Uninterruptible Power Source/ UPS Test

From the complete thread: UPS Test

 

fishfever:
This might be of some interest to anyone that experiences power outages from time to time and was considering a UPS purchase to keep things running during a blackout. On Black Friday I purchased a couple of UPS (Uninterruptible Power Supply), mostly to replace some aging ones I had for computer backup but since the sale price was so good I ended up getting 2 spare ones. I wanted to try one out for aquarium back-up power to run at least 1 filter in my tank when power is lost. The model I got was an APC 550VA/330W; it just so happens that one of the larger aquarium supply businesses also sells this model for this purpose (I can't recall who now) but I didn't know this at the time of purchase. Anyway I've been experimenting a bit with it off and on when I have spare time.

As a reference, I used a 40-watt lamp as my first test. To test the UPS, it must be fully charged (overnight) then unplugged from the wall to simulate power loss. The 40-watt lamp lasted exactly 90 minutes which was pretty good.
Next I tried some air pumps and HOB filters each by them self to isolate each test load. The results weren't nearly as good. Of the 3 air pumps I have only one ran normally. The other two did not run or the output dropped to virtually nothing. I also have 3 HOB filters and only one ran completely normally. One ran acceptably but seemed a little erratic and one did not run.


TEST
----

• 40-watt lamp - ran 90 minutes
  
• no-name small air pump that came with 3g hex tank - did not run
  
• TopFin (Petsmart brand) 60g air pump - did not run
  
• Whisper 10 air pump with 2.2-watt motor - ran normally for 65 minutes
  
• Aqueon 10 HOB - ran, forgot time but seemed to have trouble starting at first so had to cycle power manually
  
• Aqueon 30 HOB - did not run, made clicking noises
  

Via Aqua Vitalife 200 HOB with 8-watt motor - ran normally for 60 minutes



Although we generally have reliable power here and it's rare to lose power for more than a few hours, I thought this would be a nice addition to my setup; it also has surge protection built-in (it has surge-protected outlets that are not battery-backed) so it may help pro-long the life of other things (heater, UV sterilizer, etc.).
You may be wondering why the run times (for those that ran normally) are so low, especially compared to the 40-watt lamp which would appear to be a much bigger load. Without getting too technical, this particular UPS (and most of the cheaper ones) when on battery backup attempts to synthesize the normal household waveform (sinusoidal) instead of generating a true sinusoid. This can play havoc with some kinds of loads (motors in particular). Other possible problems that could reduce motor efficiency are harmonics and low power factor but I suspect the main problem here is the approximated sinusoid output (which can increase harmonics as well).

I also tried piggybacking my spare UPS to see if I could get double the run time but the second UPS would make clicking noises every few seconds when I unplugged the first UPS from the wall (meaning it was confused by the non-sinusoid output and was switching back and forth between normal and lost power mode). It should be noted that if you are wanting to run a resistive load like a heater with a UPS it should be fine (although I can't see the need for this since temperature drops rather slowly).


Carl
Interesting post fishfever

This is subject that I have actually had a lot of experience in, in part dating back to the 1980s to find a way to protect my clients tanks from power failure where I designed my own DC pumps with electric switches (I purchased at Radio Shack) to hold the circuit open while plugged into House Current (loss of power closed the circuit)

Later as I became an avid RV camper I experimented with Inverters (which convert DC to AC power), in the late 90s.

The first "cheaper" inverter (rated at 400 watts) although claiming to be a modified sine wave inverter, could not run many pieces of equipment such as drills and motors.
The second model although rated at 400 watts too (800 surge) had no problem loading up to its rating.

Both were connected to Deep Cycle batteries in parallel; by parallel, this means placing the batteries in connection in such a way that keeps the batteries at 12 volts, but nearly doubles amp output versus in series which would make the output 24 volts (4 1.5 V 'D' Batteries in series would be 6 V).
By using the same energy source, I knew this ruled out the lack of energy since you (for example) cannot connect a 1500 watt inverter to one standard auto battery.

I know true sine wave inverters are available, but according to reviews I have read from RV Journals, these give little more than the modified sine wave but for the most delicate of electronics, which your aquarium equipment does not qualify as.

