(from Alex McEachern) Electric Power Quality Newsletter - December 2012

ARPA-E -- the Advanced Research Project Agency of the United States Department of Energy -- has selected a team to investigate a fascinating question: What can we learn from tiny phase angle shifts, or twists, that occur along an AC power distribution line? Especially when there are solar generators or wind turbines along the way?

PSL has developed the new ultra-precise instrument for this project. It's a daunting metrology challenge - measure tiny fractions of a phase-angle degree, in uncontrolled temperatures, simultaneously at a large number of locations.

When the measurements need to be this precise, unexpected questions pop up: for example, just what is the phase shift on a distribution transformer when we're measuring at the milli-degree level, and how will it change with varying loads? Combine this precision with a requirement for ultra-low-cost, and you've got the kind of instrument challenge PSL enjoys!

Our new instrument is just one part of the project. Another part works on new communication channels and a new type of network, and the main focus of the project examines distribution diagnosis and control: everything from system situation awareness to possible power oscillations to islanding.

The project is led by Dr. Carl Blumstein and Dr. Alexander von Meier of CIEE, both old friends of Power Standards Lab. Rounding out the team are a great group of researchers from Lawrence Berkeley National Lab, Prof. Culler of U.C. Berkeley, and Prof. Mack Grady, who is now at Baylor University. It's a cheerful, thought-provoking group with a truly stunning level of expertise.

We're looking forward to this fascinating new instrumentation project. Stay tuned.

A couple of weeks ago I learned that I've been elevated to Fellow of the IEEE, the highest membership status, for "contributions to power quality measurements and immunity". I'm grateful, and proud.

And yesterday I learned that my mentor and guide, Francois Martzloff, has received the IEEE's Lifetime Achievement Award "for a lifetime of integrity, leadership, and mentorship...". It was a special pleasure to read the mention of his graceful writing in the announcement of Francois's well-deserved award.

With 400,000 members in 160 countries, the IEEE is a truly great engineering organization. I remember my pride when Francois first suggested that I become a Member, and, many years later, when Jerry Heydt nominated me as a Senior Member. It didn't occur to me that I could rise further. (Another favor Jerry Heydt did me, all those years ago: he introduced me to a new concept called "email", during as escalator ride at an IEEE conference.)

As usual, I find myself agreeing with Math Bollen, who told me that becoming a Fellow was among the proudest moments in his life. He's absolutely right.

My thanks to everyone like Francois who has helped me develop as an engineer: Math, Mark, Erich, Bill, Bob, the other Alex, Fred, Alec, Marek, Matthew, Dan, Ron, Jean and so many others - thank you all from the bottom of my heart. And I must mention my gratitude to my grand-dad, an electrician on the Canadian Pacific Railroad, who -- to my mother's dismay, I now think -- gave me my first soldering iron when I was just 7 years old. It's good to get children launched early, isn't it?

	PQube Sag waveform	PQube Sag RMS

In my last newsletter, I showed this voltage sag recorded by a PQube in London during the Olympics, and asked if anyone could explain the strange voltage phase shifts during the sag. (The PQube instrument produces nice Excel files, and PQDIF files too, so the raw data is readily available.)

I received many responses, but two stood out.

Dan Sabin of Electrotek used PQView -- a great program -- and his excellent engineering intuition to convert the PQube current waveforms into positive-, negative-, and zero-sequence components. He used those symmetrical components to demonstrate that this was a fault downstream of the PQube that began as a single-phase-to-earth fault for about a half-cycle, then developed into a phase-to-phase fault for another half-cycle or so, then finally developed into a 3-phase fault that endured for about 4 cycles until the protection device removed the fault. I particularly like Dan's approach because he converts to a domain we don't use often enough: symmetrical components (if you want to learn more about this, download my free Power Quality Teaching Toy program at http://www.powerstandards.com/PQTeachingToyIndex.php.) We're all used to working in the time domain and the frequency domain, but it sure is useful to remember there are other ways to think about these problems. Thanks, Dan!

My old friend Patxi Pazos of Iberdrola in Spain took a completely different approach to reach essentially the same conclusion. Patxi developed an impressive circuit simulation, including a couple of interesting home-brew arc simulatons, and could duplicate the real-world voltage and current waveforms almost exactly. Now that takes a real power engineer.

(If you'd like to contact Patxi or Dan, please let me know by replying to this email. I will connect you.)

As we've noted in Edition 3 of -4-30 (power quality measurement methods), current itself is not a power quality phenomena; but current waveforms can certainly be useful for understanding the causes and solutions for some power quality disturbances. And unlike voltage, when it comes to current measurements you have a choice of several different sensor technologies.

It turns out that it's much easier to measure voltage than current. For that reason, all current sensors are, ultimately, just a way to convert an unknown current into a measureable voltage.

The simplest current sensor is a known resistor, with the current flowing through it. At Power Standards Lab, we use a 4-terminal resistor when we need the ultimate in precision current measurements (if you don't know about 4-terminal resistors, here's my quick explanation). The resistor in the photo to the left is one of PSL's two 0.001000 ohm reference 4-terminal resistors. Three challenges with using resistors to measure AC current, though:

1. Installation - to install the resistor, you must break the circuit you want to measure. Not good for data centers, or anywhere else where reliability is a big concern.
2. Heat - Even with a 0.001 ohm resistor, if you measure 400 amps you get 0.4 volts, which means 160 watts of heat in the resistor. And heat causes the resistor value to vary, slightly. And the heat goes up with the current squared...
3. Common-mode noise - In the example above, you're trying to measure 0.4 volts, probably to within a few hundred microvolts, on a resistor that may be bouncing up and down by several hundred volts. Possible, but difficult.

A popular current sensor, the Rogowski coil, converts the rate-of-change-of-current flowing in a conductor into a small voltage. Rogowski coils are flexible, so they're easier to poke into nooks and crannies to get them wrapped around a conductor. But they have some challenges:

1. Requires an integrator - the output of a Rogowski coil is proportional to the differential (with respect to time) of the current. To get the actual current signal, you have to integrate the Rogowski signal. Integrators have three problems:
   1. They introduce a time delay, which is -- in power terms -- a phase shift. If you're interested in precise power measurements, this is a problem.
   2. At any instant in time, the output of an integrator is a function of its past inputs. Not a big problem if you're dealing with steady-state sine waves, but can be a serious issue during disturbances.
   3. It's more difficult to interpret oddities in the signal: when there's something strange, was it in the current, or in the Rogowski coil, or in the integrator, or the power supply for the integrator, or some combination, or...?
   4. DC offsets must somehow be removed from the integration process -- otherwise, the output signal runs away to one of the rails. Removing the offset, though, introduces an imperfection in the integral. It's a difficult trade-off.
2. Specifications - Almost always, Rogowski specs are given for RMS accuracy with a steady-state current, usually sinusoidal. It's hard to find Rogowski specs for transient current waveforms, phase delays, and (sometimes) harmonics. (There are good commercial reasons that explain why these specs are difficult to find.)
3. Interference - Because a Rogowski coil has an incomplete magnetic loop, its output signal can sometimes be affected by large nearby currents. If you've got nice balanced three-phase currents, Rogowski coils work well; if you don't, you might get confusing results.

Solid-core current transformers are the most straightforward AC current sensors: they output a smaller current that is an exact ratio of the measured current. (Some solid-core CT's have a built-in burden resistor that converts the secondary current into a voltage.)

1. Installation - You're forced to break the conductor you want to measure. So solid-core CT's are usually installed before equipment is put in service.
2. Accuracy advantage - Solid-core CT's have a perfectly-closed magnetic path in their core, and their secondary winding can be equally distributed around the core. The former improves their noise immunity, and the latter makes them less sensitive to the physical position of the primary conductor.
3. Shorting / Safety - For CT's with current outputs, it's important to short the secondary at all times. That's why I prefer CT's with built-in burden resistors...

A good compromise, I think, is to choose a split-core current transformers. They function like solid-core CT's, but are easier to install. They can be snapped around a live cable (with the proper safety precautions, of course).

You can find them with overlapping-core-laminates where the core is split (left photo), or with polished core-joint surfaces (right photo). The overlapping-laminate makes more sense to me for permanent installations, but the polished core-joint surfaces work well if they're kept clean and undamaged.

1. Easy installation - Just snap it around the conductor. Here's a useful trick: if you already have CT's installed that have 5-amp secondaries, you can snap a PSL 5-amp split-core transformer like the SCN2-5A around the 5-amp secondary wire, and never even go close to the main conductor.
2. Accuracy - Excellent accuracy - on the order of 0.2% and 0.2 degrees - here's a full list of high-accuracy current transformers for PQube instruments.
3. Closed magnetic core - Better immunity to external magnetic fields.

There might be significant changes to the Class A requirements in IEC 61000-4-30 - if that's important to you, or to your clients, at PSL we're suggesting that you wait until the March 2013 to make any important decisions. The next meeting of Working Group 9 takes place in Bilbao, Spain, on 18-19 March 2013. I expect that meeting will answer all the questions about what, exactly, will be required for the future of Class A. Please reply to this email if you would like an update.

If you have PQube instruments, the PSL engineering team has added a lot of great new features in PQube Firmware 2.1. You can set thresholds for temperature and humidity events, define your own depth-duration trigger curves, even configure your PQube to send different emails to different people. As always with PSL, firmware upgrades are FREE - just download one file onto an SD memory card, pop the card into your PQube, and your PQube will take care of the rest...

With my best wishes to you for a peaceful, healthy Winter holiday -

Alex McEachern Alex@PowerStandards.com Power Standards Lab 2020 Challenger Drive       Alameda, California 94501 USA TEL ++1-510-522-4400 FAX ++1-510-522-4455

(I have sent this e-mail to you at '[recipient]', because you are on my personal world-wide list of 23,240 engineers, educators, and students interested in power quality. If you no longer wish to receive it, please let me know.)