Power supply inrush current

[Rene Lamontagne]

I am curious about approximately what is the value of this on a cold start.

What brings this up is seeing my ceiling fixture blink when I power on
the machine in the morning, The ceiling fixture is an LED 100 watt
equivalent drawing about 13 watts.

The computer has a Cooler Master M850 watt gold power supply and the
machine is drawing 0.5 amps in idle mode.

There being no way I can measure this without $20,000 worth of lab
equipment is there any way I can calculate or estimate this without a
schematic of the power supply input circuits?

This is just one of those food for thought questions to satisfy my
curiosity.

Rene

[Grant Taylor]

On 1/2/22 11:04 AM, Rene Lamontagne wrote:
> I am curious about approximately what  is the value of this on a cold
> start.

I would expect that this to be labeled on the power supply. At the very
least it should be within the maximum the power supply will draw.

> What brings this up is seeing my ceiling fixture blink when I power on
> the machine in the morning, The ceiling fixture is an LED 100 watt
> equivalent drawing about 13 watts.

I've found that many LED lights are /very/ /susceptible/ to any voltage
change on the line. As in a change from 120V to 119V will unbalance
many LED lights for a moment. Most of the LED lights that I've seen
tend to /reactively/ adjust their voltage regulation. So if the voltage
on the line changes at all, they have to react to it. This reaction
usually manifests itself as a flicker of some sort.

> There being no way I can measure this without $20,000 worth of lab
> equipment

Sure there is. You should be able to get a quite good idea with a $20
multi-meter that measures current. A $200 multi-meter will be even more
accurate and respond even faster.

Here's a hint. Your inrush current is less than the current required to
trip your circuit breaker. If you're in the U.S.A., that's /likely/ a
15A at 120V circuit.

Admittedly, an inrush current /may/ be short enough that the circuit
breaker isn't responding to the over current situation for a split
second. This is especially true for older circuit breakers and / or
circuit breakers. But almost all power strips that have overload
protection will pop /before/ the circuit breaker does. So if neither of
those are popping, your inrush current is almost definitely below the
circuit capacity.

Evil hint: You can play with this by purposely adding additional known
load to the circuit to get a rough approximation on what the computer is
using. E.g. if it's a 15A circuit and you have 10A of other load and
the circuit doesn't trip, then you have a really good idea that the
computer is not drawing more than 5A.

> is there any way I can calculate or estimate this without a schematic
> of the power supply input circuits?

As stated above, this should be labeled on any reputable power supply.

Contact the power supply manufacturer.

Use a $20 multi-meter.

> This is just one of those food for thought questions to satisfy my
> curiosity.

Hopefully this helps scratch that itch.



--
Grant. . . .
unix || die

[Paul]

Even if you knew the value, the info provided does not document
the duration. The Cooler Master site gives scant info (almost as
if the PSUs are contract manufactured).

Next, you find a review.

https://www.tomshardware.com/reviews/cooler-master-v850-gold-v2-power-supply-review

Inrush Protection NTC Thermistor 16D-20 (16 Ohm) & Relay

So it probably isn't the 40A or 80A value quoted for some PSUs.
The 16 ohms would be the "cold" value, before the thermistor
heats up.

The only time I've heard of inrush being an issue, is a
review unit at Anandtech tripped the breaker, every time
it was turned on. That means the duration of the inrush
was sufficient to trigger the lab breaker. Most of the
time, that doesn't happen, even if the quoted figure
is ridiculously high.

And when they quote a high figure, I've seen enough transient
behavior here to know some aspect of it is realistic.

If a LED lamp is dimmable, perhaps the inrush transient is
confusing the dimming, rather than the rails flattening
enough for the LED to stop putting out light. A cap dropper
LED lamp, you might see an inrush transient for one of those, but
they use that for LED lighting devices at 3W or less perhaps. Not
your 13W quoted value. A tiny SMPS is used for 13W ones.
It should regulate the output, and practically no short term
event should affect the light.

