GFCI & heater problems

Status
Not open for further replies.
Riograndeelectric said:
Marc, how do you measure the leakage current ? I hate to sound stuipd but I have always gone out checked the gfci / replaced it and told the customer it is a problem with the equipment pluged into the GFCI
Click to expand...

You take a megger and turn it all the way up until you melt the GFCI. Then replace the GFCI if the problem is still there then the heater is the problem. :D
 
I tell customers all the time that if I did not provide something, I am not responsible for it... I didn't provide the heaters, and infact, they were not even purchased when the electrical work was done.
Click to expand...
Well that says it all ! :smile: I need to remember that Line...
 
Riograndeelectric said:
Marc, how do you measure the leakage current?
Click to expand...
One of several ways. If you have a sensitive enough Amprobe, you can do sort of a zero sequence test by snapping around the hot and neutral together only and see what you read in milliamperes. That number is the measured leakage current. You can also snap around just the EGC to measure the flow on that, but the leakage may not necessarily be just onto the EGC. Fluke has a pretty neat, cheap, gizmo to do this. I think it's the Fluke 360. There are also dedicated appliance leakage meters that you just plug the appliance's attachment cord into and read the screen. You can also do a megger check, read the megohms, them do the math to get the leakage current.

Riograndeelectric said:
Ihate to sound stuipd but I have lawas gone out checked the gfci / replced it and told the customer it is a probal with the equipment pluged into the GFCI
Click to expand...
I think that is probably how most people approach it, but wouldn't it be nice to know for sure rather than wonder and guess? I hate having to tell the customer what I think the problem "probably" is. I want to tell them with a firm degree of surety.
 
measuring leakage current

measuring leakage current

If you're a little squeamish about popping the internal 200 mA fuse in a DVM's current measuring circuitry, put a 1K or smaller, +/- 5% resistor in series with the ground lead and measure the v drop.
5 mA through 1K = 5 v across.
 
Lxnxjxhx said:
If you're a little squeamish about popping the internal 200 mA fuse in a DVM's current measuring circuitry, put a 1K or smaller, +/- 5% resistor in series with the ground lead and measure the v drop.
5 mA through 1K = 5 v across.
Click to expand...
Measuring the current on the EGC wouldn't be the most useful measurement to take, as I already pointed out. The current might be, and likely is, going other places too. The zero-sequence type measurement of the line and neutral together would be more useful.
 
zero-sequence type measurement

zero-sequence type measurement

Please 'splain' or show diagram.
:)
 
Lxnxjxhx said:
Please 'splain' or show diagram.
Click to expand...
Sounds like a fool's errand. Why would I want to do that for a guy who's not even an electrician, and will never need to perform such an operation? Must you be spoon-fed everything you've ever wondered about? Here, I'm feeling generous anyhow, so just read this. Pay particular attention to "unintentional paths to ground", which is some of what happens with a pond heater.

http://www.newarkinone.thinkhost.co...and_solutions/Leakage_current_measurement.pdf
 
Last edited:
080518-2049 EST

Greg:

One means to test your GFCI is a resistor from the hot output to the ground pin.

National Semiconductor has a plot of their typical trip time curve for this function and the UL specification. See
http://cache.national.com/ds/LM/LM1851.pdf
Notice the trip-time curve on p4, and a typical circuit figure 2 on p6. The UL specification is rather long at 5 MA, about 5 seconds, and to get 1/2 cycle operation it is about 200 to 300 MA.

A 5 MA sensitivity is mentioned. So I would experiment with 10 MA and 6 MA to check that trip occurs. This would correspond to 120/0.01 = 12 k ohms, and 20 k ohms. Use two 1 W Ohmite carbon composition or metal film resistors in series to make your resistor. Also create a 3 MA resistor 120/.003 = 40 k ohms for a no trip test. Use your judgment on reisistor tolerance. You can measure the leakage current of the test resistor to verify you have a suitable test element. Obviously line voltage influences the current.

Note the long allowed trip time on low currents of the UL curve.

To measure leakage make a test fixture with a 1000 ohms 2 W wirewound resistor in series with the ground lead (EGC). So this is a plug, resistor, and socket. You connect you Fluke or equivalent meter in AC volts across the 1000 ohm resistor and monitor the voltage across it. The calibriation is 1 V per millampere. Obviously hot and neutral are wired straight thru. This technique is to prevent burning out the current fuse in you meter. In a Fluke the cost is close to $20 to replace the fuses.

I agree with Marc that on a service call one should identify the problem if it is within your capability. My field is quite different than yours and I work on solving whatever the problem may be. It could be hydraulic, mechanical, electrical, programmable controller, logic, electronic, parts, lubrication, metallurgy, etc.

Our LVDTs have a carbide ball tip and at most 1 # of applied force. In almost every application this will contact another device and not the part being gaged. Usually the designer of that device will use a tool steel hardened to Rockwell 60 assuming this won't be affected at this low force. But today I know from experience that a 100,000 to 300,000 cycles will cause the steel to become indented by maybe 0.001" even though this is thru hardened and about as hard as you can typically harden steel.

When we are trying for repeatability of 0.000,1" this indent can cause random errors of significance. This is somebody else's responsibility because they designed that part of the system. But when I first encountered this problem I found the problem and told the customer how to solve the problem. I did not just pull the LVDT out and demonstrate to the customer that my equipment was OK.

Usually my equipment does not have a problem and I am troubleshooting some place else in the system for the customer. One recently was a ring gear multiple that was stripping bolts. This had none of my equipment. The problem was in the offset gear box between the nut runner transducer and the bolt being driven.

.
 
mdshunk said:
Sounds like a fool's errand. Why would I want to do that for a guy who's not even an electrician, and will never need to perform such an operation? Must you be spoon-fed everything you've ever wondered about? Here, I'm feeling generous anyhow, so just read this. Pay particular attention to "unintentional paths to ground", which is some of what happens with a pond heater.

http://www.newarkinone.thinkhost.co...and_solutions/Leakage_current_measurement.pdf
Click to expand...
How come I'm getting a "page not found" on that link?:confused:
 
cadpoint said:
Well that says it all ! :smile: I need to remember that Line...
Click to expand...

To be quite honest, I have been called back to fix things I have not installed... or even seen.

Then when I tell them there will be a trip charge, or a bill attached, people freak out.

Somehow, when everything except 1 item works in an outlet... it MUST be the outlet that is causing the problem.

Greg
 
080518-2213 ESt

Marc:

You are correct that the leakage can be somewhere else.

For Greg or others. How to get an inexpensive high sensitivity current probe. Maybe get an old GFCI and tap off of the internal current transfomer with your voltmeter if it can be accessed. If not wind your own small current transformer.

.
 
mdshunk said:
Measuring the current on the EGC wouldn't be the most useful measurement to take, as I already pointed out. The current might be, and likely is, going other places too. The zero-sequence type measurement of the line and neutral together would be more useful.
Click to expand...
If you find a current probe sensitive enough to measure the imbalance you speak of, the jaw opening will be very small though. I have a 0-450ma current probe but the jaw opening is only 3/16", so you can't even clamp two #16AWG wires. One could wrap a piece of wire around the line/neutral and back through the clamp but for every loop doubles the reading.
 
mdshunk said:
One of several ways. If you have a sensitive enough Amprobe, you can do sort of a zero sequence test by snapping around the hot and neutral together only and see what you read in milliamperes. That number is the measured leakage current. You can also snap around just the EGC to measure the flow on that, but the leakage may not necessarily be just onto the EGC. Fluke has a pretty neat, cheap, gizmo to do this. I think it's the Fluke 360. There are also dedicated appliance leakage meters that you just plug the appliance's attachment cord into and read the screen. You can also do a megger check, read the megohms, them do the math to get the leakage current.

I think that is probably how most people approach it, but wouldn't it be nice to know for sure rather than wonder and guess? I hate having to tell the customer what I think the problem "probably" is. I want to tell them with a firm degree of surety.
Click to expand...
Nice call on the meter Marc I have never seen this model It will do exactly what you say and tell you where the leakage is thanks from the cow.
 
quogueelectric said:
I got the link it is a pdf file maybe you dont have adobe acrobat reader to open it you can get it for free off the net.
Click to expand...
Nope that's not problem as I have no problem with PDFs. It can't find the page with the file.

Forget that it worked but I've had that file saved long time ago!
 
Status
Not open for further replies.
Top