AC measuring - 7 YEAR battery life!!

JoeJester

Joined Apr 26, 2005
4,126
Kent,

A problem you haven't considered is once you solve the "interference" from 60 Hz, you'll have to deal with the "interference" from 180 Hz.

Attached is a spectrum view of a 60 Hz signal at a location in Western Oklahoma from February 1999. I've only included the primary to the fourth harmonic. I have a panoramic view out to the 24th harmonic and will include that upon request.

Your problem may not be as simple as getting rid of the 60 Hz and taking a reading on the signal of interest.
 

Attachments

Thread Starter

a_kent

Joined Jun 12, 2007
30
Humm, now that is interesting!!
I had not thought that much about harmonics.

What I did do was add a low pass filter before it enters the ADC of the processor. It's -3db at 200hz, though...

I suppose I'll likely need to do some filtering in software, too...

I am using an AVR that has bipolar inputs to the ADC.

Thanks,
Kent
 

JoeJester

Joined Apr 26, 2005
4,126
Kent,

If you have access to a signal generator, you certainly can test it at varying frequencies. You'll certainly need a sharp roll off for your low pass filter.

Remember, your signal of interest [the neg 0.7] is probably 40+ dB down from the interference. If you were to work on keeping the interference 60 dB down from the signal of interest, that's a 100+ dB attenuation.

I hope you read about the noise problem in the other thread ... one story illustrates the need to eliminate the problem at the source of the problem.

J.A. Pierce, who chose the frequency for the Omega Navigation System [no longer available] stated he was upset because he chose the 170th Harmonic of 60 Hz as the frequency ... obviously 60 Hz or harmonics had some localized interference to Omega.

Here's the panoramic view. I annotated the graphs with the levels and frequency ... for your viewing pleasure.
 

Attachments

pebe

Joined Oct 11, 2004
626
Hi Kent
I’ve read through this thread over and over again, and I’m still not sure what you want to measure.

In post #9, you said
Right now, technicians must go out in the field and measure the voltage between the pipe and the buried half cell with a voltmeter…
….The goal is to measure these voltages and send them in over a cell phone, so the tech won't have to go out into the middle of nowhere to measure it.
Is that the main purpose of the measurements and is that DC voltage generated with the half cell?
Also, it's a safety thing, as tech's have been zapped before...
Is that the only reason for measuring the induced 60Hz from mains cables? If so, why do you need to accurately measure it?

Then in post #16 you said
The rectified DC comes from rectifier stations that are placed along the pipe. Many are many KM (miles) from each other. They strive to keep a negative potential on the pipe. They have anodes associated with them as well.
What is the –ve potential on the pipe relative to, ie. where is the other electrode for the applied DC? Can you say what the potential on a) the pipe, b) the half cell is, relative to ground.

If the answers are what I think they are, then the circuitry for the measurements could be quite simple. One thing surprises me though. You expect a 7-year life from the battery – bearing in mind it must transmit data over a cell phone every hour.
 

Thread Starter

a_kent

Joined Jun 12, 2007
30
Hi Pebe:
I'll try to answer you.

The half cell is at ground potential and is the measurement point for the 'anode'.
The tech in the field is interested only in measuring the DC between the pipe and half cell. One wire from each comes up into a box near the pipe. I cannot answer why they are using a half cell, except that it also creates a potential with the pipe and earth creating a battery of sorts.

I need to read the induced AC along with any DC that might be there. I must provide an RMS value of both within 1%.
I'm with you, I don't know why this kind of accuracy is needed.

The negative potential on the pipe is relative to the half cell, which is at ground potential.
The pipe is negative relative to it, and forms the cathode.
They call this whole scheme 'cathodic protection'.

Oh, the battery life is only for this measuring device. The box that contains the cell phone is powered by either a large solar powered battery, or an AC power supply, if AC is available. I cannot derive power from that, as my little box will be remote from that. I will supply a 0-5v DC signal that is relative to the RMS that I measure.

The pipe can be anywhere from 0 to -100VDC RMS, again, relative to the half cell. The amount of negative voltage placed on the pipe depends on how far the rectifiers are from each other along the pipe, and the type of ground the pipe is in. My measuring device could be placed anywhere along the pipe.

I have attached a document that describes these systems. That might help in the understanding.

I hope this helps.
Thanks,
Kent
 

Attachments

pebe

Joined Oct 11, 2004
626
The cathodirectifier.jpg you quote shows the negative voltage of 100 – 220V rectified being applied between a pipe and ground as cathodic protection.

I’ve looked at the file cathode protection.pdf and it gives little information about measurements, other than the level at the protected steel seems to be around minus 800mV relative to the half-cell. The pipesoil.gif you quote shows a graph of a survey where the voltage is between 250mV and 700mV – but is quite meaningless without some descriptive text.

You originally said that the induced AC levels in the pipe from underground power lines can be as high as 220VRMS.

So you have three indeterminate voltages. I see no way that you can combine these together to give a signal that is meaningful.

Have I misinterpreted something?
 

JoeJester

Joined Apr 26, 2005
4,126
Pebe,

Yeah, some light reading. I got the gist of it looking through the Operations and Maintenance manual ... with the engineering portion for further reading.

It boils down to the Anode material connected to the positive terminal and the pipe connected to the negative terminal. When you measure between the pipe and another ground reference, you should read a neg potential. A desired reading of 0.75 volts is obtained when sufficient current passes. Sounds like testing a diode to me. I'm sure there is a correlation between the condition of the metal structure and the current required to achieve the neg 0.75 volt drop.

I'm guessing the slang calling it a half-cell is because of the potential being one-half of a single battery cell.

Where he's getting the 100 Vac rms, I don't know, unless a particular diode shorted and allows the positive half cycle to be on the line. With the postive potential being ground, it seems to me a nearby lightning strike could short out those two diodes connected to the positive potential causing this condition. I can't tell you how many times the diodes on the return side shorted out when there were lightning strikes when I was in Oklahoma.

I wish they could post a pic illustrating the reading they are getting on the test lead. I mean a pic of the oscilloscope, not an artist rendition.

If the problem didn't exist before, and it exists now. What has changed? Masking the problem isn't solving the problem. I'm sure the law required them to keep records of the readings ... as the state isn't going to accept them saying 'The readings were normal'.
 

pebe

Joined Oct 11, 2004
626
Hi Joe,
Well, I’ve read that publication about cathode protection – or rather, I mean I’ve scanned through it and read the most interesting parts! It says that measurements using a half cell are made with the protective current switched on. With current flowing through the ground between pipe and buried anode, there must be a voltage gradient through the ground between them. I'm surprised that the half cell is buried in that ground in no specified position, yet the readings appear unaffected by that voltage gradient.

I’ve also read elsewhere that measurements made with this half-cell can also indicate the state of corrosion – but in that case the protection current has to be switched off. The more I read the more confusing it gets!

Anyway, it appears that the voltage to be read is nominally 850mV but could go to over 1V. The waveform is full wave rectified DC – unsmoothed.

Kent.
If you have not yet gone too far down the path of getting PCBs finalised, may I suggest this:
  • You say the 100V interfering signal was measured with a scope. Could it have been a measuring error like the scope was ungrounded or had unscreened leads or something? With 100V on the pipe I would have expected a big bang somewhere. Perhaps it might be better to rectify it and get it down to within the range of your A/D and treat it as quite separate from the test results.
  • The publication Joe referred to shows an analogue meter used. No mention is made of RMS measurements so it must be reading the mean DC level. So no conversion is required.
  • From what I’ve read, the ground is like a conductive bridge between the pipe and the half-cell. Its potential is somewhere between the two voltages at the test points, so you could use a DC coupled single op-amp as the input device. Something like a TC911 from Microchip wired with a gain of 3 or 4 would feed straight into your A/D converter. A simple RC filter in each leg would probably provide enough smoothing. If not the amp could be followed by an active lowpass filter.
  • For a supply, how about 4 x AA NiMH kept charged from solar cells?
  • Sending a 0-5V analogue signal any distance to the cell phone might present problems. Any chance of coding it?
I hope that may be of some help.
 

Thread Starter

a_kent

Joined Jun 12, 2007
30
Thanks for your looking into this and replying!

YES, the whole thing is confusing, and the real world and these documents are not always the same thing.
===
In an 'ideal' world, a -0.8v potential is perfect. In many places, this works and is just fine.

In the 'rest' of the world, these voltages are all over the map. As I believe I mentioned before, if a DC potential is applied to a pipe several KM away from a measuring point, it will eventually dissipate over distance, unless this potential is very high to begin with. This, sometimes can be up to 100VDC RMS applied at a particular rectifier, just to get the minimum at some distant point.

If you measure some KM away from there, it will be at a low of the perfect -0.8v, but obviously not if you measure it at the rectifier!

Also, there can be, and usually is some amount of induced AC onto the pipe, if there are HVAC lines nearby. As I already mentioned, this can be up to 100VAC RMS. Maybe higher, but this is the limit that I am supposed to design for.

There was apparently no error with the scope measurements, as techs have been zapped pretty hard from the induced AC (and/or high DC? I don't really know. Apparently noone really knows.).
===
The analog meters are in theory measuring RMS. In reality, everyone uses DVM's these days, and all measurements are listed as RMS. Note that all of this is really old tech. The active cathodic protection idea came about back in the 50's.
===
There is no chance of coding, but I did get more clarification about the output of this, as I have had the same concerns about remoteness. I was finally told that this measuring device will be in the same box as the cell phone based transceiver. But, I can use NO power from it. So I'm still battery powered, unfortunately. I don't know why they didn't tell me that before.

This is a retrofit measuring device that will use one of the transceiver's ADC inputs.
===
Pebe:
Maybe I'm misunderstanding what you are saying about DC coupling the input and an opamp. If all I had to do was measure the low level DC, then that would probably work, but as you have read, I must deal with 100VDC + 100VAC, max. I need to RMS compute both, so I need to see the full waveform at the ADC.
If there is no AC on the line, and I show no reading for a low DC level, then that's OK. If there is no AC and a high DC, then I must report that.
There are other devices that accurately measure the DC levels, but they don't measure AC at all.

Thanks for your interest and help! This is sure different sets of problems!
Kent
 
Top