The nice thing about the APC units is that they are close to free if you are
already going to purchase a UPS.

I agree that something like the Dranetz 658 would be better, but a quick peek
at eBay shows prices far beyond what I'm willing to pay.

I don't know what the internal sampling rate is.  The API is
tell me the current voltage
tell me the lowest voltage since the last time I asked
tell me the highest voltage since the last time I asked

I think I decided it's an 8 bit ADC so the resolution is far from wonderful.
(The step size on a couple of handy readings in 0.7 volts.  8 bits gives a
full scale of 180 volts.))

I have a hack that reads as fast as it can.  If nothing interesting has
happened, it adds a line to the log file every 5 minutes.  If the min or max
voltage has change enough, it logs a line now.  That gives me reasonably
accurate timing on short glitches without cluttering up the log file with
noise.

--
These are my opinions.  I hate spam.

On 1 Jan 2017 11:10, "Hal Murray" hmurray@megapathdsl.net wrote:

are

peek

Em, not a lot. My handheld true RMS Tektronix can give me the average. (One
assumes an average of RMS values).

Again,  I think if attending presenting data for others,  one wants to
avoid hacks like that. One can always post-proces to indicate the points of
particular interest.
My biggest problem is that it is not very practical to log data at the
incoming point, which is just above my back door.  If I lived on my own,  I
could set up equipment easily to do this. But sharing a house with a my
wife and a large German Shepherd dog, it is not practical to do it with the
equipment I have.

I think measuring voltage elsewhere would give someone more reason to
question its accuracy.  In my case, measuring in my lab would almost
certainly give a power supply voltage lower than that coming in.

Anyway,  short term I will use a variac to lower the voltage to test
equipment with linear power supplies.  I am less concerned about equipment
with switch mode supplies.

Dave.

Hi:

A few quick comments.

I've used a Variac for years at home to drop the line voltage for older equipment with linear power supplies that run hotter than I would like.  (My HP5370B's don't fall into that category for me but I can understand why this is an issue for some individuals.)

I've encountered situations where the line voltage has been deliberately lowered to entire buildings which has in turn caused issues for equipment I was responsible for.

In dealing with line voltage issues in Canada I've found that readings from my handheld fluke DMM seem to be accepted at face value by the individuals I've been dealing with.  Data collected from UPS systems doesn't seem to be as well accepted.

Mark Spencer

There are a couple of recent threads concerning the power line mains
voltage standards. After a bit of research and thinking, I have found
that this is a complex topic. The simple answer is:

• The standard in the US for the past 50 years has been 120/240 V +/- 5%
  RMS at the service entrance to the building. This is a range of
  114/228 V to 126/252 V.
• The load voltage could be as low as 110/220 V and as high as 125/250 V
  and be within specifications.

There are two voltage measurement points to consider:

(1) Service voltage: This is the RMS voltage measured at the service
entrance to the building (at the metering point).
(2) Utilization voltage: This is the RMS voltage measured at the load.
It might be measured at an unused socket in a power strip feeding
several pieces of electronic equipment, for example. There are of
course many different utilization voltages present in a home or
business, depending on where you make the measurement.

Most US homes and small businesses are powered by what is commonly
called a "split-phase" 240 V feed. The final distribution system
transformer has a 240 V center-tapped secondary. The center tap is
grounded, and three wires are fed to the building (actually it might be
up to around 6 houses):
(1) Leg L1 or phase A (red wire) -- This wire will measure 120 V to the
neutral or 240 V to Leg L2.
(2) Neutral (white wire) -- This wire is grounded at the distribution
system and at the service entrance to the building.
(3) Leg L2 phase B (black wire) -- This wire will measure 120 V to the
neutral or 240 V to Leg L1.

Large appliances and HVAC systems are usually connected across L1-L2
(240 V), while most sockets are on circuits either connected across L1-
neutral (120 V) or L2-neutral (120 V).

The voltages I have described are the current standardized values for
the service voltage which have been in general use for about 50 years
(120/240 V +/- 5%). I believe that the original systems installed before
1940 were designed for a 110/220 V nominal service voltage, but after a
report in 1949 the nominal service voltage was increased to 117/234 V,
as specified in ANSI C84.1-1954. After research in actual buildings, in
the 1960's the nominal service voltage was increased again, to 120/240 V
in the ANSI C84.1-1970 standard. The purpose is to keep the utilization
voltage at the load above 110/220 V.

The voltage at the service entrance should in most cases be in Range A
(120/240V +/-5%). On each 120V leg the service voltage should therefore
be between 114 and 126 V. The utilization voltage at the load should be
between 110 and 125 V due to losses in building wiring.

See details of the current specifications at:

http://www.pge.com/includes/docs/pdfs/mybusiness/customerservice/energystatus/powerquality/voltage_tolerance.pdf

These voltage specifications were designed for resistive loads and
measurement of the true RMS voltage. In most electronic equipment built
over the past 50 years, the power supply input circuitry is basically a
rectifier connected to a smoothing capacitor. This leads to high input
current surges during the peaks of the waveform, so that the peak
voltage is reduced much more by the building wiring resistance than if
the load was resistive for the same power consumption.

