuel depot has low emissivity and likely fairly high albedo, these are usually more or less inversely proportional.  So it reflects a lot of radiation (high albedo), and therefore absorbs very little , but for the radiation it doesn't reflect, it had a lot of difficulty getting rid of, so it has to heat up quite a bit. 

In another discussion it was mentioned that there is something as simple as a paint that combines both properties. It is high albedo - white - in the visible spectrum where the sun emits most of its energy and at the same time low albedo in infrared where a depot needs to get rid of excess energy.

Yes, high albedo paint is readily available for roofing.  Commercial sites rate this paint at 0,9 emissivity and 0,9 albedo.  However, it seems that it breaks down somewhat with time and that a more reasonable value for the albedo would be 0,5 or 0,6.  This still brings down the surface temperature to something like 240K, or -32C. 
But that doesn't tell the whole story, because as aluminium is very conductive, and in space there is no external surface convection to insulate it, it will lose (or gain) energy at a rate of 200 W/m2K.  So what happens then is that the surface temperture goes down to an equilibrium value between the fuel temperature and temperture from the solar heat gain, and the tank loses energy.  With the methane fuel at 160K, even a surface temperature as low as 161K would mean 200W/m2 of heat gain! (in a detailled analysis, the methane itself is not such a good conductor, and will not provide heat at this rate; in a sense the methane will act as its own insulator). Therefore, some form of insulation is needed, no matter what .  The good news is that even a thin sheet of aluminised mylar will reduce the heat gain significantly or any form of trapped gas in a foam or blanket will quicly bring this down one or 2 orders of magnitude.
 

Offline philw1776

  • Full Member
  • ****
  • Posts: 957
  • Seacoast NH
  • Liked: 563
  • Likes Given: 267
Re: MCT Speculation and Discussion Thread 4
« Reply #308 on: 07/02/2015 08:01 PM »
One way to speculate about the BFR launcher for the MCT is to look at mass fraction to LEO efficiency.  The F9 has a mass fraction of ~2.7% to LEO.  Despite full re-usability most of us fanboys/girls here expect that SX will somehow improve on that with the fully re-useable BFR. The tables below 1st assume 180mT to LEO with the dry MCT massing 80mT and the 2nd table assumes a dry MCT at 100mT.  Various optimistic mass fractions yield different BFR takeoff weights.  A T/W ratio of 1.2 is assumed yielding 1st stage thrust and dividing by 500KLBs/Raptor, the # of Raptors needed.

MCT + Payload = 180mT to LEO 

MASS FRACTION   BFR         BFR TAKEOFF      
TO LEO          WEIGHT mT      WEIGHT M LBS   THRUST M LBS     # RAPTORS @ 500KLB
5.0%                3500              7.7                     9.2                       19
                     
4.5%                3889              8.6                    10.3                       21
                     
4.0%                4375              9.6                    11.6                       23
                     
3.5%                5000            11.0                    13.2                       26
                     
3.0%                5833            12.8                    15                         31
                     
MCT + Payload = 200mT to LEO                    
MASS FRACTION                  
TO LEO        BFR                 BFR TAKEOFF      
                WEIGHT mT          WEIGHT M LBS     THRUST M LBS       # RAPTORS
5.0%                4000               8.8                   10.6                      21
                     
4.5%                4444               9.8                   11.7                      24
                     
4.0%                5000             11.0                   13.2                      26
                     
3.5%                5714             12.6                   15.1                      30

Assuming Raptor engine bells are ~1.6m wide, it's likely that 1st stage diameters of over 10m are preferred with 12.5m or even better 13.5m best to allow for max # of engines in case mass fraction drops.  A smaller MCT dry weight really helps reduce BFR mass & # of engines as would be expected.


“When it looks more like an alien dreadnought, that’s when you know you’ve won.”

Offline Impaler

  • Full Member
  • ****
  • Posts: 1283
  • South Hill, Virgina
  • Liked: 363
  • Likes Given: 0
Re: MCT Speculation and Discussion Thread 4
« Reply #309 on: 07/03/2015 05:16 PM »
For comparison here are the numbers for SLS

70 mT to LEO

MASS FRACTION   SLS         SLS TAKEOFF     
TO LEO          WEIGHT mT      WEIGHT M LBS   THRUST M LBS     

2.8%                2500              5.5                    8.4 

130 mT to LEO

MASS FRACTION   SLS         SLS TAKEOFF     
TO LEO          WEIGHT mT      WEIGHT M LBS   THRUST M LBS     

4.3%                3000              6.5                    9.2                     
                     
                   

