abcha0s' 300G Ultimate Reef

[abcha0s]

The target temp for my tank is:

Low Temp: 78F (25.56C)
High Temp: 78.8F (26C)

The goal is to maximize temperature stability within this range. The tank temperature should fluctuate less than 0.5 degrees Fahrenheit.

Cooling
As it turns out, the more energy efficient hardware is, the less heat it produces. All of my hardware was selected based in part on energy efficiency. Most notable is the LED lighting which is significantly more energy efficient then alternative lighting sources. The upside to this is very little heat transfer into the water. For most of the year, I do not need any form of cooling.

The tank is located in my basement which tends to stay slightly cooler than the rest of the house. There is constant evaporation which translates to heat loss, so the tank tends to find a temperature balance slightly below the ambient room temperature. For the occasional heat wave, I have a fan in the sump that blows over the water surface to increase evaporation and cooling. The fan is connected to the Apex controller for temperature control.

Heating
Ironically, the downside to energy efficiency is also a lack of heat transfer into the water. My tank water stabilizes at room temperature and needs to be constantly heated (day and night). The energy consumption of the heaters required to keep the tank at a constant 78F cancels out a significant portion of the savings gained through energy efficient hardware. In my case I estimate this to be approximately 700 watts (~6 Amp/h) of heating.

Background Reading

BeanAnimal - Aqarium Heaters: What you need to know!

BeanAnimal - Thermodynamics for the Reef Aquarist

abcha0s - My Marineland 300W Stealth Pro Heater Exploded!

Heating my tank

I am using 4 300W Titanium Heating Tubes.

• The Heaters are Finnex Titanium Heaters. I don't recommend this brand, nor expect it to work indefinitely. With that said, I haven't found any brand of heater where people haven't complained about failures. Ebo Jager seems to have the best overall user reviews, but it's still not hard to find incidences of failure.
• Based on a very general rule of thumb, I need 3-5W of heating per gallon of tank water. For a 300G tank, this equates to 900W to 1500W.
• I had heard that Titanium heaters were more efficient than glass heaters. After testing the two types of heaters I can't really see much of a difference. If the Titanium heaters are more efficient, it is a very small percentage.
• I went with a heater without a thermostat because the thermostat is likely to fail anyway. One less component to worry about.
• The Apex Energy Bar 8 (EB8) is rated to 5Amps per outlet. The basically means that the biggest heater I can use is 500W.
• Several small heaters are better than one larger heater.
• I went with 4 of the 300W heaters instead of 3 of the 500W heaters for efficiency and redundancy reasons. The 300W heaters have a 10 inch heating tube. The 500W heaters have an 11.5 inch heating tube. My assumption is that the diameter of the tubes is the same; therefore, the 300W version has 30W per inch. The 500W has 43.5W per inch. More surface area should make the 300W more efficient.

I know from observation that two 300W heaters (600 Watts total) running continuously will heat my tank about 9F (5C) above the ambient room temperature. To maintain a stable tank temperature, I really need a minimum of three 300W heaters (900 Watts total). In this configuration, the additional 300 Watt heater operates about 75% of the time.



This graph shows three interesting things.

• The data to the left of 8:45pm is my tank with 2x300W heaters. You can see that it struggles to keep the temp between about 76.5 and 77.5. At this point, the heaters are on constantly and are unable to raise the temp to the target of 78F.
• The drop in temperature at around 8:45 occured when I turned the heaters off to do some maintenace in the sump. The temperature dropped very rapidly.
  Note: The temperature probe is in the sump. At the time these temperature measurements were taken, the return pumps were also shutoff meaning that the sump was isolated from the display. The display tank has thicker glass so is likely somewhat more insulated, but I wouldn't count on that to keep the tank warm. This could be a real problem in a power outage.
• The data to the right of about 10pm is the tank with 3x300W heaters. With the increased heating capacity, the heaters really have no problem reaching the target temp.

The heaters are installed in the return section of the sump. The temperature probe is in the chamber right before the return section. Therefore, heated water has to circulate through the tank and back to the sump before the temperature sensor is able to detect a change. There is an approximate 0.5F temperature differential between the display tank and the sump.



The 4 Titanium Heaters are split evenly between two Apex controllers on independent EB8s. Each EB8 is connected to a separate 15A circuit. This should allow for the failure of three heating tubes before the tank water temperature begins to drop. The failure of one Apex/EB8 or 15A circuit would result in the loss of 2 heating tubes. The heating tubes don't have a thermostat, so they are basically stuck on already. If for some reason the EB8 outlets on one Apex stuck on, the other Apex should shut both heaters off. If one Apex completely failed, the surviving controller could maintain the tank temperature. This equates to "no single point of failure".

Another benefit of using two Apex controllers with independent temperature probes is the ability to identify any calibration drift between the probes. As the probes are both installed in the same sump chamber, the temperature readings should always be within about 0.1F of each other. If this drift increases, it is an indication that one of the probes is not operating at peak performance and that maintenance should be preformed.

Finally, if the temperature of the tank does deviate either up or down, both Apex controllers are configured to send email alerts to my phone.

When operating normally, the temperature characteristics should look like this: