Our mission

GFCI Reminder (and current monitors, and redundancy)

jccaclimber

Supporting Member
I'm still here in SF with my fish in DFW, TX. Yesterday my wife was kind enough to notice the 50 gallon that I use as an ATO had run low. Our internet was out, so I hadn't gotten a notification. She refilled, and apparently spilled some water. An hour later the puddle had migrated to the water alarm, but when she went to look it was a thin film, more like condensation than a puddle.
Fast forward a couple hours and the GFCI trips. Half the lights go out and take the return pump and skimmer with them. The other half are on an independent circuit. Some trouble shooting later we determine that unplugging the skimmer allows the GFCI to be reset, and it trips again when plugged in. Unfortunate as I won't be home for 2 weeks and will be out $300 for a new skimmer pump, but things will live.
My skimmer is a long way away from the EB8, which is mounted above the tank (where all unsealed electronics should strive to be, ie out of the drip, overflow, and splash paths). As a result there is an extension cord going from the EB8 to the skimmer, and I foolishly did not mount that well. It had fallen to the floor and as you might have guessed by now, the creeping puddle of water picked up some salt and wicked in to the extension cord connection. Apparently the GFCI trips faster than the breaker (this is not a surprise), because it tripped first.
This afternoon I had her find that connection and clean it with fresh water, dry, and reassemble. Sure enough the skimmer is no longer tripping the GFCI.
Once the internet came back up I pulled the EB8 power data and found this:

83725586-1211-4F4B-83C3-EE83941A2892.png


I suspect some interference because any of these should have tripped either the breaker, or eventually the one on the EB8.

Observations:
1) The GFCI allowed me to detect a hazardous situation before it resulted in a hazard to human life by electric shock or fire.
2) To emphasize it, this may have prevented a fire. I've definitely seen high resistance shorts that resulted in burnt or melted plugs without actually drawing enough current to trip a breaker. If you put 230 W across a plug it will absolutely melt it, but you're still only drawing 2 A, which isn't nearly enough to trip a 15 A or 20 A breaker.
3) A reminder that non-waterproof electrical connections should be in an area that is not at risk of water exposure, including salt mist.
4) Had there not been a GFCI this should have tripped either the EB8, or the main breaker. Having the GFCI there is nice as you don't have to walk to the garage to flip the breaker.
5) In the event that I had a partial short, but not enough to trip a breaker over-current monitoring could detect it, and shut off the outlet as well as send an alarm. Once I'm up and running post move with the new Apex and an EB832 I should put over-current shutoffs and alarms in addition to under current ones.
5) Backups are good. I'm already out of town for several weeks. Had she been gone as well I wouldn't have known due to the internet outage. It turned out something went wrong with the coax connection outside the house, so it didn't come back on until a new cable was run from the curb, which would not have happened without someone being there. I would have assumed it was just an internet issue, so it would have been a day or two before one of my fish sitters checked in. I run redundant circuits, so I would still have had half of my lights, as well as circulation pumps in the display and sump (also frag system), and some of the heaters. None of the second circuit heaters are in the sump, I should fix that. I also will be putting a redundant return pump on the next setup.
6) Backups are still good. Had this been a single circuit system I would still have gotten a half day of circulation from the backup batteries on the circulation pumps.
 
Last edited:
That’s awesome she was able to troubleshoot that for you! I don’t think I could get my wife to do that. Maybe buy a new extension cord at best...
 
She resisted due to being involved with the fish tank, but I’m sure her EE degrees helped. To her credit she has been very helpful while I’m away, must be a keeper.
 
Jccaclimber,

By “over and under current protection”, do u mean an apex outlet can be coded to shutoff based on current drawn? What will the codes be like?
 
Jccaclimber,

By “over and under current protection”, do u mean an apex outlet can be coded to shutoff based on current drawn? What will the codes be like?

You can use this syntax ONLY if you have the new Apex or Apex EL with an EB832, will not work on EB8 or EB4s

If Output [name of output] Watts [or Amps] > [or <] xxx Then OFF [or ON]
 
^What he said. I can't use it on the EB8 on my old Apex, but can use it on the EB832 with my new Apex. I haven't checked, but I wonder if there is a way to do feedback based on current of the entire strip with an EB8?
 
The older EB8s don't measure individual current draw per outlet, but I thought they measured the total EB8 current draw.

That said, would be nice if there was a fuse or something for these outlets, 10 cent fuse could save very expensive piece of plastic from becoming non functional.
 
They do measure total draw, in just not sure if you can control by it. They have a 15 A breaker on the end that should pop out, but I’m sure there is some delay. My guess is that the GFCI got there first, although I can’t explain the other spikes.
 
Well hard to tell from the graph how long those surges were, presumably they were spikes as it's been mentioned that the software likes to "smooth lines". And is entirely possibly that the readings themselves are false, sure it may have spiked a bunch but 43amp direct to ground type of spike? Yeah kind of sketchy, I mean sometimes you hook up a multimeter and it initially reads some very high (or low) reading as it takes a second to adjust, of course not knowing exactly how it's reading current couldn't say.
 
I agree they were likely spikes and may be incorrect all together. The thing is that I only expect two spikes. One where it initially popped, and another some time later when she was working to identify the fault for 10-15 minutes.
 
what is "some time later" it looks like a 2 hour window between first spike and last (unless I'm reading that graph wrong).
Also to note, that if there's water that creates a short from hot to neutral it won't trip a GFCI since all the current is accounted for, if it moves from hot to ground then yeah it'll trip, so if the path of the water somehow "shifted" it could explain some of the spikes.

Overall unless you do the Hangnail™ check, it's often hard to isolate your problem if it's leaky current (Cracked housing, insulation, etc)
 
I'm not sure as I wasn't there, but I got the impression it was at least an hour during which the current should have read 0 as the GFCI was popped, although it certainly doesn't look that way in the data.
Agreed that only a hot to ground short would trip the GFCI. In order to know if an intermittent hot to neutral would trip the breaker I would need to know more about how fast the Apex can detect a peak vs. the thermal overload characteristics of the house breaker and also the overload characteristics in the EB8. It is an interesting point that the EB8 outlets are only rated to 5 to 10 amps (5 in this case), but as far as I know the only over current protection built in is the 15 A for the bar as a whole.
 
Funny that you mention the hangnail check, I know exactly what you mean. The "my hair is wet and occasionally brushing your improperly grounded DIY LED heatsink" check is about as much fun, except you get to feel it in your head. Took us a little bit to track that one down. Note, not my tank and not my improperly grounded DIY LED.
 
Back
Top