How to record AC main blinking or dropoffs -- ie how to record AC mains events

[ez24]

Hi

I have a problem with my AC electric.  I get short brownouts (dropouts ?) that just blink the lights.  Only one time did it knock out the radio, no effect on the TV.  So the time must be in millseconds.

I have not called the electric company yet because I do not want to get blamed that it is my house wiring (over 100 years old).  I have noticed that it happens on wet days and I think it is a loose connection at the pole.  The problem is access to the pole is very hard.  They will have to hike into a canyon and cut brush to get to it, so they are not going to be happy to help.

So I thought I would get a recording multimeter (via recommendations in a post) and came up with Owon B35T.  Nice and cheap meter.  While setting it up to record on Android, I found out the min recording period is .5 sec.  Then I realized if I want to record a millisecond event, the meter would have to record in milliseconds.  Thus the Owon will not work for me.

I want to correlated the dropouts to rain events.

Now I think that a specialized meter would be required OR maybe my Rigol DZ1054z?   If the Rigol can do it, can someone give me a hint on how to set it up to record millions of data points over days or if it will work or not?  I think a CVS can have only 64,000 data points (rows).  I am really spooked to hook it up to mains voltage because the last time I did that the probe exploded (different house).

Typically I will get from one to a dozen events  a day during the winter (I do not know what to call these blinks ?)

ps it is not the connection between the house and the drop wire.  That went out about 10 years ago during a storm and the electric company spliced the wires with a gas powered hydraulic press that fused the wires.

[technogeeky]

Hi

I have a problem with my AC electric.  I get short brownouts (dropouts ?) that just blink the lights.  Only one time did it knock out the radio, no effect on the TV.  So the time must be in millseconds.

I have not called the electric company yet because I do not want to get blamed that it is my house wiring (over 100 years old).  I have noticed that it happens on wet days and I think it is a loose connection at the pole.  The problem is access to the pole is very hard.  They will have to hike into a canyon and cut brush to get to it, so they are not going to be happy to help.

So I thought I would get a recording multimeter (via recommendations in a post) and came up with Owon B35T.  Nice and cheap meter.  While setting it up to record on Android, I found out the min recording period is .5 sec.  Then I realized if I want to record a millisecond event, the meter would have to record in milliseconds.  Thus the Owon will not work for me.

I want to correlated the dropouts to rain events.

Now I think that a specialized meter would be required OR maybe my Rigol DZ1054z?   If the Rigol can do it, can someone give me a hint on how to set it up to record millions of data points over days or if it will work or not?  I think a CVS can have only 64,000 data points (rows).  I am really spooked to hook it up to mains voltage because the last time I did that the probe exploded (different house).

Typically I will get from one to a dozen events  a day during the winter (I do not know what to call these blinks ?)

ps it is not the connection between the house and the drop wire.  That went out about 10 years ago during a storm and the electric company spliced the wires with a gas powered hydraulic press that fused the wires.

I did a bunch of A/C measurements last week (A/C pains, generator-produced A/C, inverter-produced A/C). I mostly did current (inrush and average), but I also did some voltage measurements.

I wanted to capture the time spent compressor-on time period for two air conditioner units, but I did not see a way without PC capture of data through the scope to do this, and I didn't have that prepared (I am on Linux at home, and OSX on the go -- so some tools aren't available for me anywhere, some not available everywhere).

I think it could be done with the rigol, and you'd have to write a tool to do a long-term capture (or more likely adapt an existing one in python or C or something).


Since you mentioned safety, I "guaranteed" my safety by floating the scope using a UPS on battery power. Better (though more expensive) solutions would include a differential probe (several hundreds of dollars), an isolation transformer (or or two hundreds of dollars).

People always say, ideally, you would isolate the target device, but since we're looking at mains, that is simply not possible.

In any case, this reminds me of an effective and still cheaper solution, if you *truly* just want to know the number of brown-outs: a UPS.

Many UPS devices (many APC and CyberPower) have software suites that connect with the UPS, and they measure the number of power quality events (brownouts, etc -- anything that would trigger the battery to come on). Depending on the severity of the brownouts (and the likelihood of them tripping the UPS) this could constitute proof of the problem.

I don't think this is a better way then trying to use the Rigol, but you may already own a UPS capable of recording this information.

Cheers,
-tg

[retrolefty]

At our refinery we would rent a Drantz power analyzer when we needed short term testing.

http://www.dranetz.com/

[joeqsmith]

Not a big fan of the Drantz myself but have not looked at a new one in several years.   I do like the HIOKI products.

http://www.ebay.com/itm/HIOKI-3196-600V-500A-Clamp-on-Hi-Tester-/380889300297?hash=item58aec57d49:g:bJIAAOxyBjBTVLzb

[ez24]

http://www.ebay.com/itm/HIOKI-3196-600V-500A-Clamp-on-Hi-Tester-/380889300297?hash=item58aec57d49:g:bJIAAOxyBjBTVLzb

Sorry I should say I do not have a lot of money.  Some of the solutions are more expensive than rewiring my house.

I do not think an UPS would work because they are so short.  They do not affect my Samsung TV, but they knocked (made it turn off - it was ok) out an old Bose radio one time.

We need a Rigol fan to weigh in on this.  Oh I probably could get an isolation for under 50 if I try for a couple months.  Missed a few already.  I am trying to find a 2 wire one so I do not have to rewire it.  So I will do this for sure.  But since I have a 2 wire house (the one with the exploding probe was 3 wire) maybe I do not have to worry?

I asked about this before and never got a good answer because no one seems to have a 2 wire house anymore.

thanks

[ludzinc]

Random thought:

Why not measure the output of a low voltage transformer?  110V AC on the input, 5V or what ever on the output.

Good isolation, more forgiving on you if you make a mistake, just be careful of the 110V side, which could be permanently insulated.

Log output voltage with a micro with RTC and timestamp "loss of voltage" events?

[med6753]

I asked about this before and never got a good answer because no one seems to have a 2 wire house anymore.

thanks

Are you saying that your residence is 2 wire service? That is, 1 hot and a neutral line? Damn, that's old! 

[ez24]

Why not measure the output of a low voltage transformer?  110V AC on the input, 5V or what ever on the output.

This is a good idea, I have some old AC wall wart I could use
thanks

[technogeeky]

http://www.ebay.com/itm/HIOKI-3196-600V-500A-Clamp-on-Hi-Tester-/380889300297?hash=item58aec57d49:g:bJIAAOxyBjBTVLzb

Sorry I should say I do not have a lot of money.  Some of the solutions are more expensive than rewiring my house.

I do not think an UPS would work because they are so short.  They do not affect my Samsung TV, but they knocked (made it turn off - it was ok) out an old Bose radio one time.

We need a Rigol fan to weigh in on this.  Oh I probably could get an isolation for under 50 if I try for a couple months.  Missed a few already.  I am trying to find a 2 wire one so I do not have to rewire it.  So I will do this for sure.  But since I have a 2 wire house (the one with the exploding probe was 3 wire) maybe I do not have to worry?

I asked about this before and never got a good answer because no one seems to have a 2 wire house anymore.

thanks

My UPS catches two or three brownouts per week that usually aren't even visible to my eye. These brownouts do not turn off any equipment in my house that I'm aware of. So I think a UPS will detect a lot more of these power problems than you think, but you are right that it still may miss some.

The point is, the UPS would give you a way to record the events (the UPS will record them anyway) and log them too (using the software they provide). And it will also give you a way to isolate your scope, for a short period of time at least. The Rigol does not seem to care about the quality of incoming power that much.

[technogeeky]

Why not measure the output of a low voltage transformer?  110V AC on the input, 5V or what ever on the output.

This is a good idea, I have some old AC wall wart I could use
thanks

This is a good idea. Especially if you have an AC to AC wall wart. I have done this with the Rigol in the past, with no problem.

Now all you need to do is:

• find the right vertical setting to be able to see the short brownout event
• set the horizontal to as long as possible
• set the trigger mode to standard and Peak, and just slightly lower than 115V (or lower than whatever voltage of the wall wart)
• record directly with the scope (record option), and get lucky, or
• record using some software (UltraScope?) and log the data
• 
• profit!

[rs20]

I once did this simply by setting up a 10 megaohm resistor (several good quality resistors in series) on the hot lead, forming a resistive divider with earth. Yes, earth, not neutral. I also did the same with neutral. Lots of heatshrink and so-on. This terminated into a 3.5mm jack into the line-in on my computer.

Some of you might be thinking "oh my god! you'll kill yourself". But if so, please provide an example situation where it could cause harm, because:

Earth leakage of 240 V * rt2 / 10 megaohm is well below any level of concern w.r.t. safety and/or tripping RCDs.
The only low-impedance path is the Earth path, which can only be a good thing.
Computer can subtract its left and right channels (live and neutral) to get a measurement of live-neutral, which is what you want. Assuming here that hot and neutral don't both take excursions well away from earth, which is a fair assumption in a neutral-bonded-to-earth-at-the-house scenario like mine.
Can record for days if needed, or do live analysis (on a live website, as I did).
Built from things you'll have lying around in your house already.

Now of course, it'd be better to add an isolation transformer into the mix, or even just the AC wall wart already suggested, but not strictly necessary.

[george.b]

Random thought:

Why not measure the output of a low voltage transformer?  110V AC on the input, 5V or what ever on the output.

Good isolation, more forgiving on you if you make a mistake, just be careful of the 110V side, which could be permanently insulated.

Log output voltage with a micro with RTC and timestamp "loss of voltage" events?

That's exactly what I was thinking. Make the micro sample the voltage into a circular buffer that gets saved in its entirety in the event of a fault.

[Assafl]

From the OP - it seems he wants to prove the problem is on the electric companies side and not the 100 year old wiring.

So both the resistor and transformer methods would have to be wired to the panel directly. Connecting to a socket still gives the electric company wiggle room to blame the wiring

[ez24]

From the OP - it seems he wants to prove the problem is on the electric companies side and not the 100 year old wiring.

So both the resistor and transformer methods would have to be wired to the panel directly. Connecting to a socket still gives the electric company wiggle room to blame the wiring

I thought about this but if I can get the data into a spreadsheet then I could correct the numbers to make the output be the main voltages.  I would not tell them how I collected the data.

[FrozenHaxor]

Maybe you could get yourself a scope like Rigol DS1054Z and set up a zone over your sinusoid with a pass / fail counter. Would give you exactly what you need.

[ez24]

Maybe you could get yourself a scope like Rigol DS1054Z and set up a zone over your sinusoid with a pass / fail counter. Would give you exactly what you need.

I have one.  After reading the manual (it will only record waveforms and settings) I think this is my only solution.  The zone would have to be lower values than normal waveforms, I have not figured this out yet.  If I cannot, I will post a question on how to do this.  I am not smart enough to build a MCU to do this.

I did find this
http://www.surgex.com/pdf/white-papers/APC_Whitepaper_UPS_7-types-of-power-problems.pdf

and it looks like I have either - sag or interruption

Now I am wondering how I can recreate the problem.

thanks

[blacksheeplogic]

I thought about this but if I can get the data into a spreadsheet then I could correct the numbers to make the output be the main voltages.  I would not tell them how I collected the data.

If you can't get the supplier to send a crew out to take a look at the pole using photographic evidence of a possible issue you will need a report from a qualified electrician.

An amateur report, especially where the person is coy about how they obtained/manipulated the 'data',  and/or equipment used most likely will get a generalized response and a suggestion to contact a licensed electrician.

Call your supplier and discuss your concerns honestly with them. It will be the same response (or possibly better) than you will get if you roll up with your proposed manipulated spreadsheet.

[Kalvin]

Use a transformer for galvanic isolation + safety and measure the secondary voltage in order to find out the actual mains voltage. You can even use the transformer for supplying the power for the measurement unit (arduino / raspberry pi / whatever) if you provide some kind of battery backup to make it survive over those mains problems.

[Kalvin]

You can also measure / monitor the approximate mains voltage by a lamp/LED and a LDR. Optocoupler can be used, too.

[Jeroen3]

Get a PQube. There are some on eBay. Yes I know, not cheap, none of this stuff is.

Do you see yourself mailing Rigol screenshot to the enery company saying: "look, my uncalibrated setup recorded a glitch!" ?

[ez24]

-  FrozenHaxor

Maybe you could get yourself a scope like Rigol DS1054Z and set up a zone over your sinusoid with a pass / fail counter. Would give you exactly what you need.

I have a Rigol and this is what I will try


- blacksheeplogic

If you can't get the supplier to send a crew out to take a look at the pole using photographic evidence of a possible issue you will need a report from a qualified electrician.

An amateur report, especially where the person is coy about how they obtained/manipulated the 'data',  and/or equipment used most likely will get a generalized response and a suggestion to contact a licensed electrician.

Call your supplier and discuss your concerns honestly with them. It will be the same response (or possibly better) than you will get if you roll up with your proposed manipulated spreadsheet.

Too far to take any pics.
There will be no "data" so no manipulation
Will be photos of the waveform


- Kalvin

provide some kind of battery backup to make it survive over those mains problems.

The problem is not bad enough to need backups

- Jeroen3

Do you see yourself mailing Rigol screenshot to the enery company saying: "look, my uncalibrated setup recorded a glitch!" ?

I see myself dropping off a Rigol screenshot saying " look at my oscope waveforms of glitches when it rains"  I do not think I will say "uncalibrated", plus I do not think it is important to note the voltage values, just prove the interruptions happened during rain events.

[rs20]

Get a PQube. There are some on eBay. Yes I know, not cheap, none of this stuff is.

My suggestion, as well as the wall wart suggestion, both cost $0.00, assuming the OP has typical hobbyist components (read: some old resistors, a broken headphone cable, and a computer) lying around.

Do you see yourself mailing Rigol screenshot to the enery company saying: "look, my uncalibrated setup recorded a glitch!" ?

How could a so-called "lack of calibration" cause a spurious waveform that looked like a glitch or brown-out? Mains electricity is a mere 50 Hz, this stuff is trivial. If the power company are going to be belligerent with oscilloscope and sound card readings, why would they suddenly bend over when this "PQube" thing turns up?

[Jeroen3]

I know from experience it's very hard to get the grid folks to come and fix the problem. It's never their fault.

This one time we found that a phase was missing on the hot side of the low voltage breaker at a customer. They kept insisting is was in the customers installation. 

[Berni]

That pole should be looked at since it could start a forest fire if it is in such a overgrown area.

I had the bad connection on a power pole for my house. Randomly the lights would brown out but very rarely did it make a PC reboot. Then after a month of that the connection finally burned itself trough and one of the 3 phases was gone. While the slow govorment owned power company was taking 3 days to come around to fix it I rewired the faulty phase on to a working one in my fuse box so that everything would work until they fixed it(they had to replace the entire pole)

So it's better get it fixed before it leaves you completely without power. Let the power company know. If you know how to safely work with mains open up all the junction boxes and check for any burnt stuff as the location of the fault gets very hot, give the wire a wiggle while under load to see if you hear sparking in it. If you are not experienced with mains then have an electrician look at it. You don't want to kill you self, but you also don't want your house to burn down due to an electrical fire.

But if you want proof of brown outs I would connect your rigol to to a 12v transformer. Set the probe ratio to 1:10 so you get 120v on screen. Then also connect your 2nd channel to a half wave rectifier to "AM demodulate" the peak voltage on the transformer to a DC level That way you can trigger the scope when that rectified voltage falls too low while you can see what happened with the 50hz sine wave on your 1st channel.

Since your rigol still has channels left you could also use a 3rd channel on another 12v transformer that is plugged in to another wall outlet on a different fuse to prove if the fault is before or after the fuse box.

[ez24]

That pole should be looked at since it could start a forest fire if it is in such a overgrown area.

If you are not experienced with mains then have an electrician look at it. You don't want to kill you self, but you also don't want your house to burn down due to an electrical fire.

But if you want proof of brown outs I would connect your rigol to to a 12v transformer. Set the probe ratio to 1:10 so you get 120v on screen. Then also connect your 2nd channel to a half wave rectifier to "AM demodulate" the peak voltage on the transformer to a DC level That way you can trigger the scope when that rectified voltage falls too low while you can see what happened with the 50hz sine wave on your 1st channel.

Since your rigol still has channels left you could also use a 3rd channel on another 12v transformer that is plugged in to another wall outlet on a different fuse to prove if the fault is before or after the fuse box.

My "problem" only happens when it rains so I do not worry about pole fires.  I ordered a 12v transformer - thanks.  My scope has an envelope trigger (I have not figured it out it).  If that does not work I will try the rectified solution (sounds good).  I have 60hz here in the US.  I have until November to figure this out.

thanks