Offline philw1776

  • Full Member
  • ****
  • Posts: 957
  • Seacoast NH
  • Liked: 563
  • Likes Given: 267
Re: MCT Speculation and Discussion Thread 4
« Reply #310 on: 07/04/2015 05:11 PM »
Great perspective with the SLS #s.
Going with the 200mT to LEO target or possibly less it looks as if the BFR will have somewhere in the mid to high 20s # of 1st stage engines.  This likely means a single core (Musk's statement) would have over 10m diameter to house the # of engines.  Around 12.5 to 13.5m seems like a reasonable guess.  As stated by Elon, the BFR mainframe will be manufactured "close" to the launch site. This BFR will likely be a short squat looking beast compared to the taller Falcon 9.  The next question is...how wide is the upper stage, assumed to be the MCT itself?  Rockets like the Falcon & others have wide "payload" fairings. Wider is better for carrying several tens of people for several months along with many tens of metric tons of cargo.  But atmospheric entry requirements will likely drive  MCT form factor design decisions for a vehicle that goes to Mars & back from the Earth's surface.

Of course this assumes that the whole thing (MCT) goes to Mars & lands, the most conservative assumption based on what little has been said.  I would not be surprised if a very different approach was announced later this year or next year such as a SEP planetary transit vehicle. Right now, BFR & MCT have to be on Elon's back burner.
« Last Edit: 07/04/2015 05:12 PM by philw1776 »
“When it looks more like an alien dreadnought, that’s when you know you’ve won.”

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 6571
  • Germany
  • Liked: 1687
  • Likes Given: 1599
Re: MCT Speculation and Discussion Thread 4
« Reply #311 on: 07/04/2015 07:36 PM »
The clearest proof that landing is the plan was the statement that for the first crews MCT would be the habitat on Mars.

Offline Impaler

  • Full Member
  • ****
  • Posts: 1283
  • South Hill, Virgina
  • Liked: 363
  • Likes Given: 0
Re: MCT Speculation and Discussion Thread 4
« Reply #312 on: 07/05/2015 06:00 AM »
The clearest proof that landing is the plan was the statement that for the first crews MCT would be the habitat on Mars.

And their is nothing inconsistent with that statement and having a SEP transit stage.  Musk calls it the "Mars Colonial transport SYSTEM" which clearly implies multiple parts such as the BFR first stages and what ever LEO propellent depots are need, neither of which will go to Mars.

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 6571
  • Germany
  • Liked: 1687
  • Likes Given: 1599
Re: MCT Speculation and Discussion Thread 4
« Reply #313 on: 07/05/2015 06:21 AM »
The clearest proof that landing is the plan was the statement that for the first crews MCT would be the habitat on Mars.

And their is nothing inconsistent with that statement and having a SEP transit stage.  Musk calls it the "Mars Colonial transport SYSTEM" which clearly implies multiple parts such as the BFR first stages and what ever LEO propellent depots are need, neither of which will go to Mars.

True, but it is by no means a positive proof for that concept. The infrastructure of depots in LEO and ISRU propellant production on Mars is a system by itself.

I recall Elon Musk saying something like it can be done with chemical propulsion, no advanced propulsion systems are necessary. As far as I know he never repeated that statement and it does not preclude SEP. I would not be too surprised if it is added at some point in time to improve efficiency. But I am very sure it will not be part of the initial system at the time when a first base is set up because it adds complexity.

Offline Ionmars

  • Full Member
  • ****
  • Posts: 785
  • North Carolina, USA
    • The Mars Pioneer
  • Liked: 221
  • Likes Given: 492
Re: MCT Speculation and Discussion Thread 4
« Reply #314 on: 07/05/2015 12:24 PM »
...
...
I recall Elon Musk saying something like it can be done with chemical propulsion, no advanced propulsion systems are necessary. As far as I know he never repeated that statement and it does not preclude SEP. I would not be too surprised if it is added at some point in time to improve efficiency. But I am very sure it will not be part of the initial system at the time when a first base is set up because it adds complexity.
Given the depot system discussed here and given that Elon thinks an all propulsive system could work, consider the following:

Four fully-fueled Tanker-MCT (TMCT) are clustered around and attached to a Mars-Bound-MCT (MBMCT). The engines of the four TMCTs are lit and push the cluster to HEO using about 1/2 of their fuel. The MBMCT is released and begins its TMI burn while the remaining four return to LEO and then individually RTLS. Alternatively, just to a fuel depot in LEO.

Feasible?

Edit: Starting point is LEO.
« Last Edit: 07/05/2015 12:46 PM by Ionmars »
* Mars' orbit: a convenient service station for an asteroid-sized spaceship en-route to Ceres. *

Offline philw1776

  • Full Member
  • ****
  • Posts: 957
  • Seacoast NH
  • Liked: 563
  • Likes Given: 267
Re: MCT Speculation and Discussion Thread 4
« Reply #315 on: 07/05/2015 12:39 PM »
Gwenn Shotwell in a 2015 video interview, "We're looking at SEP"

Anyone who thinks that any aspect of MCT is cast in stone right now is likely mistaken.  It's in the concept stage where many alternatives are considered, not deep into the design stage.