So the waveform shape at different utilization locations in a building
(with active equipment loads) may be different, so the voltage measured
by different AC measuring instruments can differ. Many meters are full
wave average measuring but calibrated so they only read RMS voltage
correctly on pure sinewaves. Other meters are true RMS measuring and
will read very close the correct RMS voltage even if the waveform is
distorted.

Bill Byrom N5BB

On Sun, Jan 1, 2017, at 12:16 PM, CIW308 VE6OH wrote:

Thank you for the detailed analysis, Bill. The voltage measurements I made
in my garage laboratory were duplicated by the utility with their meter,
which was connected at the service entrance. We each showed voltage in
excess of 126 VAC. Date from the (U of Tennessee) Frequency Disturbance
Recorder also showed voltages in the 124-128 VAC range. The insignificant
voltage drop in the lab was due to the 200 Amp service (the house was
originally "all electric") and minimal load. In response to my concerns,
the utility dialed the voltage down to about 123 VAC where it remains today.

Jeremy

On Sun, Jan 1, 2017 at 8:49 PM Bill Byrom time@radio.sent.com wrote:

Sent from Gmail Mobile

On Mon, Jan 2, 2017 at 4:49 AM, Bill Byrom time@radio.sent.com wrote:

When someone here previously mentioned observing high voltage, one
possible cause for this in this common "split-phase" configuration  is
that if the neutral wire is overloaded, damaged, poorly connected, or
otherwise has high resistance,  the voltage on the two legs will swing
wildly and in opposite directions depending on load.

So, e.g. if you put a 1kw load on L1 while L2 is nearly unloaded then
perhaps L1s voltage drops to 108v while L2 rises to 132v.

The reason for this is that, e.g. imagine that the neutral were
removed completely you would effectively be connecting your appliances
in a parallel-series circuit (all on L1 in parallel, all on L2 in
parallel, the both in series) across the 240v feed.

I've had issues with neutrals several times in the past, and in one
instance, temporarily dealt with it by moving as much of the load to
240v as I could,  manually balancing the remaining loads, and then
using a digital multi-meter to dynamically control some additional
load to keep the voltage sane on each side.

I think the fact that you can end up with a much higher voltages at
the outlet if the neutral has problems is one of the more unfortunate
properties of the split-phase approach.

It's not a split phase system in US residential power, it is a
center tapped 240V single phase system.  Split phase systems have
historically had a 45-90 degree phase difference between the split
phases.  The US system, depending on which wire lead you take as
your reference, has a 0, or a 180 degree, phase difference.

The reason it is done this way, is for safety.  The center-tap
of the mains transformer is grounded to earth, as is the neutral,
and service entrance panel grounds.  This way, if the power company
installed grounding system is working properly, the highest voltage
that any residential customer could accidentally encounter would
be 120V to ground.  (It is not an accident to go mucking around
inside of a 240V range/dryer socket, or the service panel!)

As it has been noted, if a US based system is defective, people can
get hurt. The same is true for the European system.

Back in the dark ages of ~220V electrical distribution systems in
Europe, the reaping due to unintentional grounding of a ~220V wire
was  so common and extreme, whole house ground fault interrupters
were mandated for all residential/small business power systems
therein.

And, in so far as properly functioning GFI protectors are in use,
and can be maintained, they have been wildly successful!

I will have to leave discussions of which system is better/safer/
cheaper/more reliable, for another time and forum...preferably one
where there is beer and loud music.

-Chuck Harris

Gregory Maxwell wrote:

In message 586A8B40.4050806@erols.com, Chuck Harris writes:

Close, but no cigar.

The main problem was that in many countries outlets did not have a
protective ground terminal.

That meant that an internal fault in your appliance had a 50/50
chance of lighting up some exterior metal part you could touch.

The "obvious solution" isn't obvious in countries where the geography
does not allow you to obtain proper "protective ground".  Norway being a
good example.

But even countries with the "obvious solution" of protective ground
in all outlets saw problems, because it took 10-16 ampere misdirected
current to blow the fuse, and you can light most flameable stuff
with a lot less energy than that.

The "Residual Current Device" solved both problems.

RCD's even protect you from internal faults where proper protective
ground is not available, by providing neutral from "outside" the
RCD as PG in the installation.  You'll still be (horribly!) exposed
of an accident in the distribution grid (or lightning!) fires up
the neutral, but that's simply life - or death - without a grounding
rod.

--
Poul-Henning Kamp      | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG        | TCP/IP since RFC 956
FreeBSD committer      | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.

What modern loads are actually sensitive to high (say, +10 to +20%) line
voltage?

Old incandescent light bulbs were among the most sensitive loads in the
past (so much so, that 130V light bulbs were commonly available from the
industrial suppliers).

I would naively expect the modern CFL's and LED replacements to be fine
with higher line voltage because they have their own built-in switching
regulation.

A lot of modern electronic equipment with switching supplies, are just fine
at +20% line voltage and may even run cooler.

Tim N3QE

On Sun, Jan 1, 2017 at 11:49 PM, Bill Byrom time@radio.sent.com wrote: