Why do street lights flicker in snowy weather?

[Jeremy Parker]

[snip]

I had no idea!

Well, if they are all guilty, you wouldn't see it very often

Jeremy Parker

 

[Tzortzakakis Dimitrios]

When I lived in Massachusetts, the only times my lights flickered was
during lightning storms. Most of that occurred when demand was almost
equal to capacity.

Demand is always almost equal to capacity. If it would be more, it would be
a waste, if less there would be serious stability issues. That's the problem
with wind turbines, you just don't know when the wind blows, and you have to
cover each MW of wt with at least of 700 kW conventional reserve, because
when the wind stalls what? Stop everything?

 

[James Sweet]

What most often happens is that the lamp shuts off completely, and then
restarts, which takes it just as long to pump up as before because the
run temp is far higher than the shut off idle temp, even if after only a
couple minutes.

I used to make a joke and said that I shut them off 'with my aura' all
the time because we often saw them shutting off as we approached while
out getting a buzz, back in 'the day'.

Yeah that's what I was saying with HPS (metal halide and mercury behave
similarly). The higher pressure in a hot arc tube raises the strike
voltage beyond what is available from the ballast. It has to cool until
the pressure reduces to a point that the arc will strike.

SOX lamps are much lower pressure to begin with and will usually
restrike hot.

 

[Guest]

Demand is always almost equal to capacity. If it would be more, it would
be a waste, if less there would be serious stability issues.

-----------------------------
Please check the definitions of demand and capacity - as the above is
nonsense.

Ideally the capacity should exceed the demand by some optimal margin but as
adding and dropping on line capacity is in blocks corresponding to the
capacity or rating of individual generators, and demand is up to the
customers (predictable but not controllable) the capacity will normally
exceed the demand by a fairly large margin at times- there is no "stability"
problem. As for waste by having extra on-line generation- economic dispatch
optimization is a common procedure.
If demand exceeds capacity, then problems can occur- not necessarily
stability problems.
As for the wind turbine reserve, you are being a bit over optimistic. You
are assuming 30% availability of wind capacity. In practice, from recent
data it appears that 10-15% is a better figure and this is a statistic based
on an annual average, which means nothing if wind fails. In other words.
reserve capacity must be available for the worst case situation- 100%
failure of any generation source, concentrated as in a 500MVA fossil plant
or distributed as in 500-1MVA wind units in a region where wind diversity is
small.

 

[Guest]

With the SOX I've seen doing this, the flicker is happening when fully
run-up. I haven't watched one of these during run-up, so I don't know
if it happens during run-up. Temperature wasn't low when this happened,
and I suspect that's not relevant.

I don't have a sample dead one with this failure mode (which isn't
common), so I can't inspect the lamp to see what visible failure
indication there might be. As a pure guess, I might speculate that
the emission material is sputted off the electrodes, and it can't
sustain an arc in thermionic emission mode. The 35W SOX has an
ignitor/starter which is probably repeatedly trying to start it.
The larger ones use a leakage reactance transformer to provide both
the starting voltage and current limiting, because the arc voltage
is too high for a simple series ballast on 240V, and so don't need
an ignitor/starter. I don't think I've seen the larger ones flashing;
when the emission coating wears out, they seem to fail to light up
at all (or with only a very dim glow around the electrodes which
you can't see from the ground).

IME more common failure mode of SOX is the arc tube develops a leak
and the sodium is ejected into the outer vacuum tube, where it often
forms an opaque mirror coating on the inside of the bulb facing
the ejection point, so the light no longer escapes through part of
the bulb (can block out most of it eventually). The arc tube seems
to be able to lose a lot of sodium in this way, yet still work, but
eventually it turns into a dim red neon light which never runs-up
(nicknamed a "red burner"), as there's no longer enough sodium left
in the arc tube, just the neon starting gas.

----------------
I have seen many such lamps flicker but have never associated this with
temperature. I have not seen any more excessive flickering in cold
conditions than in warm conditions
By cold conditions, I mean central Alberta, Canada, -30 to -40C is not all
that uncommon and windchill down to -50C equivalent.
However, there could be differences in design that affect this. As an
example we don't run exterior water or sewer pipes

 

[Andrew Gabriel]

-----------------------------
Please check the definitions of demand and capacity - as the above is
nonsense.

