7

Every once in a while I have a breaker trip when I'm not home.

I have a series of computers running cpu intensive programs and sometimes alt coins which can also be hogs :-). At times there is a huge surge in power consumption and the breakers trip.

Right now the system goes from

  1. the computer's power supply unit to a

  2. wemo which in turn is connected to an

  3. outlet which goes directly to the circuit breaker.

A Belkin Wemo is a device that can be accessed remotely and can be used to turn an outlet on or off. My wemos are connected to a laptop which allows me to turn on devices remotely.

Every once in a while the circuit breaker trips and I have to wait until I get home to flip things back on. Usually this doesn't matter but if I'm away for a few days it does.

I'm looking for a solution.

I'm thinking about a "something" that would be tripped if there is a surge and I can use a wemo to flip back on. The wemo would be on another circuit.

Computer (PSU) -----> "something" 
                           |
                           |
                          wemo

If there is a surge the "something" trips (not the circuit breaker). I'll have a wemo pinging this "something." If there is no response for x seconds then the wemo shuts it off and turns it back on, effectively resetting the "something."

  • 1
    Actually the problem might be very well suited for the IT side of SE (like serverfault) because this exact problem is encountered in server rooms all the time. The circuit is oversubscribed (there is more potential for load than the breaker can tolerate, but not very often) and he wants to prevent the upstream breaker from opening, which means he needs to watch the current and open something closer to the device, preferably something that can be reclosed automatically. It just so happens devices like that exist (this is not a product endorsement): cyberswitching.com/about – Jeff Meden Apr 14 '16 at 18:35
  • 1
    You should be aware of the 80% circuit breaker derating rule. On 15A circuit, you should only draw 12A continuously. – 200_success Apr 14 '16 at 19:24
  • 2
    Its important to remember that when the nameplate says 450W, that means its capable of supplying 450W to the system before something happens (internal fuse or CB trips? overheats and efficiency goes down? maybe nothing?) but even if its supplying 450W to the system per spec, there is conversion loss so the input is 450W + (1-efficiency factor)*450. So, say its 90% efficient at 450W, its actually drawing 450W plus the losses of (1-.9)*450 or 45w, so 495W total. This could account for your missing capacity. Power factor is a question too (the cb opens to var/volts not from w/volts). – Jeff Meden Apr 14 '16 at 20:10
  • 2
    Again, all things that the nerds like me at Serverfault or Superuser would be eager to discuss. – Jeff Meden Apr 14 '16 at 20:13
  • 1
    We should make a StackExchange called "Burning Down the House" or "Death Hacks" for topics like these trying to reset/bypass human intervention and safety mechanisms/checks. I understand what the OP is asking, but I ask back: why are you not trying to make the load more efficient and reduce watts at the PC level? You are basically trying to run commercial load in a residential line atmosphere. FYI microwaves are required to run a dedicated branch circuit, and most of those are less than 1800 watts. So why wouldn't you be required as well to do the same? (ie, dedicated circuits for each pair). – dhaupin Apr 15 '16 at 15:51
18

Sounds like there are two options.

Option one is to replace the existing breaker because you think its defective. This option is good if you know that you're not using enough power to trip the breaker so something else must be going on. However, it doesn't sound like this is the case.

Option two is to run a new dedicated circuit to the room to handle the extra load you need to run. If you're using too much power, this is the only right option. The breaker should never trip, so you shouldn't need a remote way to reset it.

  • I have 4 450W PSUs on one 20 amp breaker. My understanding is that 20 amp breakers can hold 20x120 watts at max (2400) and 20% less for continuous service (1920). I'm using 1800 so I thought I was good to go. Guess I was wrong – Mayo Apr 14 '16 at 15:36
  • 1
    Is that really the only thing you have plugged in? getting some kind of plug-in watt meter would also be good to see what you're really using. I wouldn't expect a 450W PSU to consume a constant 450W, so there could be something else spiking occasionally. Also, its early for the question. You'll get more answers. – JPhi1618 Apr 14 '16 at 15:41
  • Thx. Will take a look to make certain. – Mayo Apr 14 '16 at 15:43
  • 7
    @Mayo, 450W is the output rating of a PSU. Even with the higher 80Plus certifications, you're probably only getting 85%-90% conversion efficiency at full load, so your notional "1800 watts" is an actual power draw of 2000W-2100W. – Mark Apr 14 '16 at 19:57
  • 1
    Definitely spreading the load out over 2 or more breakers is the way to go. If you are popping the breaker you are risking a fire. You might want to check the temperature of the current breaker to see if it is getting hotter than the other breakers. Heat leads to failure. – Arluin Apr 14 '16 at 22:11
24