Paul

[Paul]

On 1/2/2022 1:17 PM, Grant Taylor wrote:
> On 1/2/22 11:04 AM, Rene Lamontagne wrote:
>> I am curious about approximately what  is the value of this on a cold start.
>
> I would expect that this to be labeled on the power supply.  At the very least it should be within the maximum the power supply will draw.
>
>> What brings this up is seeing my ceiling fixture blink when I power on the machine in the morning, The ceiling fixture is an LED 100 watt equivalent drawing about 13 watts.
>
> I've found that many LED lights are /very/ /susceptible/ to any voltage change on the line.  As in a change from 120V to 119V will unbalance many LED lights for a moment.  Most of the LED lights that I've seen tend to /reactively/ adjust their voltage regulation.  So if the voltage on the line changes at all, they have to react to it.  This reaction usually manifests itself as a flicker of some sort.
>
>> There being no way I can measure this without $20,000 worth of lab equipment
>
> Sure there is.  You should be able to get a quite good idea with a $20 multi-meter that measures current.  A $200 multi-meter will be even more accurate and respond even faster.
>

The transient is the charging current going into the main capacitor
after the primary bridge rectifier. The main capacitor charges to
voltages anywhere between 300VDC and 500VDC. In some cases, the
Active PFC circuit causes voltages higher than normal to be
used with a main capacitor.

The main capacitor on the ATX PSU, is set to provide holdup time
so that the unit works properly with a SPS flavor of UPS. The
main capacitor holds many joules and you'll lose an eardrum if
it explodes. As capacitors go, the main capacitors seems to be
"well fabricated". The capacitor plague years ago, mains capacitors
were not affected. None of those leaked. The lawsuits would have
been endless, if bad fabrication techniques were used.

I put my clamp-on ammeter on "peak hold", and flipped the switch
a couple times, allowing cooling time between flips (a couple minutes)
and I could not record the transient. You'd need one of those $1000 HP
meters that records at 1000 readings per second, to record the transient.
Ordinary multimeters, dual slope integrators, they only take one or two
readings a second, and are hopeless material for transient analysis.

A Kill-O-Watt meter, the measurement circuit in it could
easily capture the whole transient (the sigma delta converter
runs at 500KHz, the circuit "bandwidth" is somewhat less). But
the software is not set up for the necessary extraction and
presentation of what happened. The only reason I
mention the Kill-O-Watt, is to show that good instrumentation
doesn't have to cost a lot of money.

It's only the old dual-slope integrating multimeters, that
suck at making measurements. You can get flash converters with
amazing circuit specifications, but the price may be prohibitive
for making a $20 multimeter. Some of the high sample rate converters
have 24 bit output. That would be a fair number of digits on
a multimeter.

A hobbyist with a lot of time on their hands, could make
this measurement :-) For the rest of us, the LED in our
room will be flashing.

Paul

[Rene Lamontagne]

Seeing inrush current only lasts for some 4 or 5 cycles or about 100 ms
you need more up to date equipment than mine, I have an older Fluke
digital multimeter, a Radioshack autoranging DMM, a RadiosSack analog
clamp on AC meter and a Kill-O-Watt meter none of which have an inrush
or peak hold feature.

So with that I will guess at 4 to 10 times running current and attribute
the flicker to a minor voltage drop at the lamp triggering a flicker due
to its inherent features, An incandescent or fluorescent lamp would
probably not be affected.

Rene

[Paul]

On 1/2/2022 2:47 PM, Rene Lamontagne wrote:

> Seeing inrush current only lasts for some 4 or 5 cycles or about 100 ms you need more up to date equipment than mine, I have an older Fluke digital  multimeter, a Radioshack autoranging DMM, a RadiosSack analog clamp on AC meter and a Kill-O-Watt meter none of which have an inrush or peak hold feature.
>
> So with that I will guess at 4 to 10 times running current and attribute the flicker to a minor voltage drop at the lamp triggering a flicker due to its inherent features, An incandescent or fluorescent lamp would probably not be affected.
>
> Rene
>

You can see the flicker on incandescent.

