Need expertise on heater function

[simplefishery]

Hello, I have a question about heaters and am hoping someone who understands the internals can help. I have a very modest/small tank (20 gallon long). It had a simple, cheap 100W heater, probably made in China but I'm not certain. Since I always feared that someday it would fail without my notice, I built a monitoring system using a submersible temperature probe and Raspberry Pi, and set it up to alert me if the temperature went outside my preferred range. (I prefer to keep the tank around 79 F.) So recently, the heater failed and alerted me. I bought a new one, also cheap and similar to the old one. But this time I decided to actually examine the data. The heater keeps the tank within a pretty tight range of about 0.3 deg, but the median temp drops at nighttime. If the heater is thermostatically controlled, then the behavior I'm seeing doesn't make sense. And yet, with the tank fairly stable and the heater element off, if I just tweak the adjustment knob a tiny bit higher, the heater element turns on, which seems to indicate that it is thermostatically controlled. I've attached a pic showing the temperature vs. time behavior for a 24-hour period. I can't complain about the range it's maintaining overall, since it's right around one full degree F. But it appears that if our nighttime temps drop significantly this winter, it could be a problem. (Our home thermostat setting won't change during the winter, but the tank is right next to an external wall.)
Does anyone know how these cheap heaters work?

 

[Colin_T]

Virtually all aquarium heaters are made in China regardless of the brand. Some brands are better quality than others and have more reliability. In my experience, Chinese branded electronics seem to have less quality control and be more likely to fail prematurely. This is based on lots of electronic devices I have bought over the last 10 years that were made in China. Devices made by Chinese owned companies have definitely been lower quality than better known brands owned in other countries.

Virtually all aquarium heaters have a thermostat and heating coil. The thermostat is at the top of the heater where the cord and temperature adjustment dial/ knob are. The heating coil is at the bottom of the heater and is a wire spring wrapped around a white ceramic rod quite a few times.

If the heater is unable to keep the temperature stable when the room temperature drops at night (or anytime) the heater is not big enough for the tank. You can compensate for this by insulating the aquarium with 1-2 inch thick sheets of polystyrene foam. Put a sheet of foam under the aquarium and tape a piece to the back and 2 sides of the tank. That will insulate those sides and reduce the amount of work for the heater. It also reduces your power bill a little bit.

You can put a coverglass on the tank if it doesn't have one. Use 4, 5 or 6mm thick glass and it will help trap heat in the aquarium and the heater won't have to work as hard or as often to maintain the temperature.

If the temperature still fluctuates after insulating the aquarium, it should be returned because it's dodgy. However, 1 degree Fahrenheit isn't a lot of variation and most places consider that acceptable. However, if the house temperature drops a lot overnight, the heater could fail or struggle with temperatures and the water might drop more than 1 degree.

With aquarium filters and heaters, better known brands that have been around for a while (Eheim and Fluval) are more likely to be better quality than a Chinese company's version that has been reverse engineered from another brand. If you get filters or heaters, look at the warranty and don't get anything with less than a 12 month warranty. Most good brands have a 2 or more year warranty and last a lot longer than that.

 

[simplefishery]

@Colin_T, thank you very much for the awesome, detailed response. That all makes sense, and I'm glad you included the info about the better brands, since I'll likely be looking for one of those. The only thing that perplexes me though, is that between around 1am and 6am during my test session, the heater seemed to be cycling on, then turning off again at about 79.25 degrees. Did you notice that in the attachment? I can't understand why it would shut off at that point and not try to keep raising the temperature during those hours. Or is it possible that there is also some circuitry involved that only keeps the element on for a specified time?

 

[Colin_T]

Heaters will come on and warm the water to a certain temperature and then turn off. The thermostat controls this and they don't add extra circuitry to change it. A red light usually comes on when they are heating and goes off when they aren't.

If there is lots of water movement around the heater it might be on for longer because more cool water (the entire tank volume) is moving past it and the heater has to get all the water warm before it goes off. If there isn't any water movement around the heater, it won't be on for long because it only has to heat the water around it and when that water is warm, the heater turns off.

Most heaters are on for 10-30 seconds and go off for a minute or two before coming back on. In a well insulated tank they might come on for 30 seconds but remain off for 5-10 minutes.

Most heaters are not well calibrated and even though the dial might say the heater is set on 79F, it might only be set on 75 or 83 or something else. The volume of water it has to heat will also vary the temperature with bigger volumes taking longer to heat but also holding temperature for longer so it won't be on as often.

Yours seems to be holding at 79.2 for the low point and 79.7 for the high point during the day. I would say the heater is set at 79.2 and during the day the house warms up a bit and allows the heater to warm the water a bit more. At night the temperature settles and people go to bed and TVs, computers, etc, are off so there is less heat being produced by other things and the heater settles at the lower level.

Most people and I doubt any companies making aquarium heaters would bother to use a probe to measure the temperature fluctuations over a 24 hours period so you have gone above and beyond to show that the heater isn't keeping a constant temperature, although it's pretty close and outside factors could be the reason there are variations.

In the wild fish live with temperature fluctuation on a day to day basis where they swim into deeper water, which is cooler than the surface water. And season to season where it rains and the temperature drops, or during the dry season when the waterways recede and the temperature goes up. Your fish will be fine with the slight temperature variation but the fact it happens would make me monitor it and have a back-up heater nearby, or have access to a shop to get a replacement at any time.

 

[simplefishery]

That's it! My probe is far from the heater. It's heating the water very close to the thermostat to the proper temp and shutting off, but by the time it mixes with the surrounding water the overall temp ends up being lower if the ambient environment is cooler. I'll experiment with some surrounding foam and see if things change. Thanks again!

 

[seangee]

For my 20 long I had a heater at either end of the tank. Also adds redundancy in the event of a failure - especially important where you have cold winters.
FWIW its also worth having some kind of shut off if the tank gets too hot. I use inkbird controllers in my tanks for this. Fish will survive for some time if the heater stops heating, but not nearly so long if it fails to turn off due to a stuck / faulty thermostat.

 

[MaloK]

I never checked a heater as close as you... But if it's not able to cope with a 10 degree slide over time, it's not good.

For example if you set your heater to 74 it should maintain that temperature even if the room goes a lot lower for extended periods.

If your room temperature is walking on the same line as your heater. you need a greater difference to see if the heater is able to cope with drops.

Position of the heater also is a great factor. In my experience the longer it stays off and the longer it stays on between cycle the less fluctuation occurs in general temperature.

 

[Donya]

The heater keeps the tank within a pretty tight range of about 0.3 deg, but the median temp drops at nighttime. If the heater is thermostatically controlled, then the behavior I'm seeing doesn't make sense.

In addition to what others have mentioned, another completely different possibility is that there are no problems with the actual heating ability or probe measurements but your built-in thermostat has a small offset (they all do really - trial and error to figure it out on setup) and the heater thinks it's supposed to heat to the lower temp you're seeing, which it does at night, and then something else is heating your tank to the higher daytime temp - such as heat from the lighting.

That said, the mini fluctuations within the larger pattern are really suspicious to me. It's extremely regular but not at a time scale that makes sense to be measuring actual on/off cycles. I've never seen such constant and extremely regular little fluctuations like that from always-on, probe-type thermometers on my 20gal tanks. I do see potentially fast fluctuations from read-only-on-demand type probes where you push a button to get a reading, but with those there's a range of variation and it's random within the error range, not cyclic like your data shows. The regularity of that little wobble in the larger pattern makes me thing there's something else going on in the system that's reading the temperatures.

 

[seangee]

That said, the mini fluctuations within the larger pattern are really suspicious to me. It's extremely regular but not at a time scale that makes sense to be measuring actual on/off cycles. I've never seen such constant and extremely regular little fluctuations like that from always-on, probe-type thermometers on my 20gal tanks.

I was wondering if that was a safety feature [sic] that would only make the heater come on for a limited time before it stopped for a rest. If that is the case it suggests the heater is only capable of hitting the target between 11 and 5. Yes that is the warmest part of the day but I wonder what time the tank lights are on.
In my 20 long the 2 heaters were 100W each. Its not as cold in this part of the UK but mine was in an unheated and poorly insulated room (converted garage).

 

[Donya]

I was wondering if that was a safety feature [sic] that would only make the heater come on for a limited time before it stopped for a rest. If that is the case it suggests the heater is only capable of hitting the target between 11 and 5. Yes that is the warmest part of the day but I wonder what time the tank lights are on.
In my 20 long the 2 heaters were 100W each. Its not as cold in this part of the UK but mine was in an unheated and poorly insulated room (converted garage).

I've had many heaters in many tanks over the years. Some very cheap, some brand name, different design types. All the ones I've had for things bigger than a small 1-2gal bowl had indicator lights somewhere to show when they were heating vs not. They all heated for a short period of time, a few seconds like Colin said; the time scale is much smaller than is on your graph. I've had some pretty high-flow systems (since I keep marine) and have never seen cycles lasting more than a 30sec or so. Even if we assume a rapid on/off heating cycle for the mini rises, it would mean the heater is stopping or at least ceasing to heat properly for around 30 full minutes before heating again. I've never had a heater behave like that. If I did see that kind of behavior, I'd consider it faulty equipment whether it's due to the heater just bein broken or tripping some kind of internal safety feature. It's just not how those things are supposed to work. If those mini swing readings are legit and due to the heater quitting for 30min at a time, that's bad and will definitely result in temperature swings that are larger when the ambient temperature is colder. Still, I would be inclined to verify both the 30min up/down mini swings and also a few points along the larger pattern with another measurement tool, such as one of those cheap always-on digital probes that can do 0.1F (although those as well have an offset that has to be figured out through some kind of calibration, but regardless of that they should still show relative up/down changes even if not giving the same numbers as the pi). If one of those cheap stand-alone probe thermometers says the tank is a stable temp then there's something going on in the pi part of things that's affecting the readout.