What I do not know is why the more costly (triple the cost) heavy duty inverter worked better than the cheap one when the ratings were the same, but I suspect that the claim of a modified sine wave was not true, as it behaved more like a square wave inverter.

My suggestion is to get one or two deep cycle batteries and connect a good modified sine wave inverter to them, your run time will be vastly longer.

In fact in 2000, I used a couple of series 24 Deep Cycle RV Batteries and a 800 watt inverter for my Aquarium Store. We had a major blackout in 2001 and were able to run all our electronics (cash registers, etc.), pumps, lights (no heaters, air though, but then it was summer in LA). This ran fine for a few hours with a full load when power returned. What was interesting is that mobile 24/7 radio reporter for LA station KFWB was driving down the street we were located on and noticed our business was still functioning while others were in the black, and interviewed an employee of mine who showed them our system.

BTW, the Azoo continuous run pump I sell has an AC motor (at least it appears so when I took it apart since it uses a vibrator motor which require alternating current to function), so this pump is essentially an inverter as opposed to a converter (or DC switch). I should note that I only took apart the pump part, but I think this is a safe assumption.


fishfever
I was out of town for a few days but had a few minutes this morning to capture some waveforms. I used an old scope and a simple resistor network to divide the voltage down to something that wouldn't overload/damage my scope.


This is the way our normal 115vac right from the wall appears or when the UPS is plugged into the wall. As you can see, it's not a perfect sinusoid (could be to the various loads on it) but it's pretty close to a pure sinusoid:








This is the output of the UPS when unplugged from the wall and using it's internal battery-powered inverter:









I also had an old Tripplite inverter in my pickup truck so I hooked it up to a 13vdc power supply and looked at it as well:

As you can see these inverters are nowhere near sinusoidal output. They call them modified sine but the approximation is really, really poor. I was expecting (or hoping to see) small stair-step approximations but this is only slightly better than a square wave. The only thing that makes this slightly better than a square wave is the duty cycle (amount of time waveform is not zero per cycle) is a lot lower than 100%. It is no wonder that the ac motors don't like to run well off of these inverters. I understand from talking to APC tech support that they make a SmartUPS series that has sinusoidal output (but is also pricier). I think I can do better by getting my own sinusoid UPS and battery combination; I'll report back when I get around to this... But in the mean time, I'd steer clear of general run-of-the-mill UPS types for aquarium back-up power as it's likely to be a big waste of money and time.


Carl
Although I have never "scoped" mine (I do not have one, but my Dad does, so maybe he can check his someday), I have not had issues with my Tripplite heavy duty modified sine wave inverter that is hard wired into my camper.
I did have problems with a cheapie one purchased at Walmart than connects to a cigarette lighter or clips to the batteries.

My understanding is that the modified sine wave is more like the square wave when scoped than the true sine wave.

Please see this article from an online RV Blog I subscribe to:
RV Doctor – Should I Install an Inverter in My RV?

Likely there are differences in quality, as I had similar issues with the cheapie Walmart inverter of similar cost to yours.
Where as the heavy duty hard wired unit cost me $400 (although this price has come down since I purchased as most electronics have)


parker002
I take it nothing running in the aquarium requires any power conditioning? I have two UPS' but they're for computers. I could run my aquarium off of one of them, but I wouldn't think of running my servers off of a homemade DC inverter.


fishfever
Power conditioning is always better than none and is easily provided using an inexpensive passive wall strip to distribute power when there is no power outage. The sine wave output inverters we are talking about are not homemade; however they are more expensive than the cheaper modified sine versions (which is really just a square wave with the output suppressed to 0 for a period as it crosses through 0).

You may find that if your UPS is not sine output (most aren't) your motorized things like pumps will either run for a lot lower time than their power consumption would indicate or not run at all (about half the stuff I tried did not run). For computers and other electronics with switching power supplies the waveform type is not critical. A switching power supply rectifies and filters (converts AC to high-voltage DC) before chopping this DC at a much higher frequency than the 60 Hz line input. So it really matters little whether the waveform starts off as a sinusoid or not since it's immediately converted to DC. But many/most AC motors are not going to run right and/or run inefficiently with the non-sinusoidal power.


fishfever
Well I've made good progress on the new UPS testing and have learned/discovered an awful lot... a lot more than I can post here without getting into rather technical stuff and boring everyone to death. LOL Anyway this new UPS I bought charges up much faster (about 3.5 hrs from fully drained to fully charged) so I was able to run a number of tests in a fairly short while. One thing I discovered is this is more than an open box UPS - it has probably been in use for about a year. I found this out when I downloaded some software from APC's web site so I can monitor and tweak some settings (for example I don't want to hear it beep every 30 secs when power is lost).

There are tons of things you can monitor and/or change (like internal temp, battery voltage, etc.). One parameter was the battery installation date which was Nov 2008. Another giveaway that this is used UPS is some sticky residue on the case where someone pulled a label off the case. This is probably where someone put the user name/password needed to log into the UPS because this is also where I put my label with my created user name/password. ;-)

So a used UPS means the battery has some degradation from age and use. Batteries have a shelf/usage life and generally age faster at warmer temps. This UPS does not have an internal fan since it's a low-end model (the smallest pure sine UPS that APC makes). The monitor screen showed the temp at 36.9 C after it had been on a while so I clipped a tiny CPU fan (maybe 1-watt) over the ventilation holes and the internal temp dropped to 24.3 C (about room temp) after a while. This should extend the battery life going forward but whatever degradation has occurred in the past can't be reversed. The battery terminal voltage also increased from 24.57 to 26.59 volts after the unit cooled down (good sign).

I ran a test with the same 40w incandescent lamp I used with the previous (modified sine) UPS. I also tested it with all my filters/pumps. Every one of them runs just fine (not surprisingly) with the pure sinusoidal power. Here are the results:

TEST, 750VA/500W Pure Sine UPS
-------------------------

1. 40-watt lamp - ran 93 minutes (this is somewhat low compared to the APC run time calculator so this can probably be associated with not having a brand new set of batteries installed, APC estimates 126 minutes for a 40-watt load)
2. no-name small air pump that came with 3g hex tank - runs fine, did not time run length by itself
3. TopFin (Petsmart brand) 60g air pump - runs fine, did not time run length by itself
4. Whisper 10 air pump with 2.2-watt motor - ran normally for 4 hours, 17 mins
5. Aqueon 10 HOB - runs fine, did not time run length
6. Aqueon 30 HOB - runs fine, did not time run length
7. Via Aqua Vitalife 200 HOB with 8-watt motor - ran normally for 3 hours, 48 mins
8. Ran tests #2, #3, and #7 simultaneously - all ran normally for 2 hours, 40 mins
9. Nothing connected to UPS - shut down after 3 hours, 44 mins


So the revelation here is that the UPS itself with NO LOAD has a finite run time similar to a small load. Why is this? The reason is the UPS inverter itself consumes power EVEN IF THERE IS NO LOAD. This is why if you compare tests #4, #7, and #9 you will see it doesn't seem to matter if a filter/pump is connected or not - the limiting factor in the run time is UPS itself! (It is a little odd that the Whisper 10 actually increased run time over nothing connected at all but there is probably a good bit of variance repeating runs, I didn't bother to explore this.) Only when I connected several filters/pumps as in test #8 did the run time drop substantially from the rest of them (in other words, now the load is big enough to have a bearing on the run time).

Here is the efficiency curve of my UPS from APC's web site (copyright APC):

UPS Efficiency

This confirms that the efficiency (effective power transferred to load) drops substantially when using it at the low end of the load range. In fact with a small 8-watt filter like the Via Aqua, we are only at 1.6% loading (8/500) which is actually off the chart (well, well below 50% efficiency if you examine the slope of the curve which intersects with zero efficiency at 0% load). So this explains why it almost doesn't matter whether the filter is plugged in; almost all the power is being consumed within the UPS!

Although I'm satisfied for now with nearly quadruple the run-time with this new UPS (I think it would be closer to 5 hours or more with new batteries), this is still not the optimum solution for emergency run power. One could go to a higher VA/watt UPS as long as the batteries were actually higher amp-hours and not just a larger inverter (and hope the UPS low-end efficiency does not consume too much of the battery). But a better (and much smaller, less expensive) solution for a single tank or small number of tanks would be a smaller inverter that was sized just big enough to run the emergency filter/pump.

This would shift the whole efficiency curve to the left AND allow more reasonable sized batteries (i.e. smaller) to run your filter/pump. Unfortunately I don't know of anything like this out there because most who would buy a UPS are trying to run high-power equipment so this would be kind of a niche product. The closest thing is probably the SunSun UPS NonStop Air Pump that American Aquarium Products sells but it would be nice if instead of the dedicated air pump they made one with just an outlet where you could plug your own small low-power filter or pump into (limited to maybe 10-15 watts). For a large number of tanks with lots of filters/pumps the larger inverter/custom battery solution is a no-brainer.

I'm of the opinion that the sine versus modified sine is primarily a factor for whether your motor runs or not based on the data I collected. It is much less a factor for determining how well the motor runs (i.e. length of run time, motor efficiency). So you can probably get away with a cheaper UPS (modified sine) if you are sure your filter/pump can run off a modified sine or are able to test your filter/pump on a particular UPS before purchasing it.

*** Several Skipped Posts***

fishfever
I have finished putting together my emergency power unit and installed it in the tank today. I have run many cycles and quite satisfied that everything is working as it should with full unattended backup power and recovery should we lose power while away on a trip or vacation. I did try the very simple battery/charger/inverter wiring connection but there some big drawbacks to this scheme. First the battery charger (which needs to be a "smart" or "staged" charger to avoid overcharging the lead-acid battery) is fooled when the inverter/load is also connected. The charge time goes up greatly and the charger doesn't behave well (hard to explain without getting real technical) which makes me believe it could shorten battery life by never going to a "float" charging mode. In addition, the inverter will not restart when the power is restored until it is reset by disconnecting the battery (which is not completely dead but measures around 10.5 volts).

So this would mean somebody would have to be present when power is restored to reset or else the filter wouldn't run even with normal power available. I was only planning to use this wiring scheme temporarily to check inverter reliability by running 24/7 for a while but after seeing these issues I went straight to my final scheme which has automatic switching circuits to overcome these problems. I have run many cycles and haven't noticed any issues. I'll probably continue running cycles for awhile more although I'm starting to get sick of testing! The switchovers (both from normal to emergency and emergency back to normal) occur almost instantaneously and I notice no break at all in the water flow from my Vitalife 200 HOB. I ordered a 7AH battery to replace the one I had been borrowing from my UPS but the battery company sent me a 9AH instead. It is physically the same size but weighs about 1.5 lbs heavier. With this battery I'm getting over 16 hours run time on the Vitalife 200 and probably over 60 hours with the small Whisper 10 air pump (but I have not bothered to verify the 60+ number).

Here is a picture of the backup unit ready to install:

The total weight is 9 lbs which is significantly lighter than the 30 lb UPS that it is replacing. I bought the base from a local Goodwill store for a whole $2! Since I had most of the small electrical items in my junk box, I probably paid about $75 for everything. I imagine with careful shopping one could put it together for about $100 from scratch not including labor cost of course. (Note: I was trying more for functionality than looks. LOL)

And here it is installed (vertically to save space):


It only uses about a 5" x 5" corner space so I have regained a good bit of tank stand floor space by replacing the bulky inverter (and more than quadrupled the emergency run time as well). My next project will be to clean up the ugly cable mess by mounting the power strip up high off the floor which will give me even more space for food, maintenance items, etc. I also plan on adding an audible alarm so that in the unlikely event we get a power interruption of more than 16 hours I can replace the battery with a fresh one (probably a 7AH from one of my spare UPS's).

To read the complete thread/blog (this is an ongoing blog): UPS Test

For another resource of related interest:
Aquarium Lighting; the most researched and updated article on the subject of aquarium lighting (including LED Aquarium Lights) found anywhere on the internet!

Or for a popular/useful article about Aquarium Silicone Use, Type, Repair, Glass Hinges, & much more information