Ideally the capacity should exceed the demand by some optimal margin but as
adding and dropping on line capacity is in blocks corresponding to the
capacity or rating of individual generators, and demand is up to the

Most of the generators have variable power output, not simply on
or off.

customers (predictable but not controllable) the capacity will normally
exceed the demand by a fairly large margin at times- there is no "stability"
problem. As for waste by having extra on-line generation- economic dispatch
optimization is a common procedure.
If demand exceeds capacity, then problems can occur- not necessarily
stability problems.

Frequency drops below nominal, and conversely when supply exceeds
demand, frequency increases above normal. There's a requirement in
the UK for frequency to average out correctly long-term (so things
like synchronous clocks don't drift long-term), consequently,
supply has to exactly match demand long-term. However, since the
demand and supply can't exactly track each other short term due to
inherent lags, there are periods of both demand exceeding supply,
and supply exceeding demand. These are both inevitable due to the
supply lag with different types of plant and unexpected plant
failures on some occasions, and deliberately forced to correct for
earlier drifts on other occasions.

There is contingency reserve in addition to the supply - additional
plant spinning sychronous online ready when needed due to either an
increase in demand or an unexpected loss of supply, and yet more
plant offline ready to run up and cut in with a bit more notice.

A longer article I wrote on this some years back, with some examples
of how it was applied in the UK to some specific historic events,
and how it went wrong on one occasion...
http://groups.google.co.uk/group/sci.engr.lighting/msg/99b03c9711a4f753?hl=en

As for the wind turbine reserve, you are being a bit over optimistic. You
are assuming 30% availability of wind capacity. In practice, from recent
data it appears that 10-15% is a better figure and this is a statistic based
on an annual average, which means nothing if wind fails. In other words.
reserve capacity must be available for the worst case situation- 100%
failure of any generation source, concentrated as in a 500MVA fossil plant
or distributed as in 500-1MVA wind units in a region where wind diversity is
small.

The BBC did a programme about the wind power in Denmark, one of
the highest users of wind power. In spite of installing lots of
turbines and being able to point to all the power they get from them,
they haven't been able to spin down any conventional generating
station, because they need them when the wind stops. When the wind
blows, they have an excess of conventional electricity which they
sell, but for their neighbours, it's effectively as unreliable as
the wind, since its export stops as soon as the wind stops, so it
only commands a low price as an unreliable source. This combined
with a failure of a transmission circuit, plunged much of central
Europe into darkness a couple of years ago when supply suddenly
fell well short of demand, and emergency load shedding was initiated.

 

[jmfbahciv]

Demand is always almost equal to capacity. If it would be more, it would be
a waste, if less there would be serious stability issues. That's the problem
with wind turbines, you just don't know when the wind blows, and you have to
cover each MW of wt with at least of 700 kW conventional reserve, because
when the wind stalls what? Stop everything?

Shit. I knew I wrote that badly. I'm talking about physical capacity.

/BAH

 

[[email protected]]

| In the UK we currently have cold winds and snow.
|
| A radio phone-in got a lot of people talking about flickering street
| lights.
|
| Is there a real connection between bad weather and flickering street
| lights. If so then how does it work?

 

[TKM]

In the UK we currently have cold winds and snow.

A radio phone-in got a lot of people talking about flickering street
lights.

Is there a real connection between bad weather and flickering street
lights. If so then how does it work?

What kind of lammps do you have, carbon arc, sodium vapor, mercury
vapor, incandescent, LEDs?

It makes a difference.

Vapor lamps are sensitive to the cold, as they require a narrow range
of pressure internally to function. Cold reduces the presssure,
making operation less reliable.

Cold can also shorten wires. If you have a break internally in an
insulated wire, cold could shrink the metal, weakening the connection
at the break and raising resistance, which heats the wire which
expands again, which drops the resistance, which lets the wire cool
down again, ad nauseum. That is the mechanism for flashers in
Christmas lights and automotive turn signals.

