Heaters

[Robin Goodfellow]

...but I don't follow that logic that says that by having the load divided into two heaters means you'll necessarily increase the chances (let alone double them) that one will give trouble. ...

hi.
Let me put some logics into your head, young lady!

This is a very complicated issue, and so please let me know if any of you find any mistake(s). The complication is, in addition to the issue of Poisson distribution and mean-time-to-failure parameter, we have to consider the wattage of the heaters as well as the temperature setting. Let's not do it formally with null hypothesis and other technicality, but instead, let me explain it out in pose informally.

Assumptions: The heaters has perfect thermostat, and heat transfer from the heater to the tank water is instanteous and perfect, and temperature of the water is homogenous thru out the entire tank.

1, control: a 200W heater with temperature setting of 75 degree.

2a, let's assume the heaters are two 200 W heaters with identical failure parameter as the control and have identical set point of 75 degree.

2b, two 200W heaters ... and have different set point of 74 deg and 76 deg. (temp. difference of 2 deg.)

2c, two 100W heaters ... and set point of 75 degree.

2d, two 100W heaters ... and set point of 74 deg. and 76 deg.

3a, a 100W heater along is sufficient to keep the temp. to within 2 deg.

3b, a 200W heater along is sufficient, but a 100W is not sufficient.

3c, a 200W heater is not suffiicient to keep the temp to within the 2 deg.

So now, we have this 4X3 matrix that we have to fill out. The failure rate will either be the same compare to the control, or it is less or more. So put the result in applicable cell in the matrix... (some cells are not legit.)

I will come back to the matrix, but if anyone of you get this done, be free to post the result....

 

[Merkur]

i run 400 watt metal halides as my heater.

i guess i would third the option for adding "none" to the poll. adding heat is the LAST thing i want to do. in fact, back in dallas i even had trouble during the winter!

 

[Anonymous]

This is a very complicated issue, and so please let me know if any of you find any mistake(s). The complication is, in addition to the issue of Poisson distribution and mean-time-to-failure parameter, we have to consider the wattage of the heaters as well as the temperature setting. Let's not do it formally with null hypothesis and other technicality, but instead, let me explain it out in pose informally.

Assumptions: The heaters has perfect thermostat, and heat transfer from the heater to the tank water is instanteous and perfect, and temperature of the water is homogenous thru out the entire tank.

1, control: a 200W heater with temperature setting of 75 degree.

2a, let's assume the heaters are two 200 W heaters with identical failure parameter as the control and have identical set point of 75 degree.

2b, two 200W heaters ... and have different set point of 74 deg and 76 deg. (temp. difference of 2 deg.)

2c, two 100W heaters ... and set point of 75 degree.

2d, two 100W heaters ... and set point of 74 deg. and 76 deg.

3a, a 100W heater along is sufficient to keep the temp. to within 2 deg.

3b, a 200W heater along is sufficient, but a 100W is not sufficient.

3c, a 200W heater is not suffiicient to keep the temp to within the 2 deg.

So now, we have this 4X3 matrix that we have to fill out. The failure rate will either be the same compare to the control, or it is less or more. So put the result in applicable cell in the matrix... (some cells are not legit.)

I will come back to the matrix, but if anyone of you get this done, be free to post the result....

I was thinking more along the lines of "if you flip a coin twice you will have double the chance of getting heads than if you flip it once"

 

[MarkO1]

Seamaiden, yes, even in SMashechusetts... let me explain.

My tank has been up for 20 months. It is heated:
1. Primarily from the ambient room temp (kept warm typically between 70-72º in winter and maybe a bit warmer in summer (up to 75º or so)). I live on the second floor of a condo in an old mill building http://vze2372e.phpwebhosting.com/postn ... /index.php and our heat is included in our fees! Needless to say, the heating system is very effective and the wife likes the condo warm.
2. The second heat source is from the 4 pumps. I believe these pumps increase my tank temp by about 4-6 degrees.
3. My 542 watts of lights add about another 6 degrees to my water temp.

In 20 months, the coldest it has ever gotten in the tank is 75º, and that's when the room temp drops to about 60-65 overnight. Even at 75º, I have not noticed any ill effects.
_________________
gremlin plants

 

[Texas2Turkey]

Statistically speaking, each time you flip a coin your odds are still 50:50. But the odds of hitting heads every time you flip gets worse and worse each time you flip.

Having two heaters is more like flipping two coins and hoping either they both hit heads or both hit tails each time you flip them ...?

Although I'd bet money that statistically speaking most heaters have better odds of success than that of a coin toss...

 

[Anonymous]

8O OH MY GOD! I'M SO CONFUSED!!!! :lol:

<sm walks away muttering to herself..so if we have heaters in a pose, we can make assumptions that they're 200W...?>

 

[Robin Goodfellow]

hi.
Here is what I came up with. Flame away!

Assumptions:
a, if a heater were never used, it will last forever.
b, the more you use a heater, the higher the failure rate.
c, failure rate is independent of the wattage.
d, if 50% of heaters fails in first 6 months, and if you have 2 heaters, then the probability of having at least one heater fails in the first 6 months will be 75% (1 - 1/2 * 1/2). Note: this is not a very good method to calculate it, but 75% is close enough. This is true of you will run them for same amount of time.

2a, 3a: both heaters will be on at the same time, but since the wattage is higher, they are on for lesser amount of time than the control. Failure rate is about 37.5% (75% /2), so it is better than control.

2a, 3b and 2a, 3c: see above for 2a, 3a.

2b, 3a: 2nd heater never turns on, since the first one can handle everything. Failure rate is 50%, same as when only one heater is used.

2b, 3b and 2b, 3c: see 2b, 3a above.

2c, 3a: similar to 2a, 3a, but since the heaters are of lower wattage, the failure will be 75%.

2c, 3b and 2c, 3c: see 2c, 3a above.

2d, 3a: similar to 2b, 3a, but since the first heater is of lower wattage, it will work more, and result in a higher failure rate (>50%)

2d, 3b: the first heater will be on more often than the second one, so the failure rate will be a function of how the work is shared between the heater… failure rate: don’t know, but pretty close to 50% if not higher.

2d, 3c: similar to 2d, 3b, but the work is shared more evenly between the heaters. Failure rate will be almost, but not as high, as 75%.

conclusion:

__ 2a 2b 2c 2d
3a .G . S .W .W
3b .G . S .W .W
3c .G .S .W .W

where G=better, s= same, w=worse

 

[Robin Goodfellow]

hi.
Oh, when you see the first column to be all good, keep in mind that you assume that the heater is perfect. You will be better off using one 400W heater instead of two 200Watters. So the real conclusion is that it is better to keep the number of heater to be as small as possible, and have a back up heater that you can plug in if the first one broke.

 

[Anonymous]

It is absolutely refreshing to see so many different opinions on this. I asked this question because I designed the stand for my tank before I knew anything about this hobby. I was planning to keep only my magnum 350 in the bottom. Now there is a 20 high sump, skimmer, collections jug, heaters, lights, live rock, and testing kits as well. I think I will stay with one heater in the sump, but when I go on vacation I am going to add a second heater set a couple of degrees cooler. I am also going to add a temp sender with a digital readout so I can see it from "my chair" where I sit and pick at the guitar all night.

Thanks for all the input

Bryan Flanigan

 

[Anonymous]

If you assume that heaters can be in one of three states: stuck on, stuck off, and good.

one heater you have the following

stuck on - bad
stuck off - bad
good - good

1 good 2 bad

two heaters you have the following

h1 good h2 good = good
h1 on h2 good = bad
h1 off h2 good = good
h1 good h2 on = bad
h1 on h2 on = bad
h1 off h2 on bad
h1 good h2 off = good
h1 on h2 off = bad
h1 off h2 off = bad

here you have 3 good 6 bad

Now it just depends on the chances of each good bad senerio.

 

[LFS42]

I would 4th (I think that's how many I counted) the no heater route.

winter 78-80, summer 82-84

 

[jimb432]

Statistically speaking, each time you flip a coin your odds are still 50:50. But the odds of hitting heads every time you flip gets worse and worse each time you flip.

Having two heaters is more like flipping two coins and hoping either they both hit heads or both hit tails each time you flip them ...?

Although I'd bet money that statistically speaking most heaters have better odds of success than that of a coin toss...

Doesn't the law of independent trials say that each coin flip is it's own event, and has nothing to do with the flip before it. And that there is exactly the same probability of flipping heads 1000 times in a row as there is of flipping heads and tails an equal number of times?

jb

 

[Anonymous]

Statistically speaking, each time you flip a coin your odds are still 50:50. But the odds of hitting heads every time you flip gets worse and worse each time you flip.

Having two heaters is more like flipping two coins and hoping either they both hit heads or both hit tails each time you flip them ...?

Although I'd bet money that statistically speaking most heaters have better odds of success than that of a coin toss...

Doesn't the law of independent trials say that each coin flip is it's own event, and has nothing to do with the flip before it. And that there is exactly the same probability of flipping heads 1000 times in a row as there is of flipping heads and tails an equal number of times?

jb

Doesn't the law of independent trails just tell you that no matter what happened the first 999 flips you still have a 50/50 chance of getting heads the 1000th time.

To determine the probability of how many head/tails you have to take into account the number of ways to get 50/50 and the total number of possibilities. there are more ways to get 50/50 then ways to get all heads.

It is true that is equally likely that you will flip 500 heads then 500 tails as all tails. Which is just as likely as flipping 499 heads one tail one head and 499 tails. ...

but taking these as the only possibilities you have 2x as much chance as 50/50 as you do all heads.

I'm too lazy to do the math but hopefully you get the idea.

 

[clutchcargo]

Maybe I missed the post that simply spells it out.

Assumptions:
You need a minimum of a 400 watts to heat the tank.

Scenarios:
You have one 400 watt heater that gets stuck on and kills the tank because when you get home from work you see that the temp is now 132F.

If you have two 200 watt heaters set to the same temp and one gets stuck on, one heater is not enough to overheat tank on it's own and the second heater will control the temperature. The bad thing is you actually have to look at the heaters every now and then to see one of them it has gone bad.

If you have two heaters and one gets stuck off then the temp will drop to a point that one heater can maintain. In this case you'll notice right away when you check the temperature but most likely nothing bad will happen.

The chances of two failures is slim but I guess it could happen.

The point is even if the ambient room temperature is high enough for 200watts of heat to overheat the tank, it will happen slower and give you more of a chance to get it taken care of.
Cjh

 

[Anonymous]

Now THAT was easy to understand. Thanks, Clutch!