Paul

[Rene Lamontagne]

As a true scientist I had to test this flicker further with an
incandescent lamp

With the 17 working bulbs in my 2 bedroom apartment I could not find 1
incandescent bulb, I had changed them all to LEDs.

So out of desperation I had to get down on my knees and unscrew the 25
watt incandescent lamp from my refrigerator and install it in a desk
lamp and plug iit in near my computer.

And the occlusion is (drumbeat) !!! yes Paul, it do flicker.:-)

Rene

[Rene Lamontagne]

Stupid spellchecker, doesn't know how to spell *conclusion*. sheesh

Rene

[Sjouke Burry]

Well..... My fluorescent lamps in my kitchen sure show my magnetron
cycling , Any light source converting without delay should show a dip ,

[Mark Lloyd]

On 1/2/22 14:46, Rene Lamontagne wrote:

[snip]

> As a true scientist I had to test this flicker further with an
> incandescent lamp
>
> With the 17 working bulbs in my 2 bedroom apartment I could not find 1
> incandescent bulb, I had changed them all to LEDs.

The only incandescent bulb I'm using in my house is in the oven. I do
have a few fluorescent lights (which will be replaced with LEDs).

The fluorescent behind me will flicker when I print something. The
printer is a laser that's in the next room but on the same electrical
circuit (these are all 20A breakers).

> So out of desperation I had to get down on my knees and unscrew the 25
> watt incandescent lamp from my refrigerator and install it in a desk
> lamp and plug iit in near my computer.

I've been using LEDs in my refrigerator and microwave for several years
without any problems. Only the oven requires incandescent.

> And the occlusion is (drumbeat) !!!  yes Paul, it do flicker.:-)
>
> Rene

--
Mark Lloyd
http://notstupid.us/

"[Fundamentalists] never wonder why, if herpes is sent by 'god' to
scourge "adulterers," whooping cough and measles weren't purposely
created to lambaste children." [Fred Woodworth]

[Rene Lamontagne]

On 2022-01-02 3:09 p.m., Mark Lloyd wrote:
> On 1/2/22 14:46, Rene Lamontagne wrote:
>
> [snip]
>
>> As a true scientist I had to test this flicker further with an
>> incandescent lamp
>>
>> With the 17 working bulbs in my 2 bedroom apartment I could not find 1
>> incandescent bulb, I had changed them all to LEDs.
>
> The only incandescent bulb I'm using in my house is in the oven. I do
> have a few fluorescent lights (which will be replaced with LEDs).
>
> The fluorescent behind me will flicker when I print something. The
> printer is a laser that's in the next room but on the same electrical
> circuit (these are all 20A breakers).
>
>> So out of desperation I had to get down on my knees and unscrew the 25
>> watt incandescent lamp from my refrigerator and install it in a desk
>> lamp and plug iit in near my computer.
>
> I've been using LEDs in my refrigerator and microwave for several years
> without any problems. Only the oven requires incandescent.
>
>> And the occlusion is (drumbeat) !!!  yes Paul, it do flicker.:-)
>>
>> Rene
>

This fridge came with the apartment so never had to change it yet, Never
thought of the oven one.

Rene

[VanguardLH]

Rene Lamontagne <rla...@shaw.ca> wrote:

> I am curious about approximately what is the value of this on a cold start.
>
> What brings this up is seeing my ceiling fixture blink when I power on
> the machine in the morning, The ceiling fixture is an LED 100 watt
> equivalent drawing about 13 watts.

What ELSE is on the same circuit? The light might be noticeable because
you can see it, but other electrics might be on and placing a load on
the same circuit. You won't see a toaster blink or dim, and your
television might compensate for brownouts. The large caps in a stereo
power supply won't let you hear blankouts or other artifacts for a very
short brownout.

> There being no way I can measure this without $20,000 worth of lab
> equipment is there any way I can calculate or estimate this without a
> schematic of the power supply input circuits?