Tom Davidson
Richmond, VA

From a practical standpoint, these days the only lamps that are sensitive to
cold temperatures down to ambient temperatures of about 0 F. (-18 C) are
fluorescent lamps. Even so, special fluorescent lamps are used routinely
at -20F (-29 C) in commercial freezers, outdoor signs and exterior building
lighting. High pressure sodium, high pressure mercury, metal halide, LED,
cold-cathode ("neon") and, of course, incandescent all work just fine. Low
pressure sodium lamps have a "thermos bottle" construction so they may just
take a little longer to warm up. Discharge lamps do require somewhat
higher voltages to initiate the arc at low temperatures, so in applications
where cold temperatures are expected, so-called "minus 20" ballasts are
specified.

The only discharge lamps that I've ever seen flicker in cold temperatures
are CFLs and unprotected or reduced-wattage ("watt miser") fluorescent types
which have been mis-applied. Rarely, a high pressure mercury lamp will
flicker due to a faulty arc tube and there's an odd type of high pressure
sodium lamp designed to work on mercury ballasts that sometimes flickers
near end-of-life. I've heard that high pressure sodium lamps operated on
old series (constant-current) streetlighting systems will flicker; but, if
so, all of the lamps on the circuit will flicker, not just one. I've never
seen that phenomenon, however.

Terry McGowan

 

[Jeff Waymouth]

I am almost scared to jump in here, after the flaming has begun, sigh,
but one possibility has not been fully explored (although a previous
poster did come near to it). European power systems normally provide 50
Hz, instead of the US 60Hz, and, therefore, the roadway lighting systems
operate with that since electronic ballasts haven't penetrated that
market very much.

In my 20's and 30's, I could definitely see 50Hz flicker from HID
lighting (We had occasion to test European systems and used 50Hz power
to drive them) A previous poster noted that the snow on the ground
is a good reflective body and I think that characteristic could make the
50 Hz flicker more noticable.

Jeff Waymouth

 

[lurch]

MY CONCLUSION AFTER SCANNING THE KNOWN SPECTRUM OF STREET LIGHTS IS
THAT HIS LOCALE HAS VERY CHEAP KNOCKOFF STREET LIGHTS AND HE SHOULD BE
CONTACTING HIS COMMUNITY LEADERS TO BRING THE LAME ASS CONTRACTOR THAT
INSTALLED THEM TO JUSTICE

Where is that list of "knock off" makers at?

You're the idiot. That is about a silly claim.

"knock off" makers are not even in this channel. They are concerned
with bigger demographics. That is aside from the fact that you do not
even know how city planners go about choosing their street lights, much
less how they buy and install them.

Again, you're an idiot.

 

[Steve Firth]

In my 20's and 30's, I could definitely see 50Hz flicker

<Cough>obblers.

 

[Andrew Gabriel]

I'd no idea that these lights operated at such high temperatures - even
hotter that an indcandescent light. I tend to think of discharge bulbs as
being "cold" compared with incandescent, but evidently this is not always
the case!

Incandescent lamp filament is about 2700K, which is quite a bit hotter
than these arc tubes.

When you switch off a HID lamp (mercury, sodium, ...) the electrodes
and arc tube are glowing bright red hot until they cool down.

 

[lurch]

<Cough>obblers.

On the Toad again...

 

[lurch]

Incandescent lamp filament is about 2700K, which is quite a bit hotter
than these arc tubes.

The color temperature of the light emitted, and the operating
temperature of the bulb are two entirely different and unrelated things.
Light bulb nomenclature, when referring to a temperature measured in
Kelvins, is referring to the light color temp, not the operating temp of
the bulb or filament.

 

[Don Klipstein]

The color temperature of the light emitted, and the operating
temperature of the bulb are two entirely different and unrelated things.
Light bulb nomenclature, when referring to a temperature measured in
Kelvins, is referring to the light color temp, not the operating temp of
the bulb or filament.

With tungsten incandescent lamps, the filament temperature and the color
temperature of the emitted light are about the same. Color temperature
refers to the temperature of an ideal incandescent radiator (namely, a
"blackbody") producing light of the color temperature in question. The
color of light from incandescent tungsten is not much different from the
color of light from a blackbody at the same temperature.

- Don Klipstein ([email protected])

 

[lurch]

With tungsten incandescent lamps, the filament temperature and the color
temperature of the emitted light are about the same. Color temperature
refers to the temperature of an ideal incandescent radiator (namely, a
"blackbody") producing light of the color temperature in question. The
color of light from incandescent tungsten is not much different from the
color of light from a blackbody at the same temperature.

- Don Klipstein ([email protected])

I used to make a black body source that used a foot long calibrated
filament.

I used to make another that used coils of 'coiled' heater wire wrapped
around the ceramic 'wool' wrap and silica 'plaster of paris', then we
wrapped around the outside of that. The base form inside all of it was a
basketball. After the form set, we bake it in a kiln and after that,
with the actual excitation coils. Put a foot long ceramic feeder tube
onto that, and one "looks" into a perfect black body cavity.

Those were fun to slather together, and quite profitable too.

 

[Archimedes' Lever]

I used to make a black body source that used a foot long calibrated
filament.

The bulb the filament was in, not the actual filament. It was
vertical, and the excitation area was about an inch long and about 3mm
wide.

 

[Guest]

Most of the generators have variable power output, not simply on
or off.

--------
In no way did I imply otherwise. However a generator is either on line or
off line as a unit so that the system "capacity" changes in blocks depending
on the rating of the machine added or removed from the system. The system
"demand" depends on the load and this demand is split between the units on
line- generally through some economic dispatch scheme.

Frequency drops below nominal, and conversely when supply exceeds
demand, frequency increases above normal. There's a requirement in
the UK for frequency to average out correctly long-term (so things
like synchronous clocks don't drift long-term), consequently,
supply has to exactly match demand long-term. However, since the
demand and supply can't exactly track each other short term due to
inherent lags, there are periods of both demand exceeding supply,
and supply exceeding demand. These are both inevitable due to the
supply lag with different types of plant and unexpected plant
failures on some occasions, and deliberately forced to correct for
earlier drifts on other occasions.

--------------------
I am quite familiar with the concepts and practice involved.
-----------------

There is contingency reserve in addition to the supply - additional
plant spinning sychronous online ready when needed due to either an
increase in demand or an unexpected loss of supply, and yet more
plant offline ready to run up and cut in with a bit more notice.

------------------
Again I have no problem with this- this is normal . Possibly there are
some word usage problems

I take capacity as the total available generation on line That is,
(ignoring power factor as it affects unit capability), if there are 2- 100MW
units and a 50MW unit on line the capacity is 250MW . The load and losses
(again ignoring pf) may be 200MW leaving 50MW on-line reserve.
I take demand in this case as the load +losses 200MW which has to be
delivered to the system. Except for transient periods when loads change the
supply and demand are the same. Only during acceleration or deceleration
will they be unbalanced. Typically generator droops, essential to proper
load sharing between units, will result in speed changes. These lead to
frequency errors and the need to correct the long term average frequency.
-----------------------------------------

A longer article I wrote on this some years back, with some examples
of how it was applied in the UK to some specific historic events,
and how it went wrong on one occasion...
http://groups.google.co.uk/group/sci.engr.lighting/msg/99b03c9711a4f753?hl=en


The BBC did a programme about the wind power in Denmark, one of
the highest users of wind power. In spite of installing lots of
turbines and being able to point to all the power they get from them,
they haven't been able to spin down any conventional generating
station, because they need them when the wind stops. When the wind
blows, they have an excess of conventional electricity which they
sell, but for their neighbours, it's effectively as unreliable as
the wind, since its export stops as soon as the wind stops, so it
only commands a low price as an unreliable source. This combined
with a failure of a transmission circuit, plunged much of central
Europe into darkness a couple of years ago when supply suddenly
fell well short of demand, and emergency load shedding was initiated.

--------------------------

In my opinion, wind is to be used when available- reducing the load on other
sources at that time- but it doesn't replace the other sources for the good
reasons that you have given. The fact that wind energy is available on
nature's timetable, not man's, is one that many wind advocates appear to
ignore.

 

[Derek Geldard]

At the start of the cold period, my house lights and those of a friend,
both also in SE England, flickered a few times. I assumed the weather
was affecting overhead power lines somewhere, which resulted in
automatic switches having to transfer load.

Doesn't that sort of thing happen when intermittent shorting or arcing
trips a breaker somewhere and the breaker makes automatic re-tries ?

Derek.