If your device is drawing too much power, and is causing the breaker to trip. It means the device is too large for the circuit, and the breaker is doing the job it was designed to do. Making an "automatic breaker turner on-er" is not likely the best solution.

There's two possibilities for why you'd trip the thermal protection of a circuit breaker. If the breaker is faulty, it could be tripping at a lower current than it is designed to trip at. In this case, the solution is to replace the breaker. If the breaker is not faulty, it means that your device is drawing a sufficient amount of current, for long enough to trip the breaker. If this is the case, then your circuit is likely undersized.

If your device is tripping the magnetic protection of the circuit breaker, there's a larger problem with the circuit and/or device. The magnetic protection of a circuit breaker should not kick in, unless the circuit is drawing many times the rated current (think short-circuit). If this is the case, you likely have a faulty device and/or wiring.

If your device is tripping the GFCI or AFCI protection of the breaker, then you'll have to investigate that further. If the device is connected to a surge protector, the normal functioning of the surge suppressor may cause a GFCI device to trip.

You'll likely want to start by monitoring the system to determine the normal current draw, as well as the peak current draw. From there, you can decide how to proceed. Without knowing the actual current draw, or why the breaker is tripping. There's not much you can do to remedy the situation.

It's not likely that an "automatic breaker turner on-er" exists, or ever would exist. What if there was a major problem causing the breaker to trip, and you continually remotely reset it? You're not there, so you can't see that a couple wires came loose and are laying in a pool of gasoline. You keep resetting the breaker, which causes an arc between the gasoline soaked wires. You'll probably continue to try and reset the breaker, as your house burns to the ground. </Dramatization>


It's not clear to me whether or not this is a single device, or multiple devices. Dividing the load across multiple individual 20 ampere circuits, might be a solution. This is much more straight forward if you're powering multiple devices, though is not impossible if it's a single device.


NOTES:

Reclosers

While reclosers do exist, they are only used on the distribution side of the electrical system. They are not available for residential systems, for the reasons mentioned above.

Circuit breaker thermal protection

The thermal protection device within a circuit breaker is quite basic, yet also quite effective. It simply opens the circuit when it gets too hot, which is hopefully before the branch circuit wires are burning. It makes no difference what type of load is connected, if the wires are getting too hot, the thermal protection opens the circuit.

This is the reason why conductor size matters. If you have 12 AWG copper conductors connected to a 20 ampere breaker, the thermal protection will open before the wires are burning (as long as the breaker is functioning properly). If you had smaller wire connected to the 20 ampere breaker, there's a possibility that the wire could be on fire before the thermal protection kicks in.

  • 2
    thank you. I appreciate your time and effort It's "obvious" now that whatever solution I come up with circumventing the circuit breaker is not the answer. – Mayo Apr 14 '16 at 17:57
5

EE here. PSUs have a non-linear relationship between current draw and voltage. Your calculations assume that the PSUs are purely resistive loads, and they are not. Assuming the circuit is not also oversubscribed, the circuit breaker can be fooled in this situation.

A circuit breaker exists to protect the wiring from overheating and starting a fire. It is the real power dissipated that causes overheating, not the temporary peak current from active loads. Because of this, you can consider installing a residential high magnetic circuit breaker which are available for most panels. These can tolerate high peak current. However, you may want to have a licensed electrician examine the situation and determine if there is a wiring problem. A circuit breaker that repeatedly trips is quite a nuisance, but is not always a false alarm. It could be saving your house from an electrical fire every time it trips.

Definitely don't mess around with anything that tries to automatically reset a circuit breaker. That will void your house insurance.

  • Thank you. Was completely unaware of non-linear relationship between draw and voltage. Will read up on it. – Mayo Apr 14 '16 at 20:00
  • 1
    Agreed, switching power supplies are non-linear loads and when enough of them are on a circuit they can wreak havoc. As the OP has discovered. – ArchonOSX Apr 14 '16 at 21:07
  • A good PSU should have automatic power factor correction, automatically making it act very close to a pure load. – JDługosz Apr 15 '16 at 8:53
  • @JDługosz Active PFCs helps a lot, but based on OP's description and context it sounds like there are some older PSUs in the mix. Even if PFC is used, many PSUs may have a rated PF of 0.9 to 0.95, meaning that there's still some non-negligible reactive current when multiplied by several devices. – jbarlow Apr 15 '16 at 20:25
4

Disregard the arm-waving guesses at how much power the PC "might" draw, and buy a Kill-a-Watt. Bonus points if you get a better type of power monitor that can data-log.

The Kill-a-Watt can also tell you a lot about power factor. Low power factor means it is causing interesting side-effects that masquerade as drawing more current than you actually are on average. Having high power factor of 99% is not a problem with a little bit of extra engineering, but it is not a priority for all builders, least of all the cheap ones.

With hard data, you can troubleshoot exactly when a PC is drawing what power - and if that's where the problem lies, you can iterate on changing the PCs' draws or improving their power factor.

For instance, this could be as simple as upgrading power supplies, or flipping a BIOS setting that tells it to spin up hard drives one at a time, instead of all at once.

  • The IT professional version of the Kill-a-Watt would be a metered PDU. Some of those would have logging and remote control capabilities. – 200_success Apr 14 '16 at 22:52
  • That would be neat, but I have a feeling that budget is a factor here. – Harper - Reinstate Monica Apr 15 '16 at 4:28
  • Yea, pick up a plug-in device in the $20 range that includes power factor and separate Watt and AVR readings. You'll use it again, chart all your appliances, and will be glad to have it in the junk drawer. – JDługosz Apr 15 '16 at 8:57
  • Ebay has metered PDUs dirt cheap (example, 20/16A for $50 with full power monitoring); for home/hobby use its a great way to get a device that does exactly what you need. You just need to do some homework on what specs to look for, and how to monitor it once you have it in place. – Jeff Meden Apr 15 '16 at 13:11
3

Based on some details that were revealed in comments, I think I should post this as an answer.

Your comment said:

Using 4 PSUs (450 each) on each 20 A circuit. 80% is 1920 and I'm using 1800. I thought I had room to spare. Apparently not.


There is an 80% circuit breaker derating rule: you are only supposed to draw 80% of the breaker's rated current continuously. This rule applies to any breaker unless it is specially marked.

Based on that rule, a 20A circuit is only supposed to carry 16A continuously.

A "450W power supply" for a computer means that it is able to provide 450W total to your computer's DC components. Assuming, slightly pessimistically, 80% efficiency, it might actually use 560W, or 5.1A at 110V. Four of those power supplies could max out at 20.4A.

Your estimation is too optimistic. 20.4A is much greater than 16A. It's no surprise that you are tripping your breaker. You need to either upgrade your circuit or find somewhere else to plug in your computer. Repeatedly resetting the circuit breaker manually, let alone automatically, is defeating an important safety mechanism.

  • 1
    A) We use 120V when figuring ampacity and such. 120V is nominal and typical voltage in North America. .... B) A computer power supply IS NOT a continuous load. ........ Continuous Load. A load where the maximum current is expected to continue for 3 hours or more – Speedy Petey Apr 14 '16 at 20:01
  • 1
    @SpeedyPetey Bitcoin mining could easily be a continuous load. In any case, these are estimates. I think it would be worthwhile for Mayo to get a metered PDU — basically a power strip with a digital current readout — to keep an eye on the situation. – 200_success Apr 14 '16 at 20:25
  • 1
    I would have to agree with 200. Bitcoin mining would max out the computer for as long as the task is assigned. 3 hours or more would constitute a continuous load and therefore the circuit should only be loaded to 80%. If the PSUs are not drawing more than 450 each the circuit should be able to handle 15 amps. However switching power supplies might be causing the problem. – ArchonOSX Apr 14 '16 at 21:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.