I took the conservative approach following Musk's long ago , "Land the whole thing" remark.
« Last Edit: 07/05/2015 12:43 PM by philw1776 »
“When it looks more like an alien dreadnought, that’s when you know you’ve won.”

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 6571
  • Germany
  • Liked: 1687
  • Likes Given: 1599
Re: MCT Speculation and Discussion Thread 4
« Reply #316 on: 07/05/2015 01:56 PM »
Four fully-fueled Tanker-MCT (TMCT) are clustered around and attached to a Mars-Bound-MCT (MBMCT). The engines of the four TMCTs are lit and push the cluster to HEO using about 1/2 of their fuel. The MBMCT is released and begins its TMI burn while the remaining four return to LEO and then individually RTLS. Alternatively, just to a fuel depot in LEO.

Feasible?

Edit: Starting point is LEO.

I don't think it is a very efficient architecture. It means several MCT with all their mass would need to be accelerated a significant part of TMI. Also you mention using half of their fuel. It would not be necessary to reserve half of the fuel for return. Injecting into a highly ellicptic orbit would give the Mars bound MCT much of the needed delta-v and brings the booster MCT back to earth basically free.

Why do you propose to get them back to LEO? More efficient to land them for a new launch with payload.

I think the most efficient way is giving MCT tanks large enough to do TMI burn and Mars EDL by themselves. Use tanker MCT to refuel in LEO either directly fuelling up an MCT or filling depots. They need that tankage and the delta-v to get back to earth from the Mars surface.

Offline Ionmars

  • Full Member
  • ****
  • Posts: 785
  • North Carolina, USA
    • The Mars Pioneer
  • Liked: 221
  • Likes Given: 492
Re: MCT Speculation and Discussion Thread 4
« Reply #317 on: 07/05/2015 05:14 PM »
...
...
I don't think it is a very efficient architecture. It means several MCT with all their mass would need to be accelerated a significant part of TMI. Also you mention using half of their fuel. It would not be necessary to reserve half of the fuel for return. Injecting into a highly ellicptic orbit would give the Mars bound MCT much of the needed delta-v and brings the booster MCT back to earth basically free.

Why do you propose to get them back to LEO? More efficient to land them for a new launch with payload.

I think the most efficient way is giving MCT tanks large enough to do TMI burn and Mars EDL by themselves. Use tanker MCT to refuel in LEO either directly fuelling up an MCT or filling depots. They need that tankage and the delta-v to get back to earth from the Mars surface.
You are right, it's not the most fuel efficient. But it might be a method for reaching Mars faster than the least-energy transfer orbit to reach Mars, perhaps in in 3-4 months rather than 6. Also a method that would represent a non-SEP architecture if Elon is serious about it. I am sure this has already been addressed somewhere and lies on someone's spreadsheet.
* Mars' orbit: a convenient service station for an asteroid-sized spaceship en-route to Ceres. *

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 6571
  • Germany
  • Liked: 1687
  • Likes Given: 1599
Re: MCT Speculation and Discussion Thread 4
« Reply #318 on: 07/06/2015 06:00 AM »
Plus you have to make ullage burns to get the stuff flowing, or else use something clever like low-temperature bladders. Rotate for (very weak) artificial gravity? That'll probably cause more problems than it solves.

Maybe a totally crazy idea. But could a fan be used to herd the propellant to the pumps? That would save the need to accelerate a depot with thousands of tons of propellant so it can settle.

Offline CyclerPilot

  • Member
  • Posts: 97
  • USA
  • Liked: 23
  • Likes Given: 0
Re: MCT Speculation and Discussion Thread 4
« Reply #319 on: 07/07/2015 01:36 AM »
Plus you have to make ullage burns to get the stuff flowing, or else use something clever like low-temperature bladders. Rotate for (very weak) artificial gravity? That'll probably cause more problems than it solves.

Maybe a totally crazy idea. But could a fan be used to herd the propellant to the pumps? That would save the need to accelerate a depot with thousands of tons of propellant so it can settle.
Good ideas.  I would think the fan could work if it covered most of the diameter.  It could be an interesting ISS or Dragon lab experiment.

I was trying to come up with a KISS transfer solution that could work between MCTs.  If one tank was actively heated/boiled and the other actively cooled to below boiling, it would all transfer except the residual gas.  There would be a constant pressure differential and flow which should keep the cold liquid on its own side.

Tags: