7
\$\begingroup\$

I'm trying to design a power backup system for my saltwater aquarium. In short, the system will be powered by a 12VDC power supply while the mains power is on. On power failure, the system will automatically switch to a battery source until the mains comes back on. At this point, it will switch back to the 12VDC power supply. I believe I have all that worked out, no issues.

The only thing I'm stuck on is the 12VDC power supply that runs the system under normal conditions. If this power supply fails, the system will detect this as a power failure and switch to the battery bank. The batteries will eventually run out because they're meant for short term life support, not for indefinite running. At this point, it seems like the greatest weakness in the system is the 12VDC power supply that runs the system during normal operation.

So, my question: would it be safe to wire two identical 12VDC power supplies in parallel for redundancy purposes only? Note, I am NOT trying to gain additional capacity from this configuration, only redundancy. One power supply could very easily handle the entire load. The power supplies in question will be 12V 40A, and the consumption of the system tops out around 300W.

Thanks in advance for any help regarding this.

\$\endgroup\$
6
  • 4
    \$\begingroup\$ Normally you would diode-OR the two supplies together, not connect them directly in parallel. (Assuming you can handle the roughly 0.3V or so drop in a Schottky diode. If not, there are FET or-ing controllers you could investigate.) \$\endgroup\$
    – John D
    Apr 29, 2016 at 17:17
  • 1
    \$\begingroup\$ For 300W of 12V, a diode-OR configuration would waste a lot of power and require fairly substantial diodes. Depending on the electronics after the diode, the voltage drop from the diode(s) could also affect their operation. \$\endgroup\$
    – user2943160
    Apr 29, 2016 at 18:02
  • 1
    \$\begingroup\$ Why not a simple relay to switch to the back-up 12VDC power supply, in the event the main 12V is lost? I can suggest a SPDT. \$\endgroup\$
    – Tim Spriggs
    Apr 29, 2016 at 18:40
  • 1
    \$\begingroup\$ added to last comment: the relay N.O. would connect the neutral leg to the 2nd Power Supply in the event of loss of 12VDC. the N.C. contacts would be for the main DC Pos., with the coil being 12 VDC of course. \$\endgroup\$
    – Tim Spriggs
    Apr 29, 2016 at 18:50
  • 1
    \$\begingroup\$ added to last comment: I think the NO and NC would be opposite as from the above, but you get the idea. And this would be easy to wire up as long as you have the power supplies and other stuff. \$\endgroup\$
    – Tim Spriggs
    Apr 29, 2016 at 19:07

3 Answers 3

Reset to default
6
\$\begingroup\$

You should re-think your topology if reliability is really so important. Your most likely point of failure is the switchover relay, not the 12 V power supply.

I would set up the system to always run from the 12 V battery, just that the power supply charges the battery when line power is available. This is exactly how the 12 V system in your car works. It is always connected to the battery, but the battery is charged when the alternator is running.

Lead-acid batteries can be held at their "float charge" voltage indefinitely without damage. The alternator and regulator system in a car is usually set to 13.6 V. Since you'll generally be charging for long periods of time, you can be a little more conservative, like 13.0 - 13.2 volts.

To achieve this, get a "12 V" power supply that can be tweaked a little. Many can. Put a Schottky diode between the power supply output and the 12 V lead-acid battery, then adjust the power supply for the desired float charge voltage at the battery. The actual power supply voltage will be a little higher due to the diode. You can even add redundant power supplies, each with their own diode if you like.

You have to make sure that your 12 V equipment can handle the full range of voltage from the battery almost depleted to the battery full and being float-charged. 10-14 V would be a good target. 9-14 V even better.

This system can even support multiple power supplies and multiple batteries. Here is what it would look like:

This scheme allows for as many power supplies and batteries as you want. There is one diode per power supply per battery for charging. That's what D1-D6 are in this example. Then there is one diode per battery to the final load. That's what D7 and D8 are in this example.

\$\endgroup\$
9
  • 2
    \$\begingroup\$ He already said he has a batt. backup. So really , to use your car analogy, he wants a backup alternator, in case it fails, of course. \$\endgroup\$
    – Tim Spriggs
    Apr 29, 2016 at 22:13
  • \$\begingroup\$ You want to run three power supplies at the same time? And no LEDs? \$\endgroup\$
    – Tim Spriggs
    Apr 29, 2016 at 23:02
  • \$\begingroup\$ @Tim: As I said, I'm showing a overall topology that scales to multiple batteries and multiple power supplies. All switchover is automatic and without any mechanical parts. The didn't say exactly how his switchover system works, but he might be using a relay, which I'm advising against. \$\endgroup\$
    – Olin Lathrop
    Apr 30, 2016 at 11:57
  • \$\begingroup\$ What do you have against relays? My AC fan relay is over 40 years old and still works. \$\endgroup\$
    – Tim Spriggs
    Apr 30, 2016 at 13:04
  • 2
    \$\begingroup\$ @Tim: You don't seem to understand how this site works. What you see are answers to a question, not entries to a contest. If you feel so strongly that you know a better way, write your own answer. Even in your own terms, you haven't "entered" the "contest". Comments are never answers, just comments. They can't be downvoted and aren't peer-reviewed, like answers, so they're pretty meaningless. They can also be deleted eventually. \$\endgroup\$
    – Olin Lathrop
    Apr 30, 2016 at 13:48
5
\$\begingroup\$

schematic

simulate this circuit – Schematic created using CircuitLab

Figure 1. Parallel PSUs with power indicators. V+ will be about 0.7 V below the power supply output voltage due to diode drop. Both PSUs on all the time and, if similar type, will probably share the load 50/50.

schematic

simulate this circuit

Figure 2. Relay switched PSU. Both PSUs on all the time but one on standby.

schematic

simulate this circuit

Figure 3. DPDT relay switches mains on second power supply (preventing it from wasting power when PSU1 is running). A second contact switches the DC side.

Note my schematics show F1 as a common point of failure. The same is probably true of your home supply. You need to take that into account when designing the system.

\$\endgroup\$
4
  • \$\begingroup\$ If you want to waste power, do it that way. My solution above kicks in the reserve power supply only when 12VDC is lost. And I use fewer parts! so few, I didnt need a schematic to explain it. \$\endgroup\$
    – Tim Spriggs
    Apr 29, 2016 at 20:58
  • 3
    \$\begingroup\$ Not sure what the problem is, Tim. The LEDs? They're optional (but useful) and will consume about 1 kWh per year - cost about €0.10 - €0.20c depending on location. Apart from that the circuit uses either two diodes or one relay. Can't get any simpler. \$\endgroup\$
    – Transistor
    Apr 29, 2016 at 21:32
  • \$\begingroup\$ You said you have two power supplies running at the same time, on one of your designs. yes, the LEDs are pretty. And the common fuse could be an Achilles heel. \$\endgroup\$
    – Tim Spriggs
    Apr 29, 2016 at 22:16
  • \$\begingroup\$ I need a similar setup for a "poor man's SAN" at work. I'm powering extra spinning hard drives outside of a computer case. If the power supply dies on the RAIDed drives, I could lose data. I'd like to tie two together like Figure 3 above. My server room is on Battery backup, so that's not a concern. A faulty power supply is though. @transistor, do you happen to have part numbers for compatible relays above in Fig. 3? \$\endgroup\$
    – user208145
    May 20, 2016 at 8:15
2
\$\begingroup\$

No, it is generally not safe to parallel two power supplies (even of the same model) unless they explicitly support such a mode of operation. Some power supply chips (such as the LTM4625) are designed to be placed in parallel on the same circuit board. If configured correctly bench power supplies can be connected in parallel for load-sharing.

I can't quickly find a wall-powered supply that supports parallel operation. Such a supply would also require the ability to operate with its output being powered by the other supply when the wall power fails. You may need to adjust your specifications or simply have larger battery backup capacity to meet your requirements. Having an alarm when battery power is in use is standard for computer/IT UPSs, which would alert you if the 12V supply or the wall power fail.

\$\endgroup\$
2
  • \$\begingroup\$ Thanks, this is good background. The capacity is quite good, as I expect it will run my setup for 10 hours with plenty of wiggle room. My main concern with redundant power supplies is when I'm away for business. I work from home and am hardly away from home for more than 10 hours at a time. Further more, power failures in my area are relatively rare. However, a failed power supply could have the same effect as a multi-day power outage if I'm away, which the system is not designed to handle. I will consider the options more and do truly thank you for your input. \$\endgroup\$
    – user108567
    Apr 29, 2016 at 17:59
  • \$\begingroup\$ If possible, wire the two power supplies from different circuits. Would it make more sense to have two life-support-systems (air pump?) running on separate power supplies? \$\endgroup\$
    – Lenne
    Jul 8, 2020 at 11:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.