Only takes an ammeter that you put inline. Because of A/C current, you
can't just wrap a sensor around the 2 lines (hot and neutral) to measure
current as they will cancel each other. You could split the cord, or
buy a cord splitter, so the ammeter can be placed around just one wire.
There are also "splitters" that you plug into the outlet, you plug the
power cord into the splitter, and the ammeter clamps around just one
half of the splitter.

An analog ammeter (I've had those) will be too slow to measure inrush
current, and it'll bounce around a lot when the current is suddenly
high. There are ammeters that will measure peak current, and hold that
reading, and they don't cost thousands. The ones I recently bought
(last October at Home Depot) were:

2,000 Amp Digital Clamp Meter
https://www.homedepot.com/p/Klein-Tools-2-000-Amp-Digital-Clamp-Meter-CL900/304423774

Line Splitter
https://www.homedepot.com/p/Klein-Tools-Line-Splitter-69409/206799961

The ammeter can measure inrush current. The splitter eliminates me
having somehow separate the power cord wires to clamp around just one of
them. I used to have a multi-point line splitter: it had multiple
openings at different magnification allowing to measure small to very
large currents without a loss in granularity in measurement (reading a
10mA current at the 2000A scale isn't going to work).

https://www.homedepot.com/p/Combination-AC-Line-Spliter-and-GFCI-Receptacle-Tester-ET200/315142123?MERCH=REC-_-searchViewed-_-NA-_-315142123-_-N

That one has X1 and X10 positions. The old one I had (can't find it)
had 5 multiplier positions. I can find several at Amazon:

https://www.amazon.com/s?k=ac+line+splitter&crid=2W5X2KFUAGWX0&sprefix=ac+line+splitter%2Caps%2C91

but they only have the X1 and X10 positions. I don't need yet another
GFI tester making the line splitter longer, so I'll probably get:

https://www.menards.com/main/electrical/electrical-tools-accessories/electrical-hand-tools/sperry-instruments-commercial-line-splitter/e1/p-1444431011716-c-6455.htm?tid=-8194613577563981708&ipos=1

[VanguardLH]

Oh, have you yet replaced the LED light bulb? Is the LED on a dimmer
circuit? When they're about to fail, they often flicker at any little
change in line voltage. LEDs are on or off. There is no persistence as
with incandescents.

https://ledlightinginfo.com/why-do-my-led-lights-flicker
and more at:
https://www.google.com/search?client=firefox-b-1-d&q=led+light+flicker

[VanguardLH]

Rene Lamontagne <rla...@shaw.ca> wrote:

> Stupid spellchecker, doesn't know how to spell *conclusion*. sheesh

Grammar checker might catch that, might not, but not a spell checker.
Occlusion is a valid word; see:

https://www.dictionary.com/browse/occlusion.

Sometimes your fingers don't obey what you tell them to do. And
sometimes your eyes, like when proofreading your own work, see what you
expect to see.

[Rene Lamontagne]

Ain't that the truth. :-)

Rene

[Rene Lamontagne]

On 2022-01-02 4:15 p.m., VanguardLH wrote:
> Rene Lamontagne <rla...@shaw.ca> wrote:
>
>> I am curious about approximately what is the value of this on a cold start.
>>
>> What brings this up is seeing my ceiling fixture blink when I power on
>> the machine in the morning, The ceiling fixture is an LED 100 watt
>> equivalent drawing about 13 watts.
>
> What ELSE is on the same circuit? The light might be noticeable because
> you can see it, but other electrics might be on and placing a load on
> the same circuit. You won't see a toaster blink or dim, and your
> television might compensate for brownouts. The large caps in a stereo
> power supply won't let you hear blankouts or other artifacts for a very
> short brownout.
>

Only the computer, monitor, and ceiling light

>> There being no way I can measure this without $20,000 worth of lab
>> equipment is there any way I can calculate or estimate this without a
>> schematic of the power supply input circuits?
>
> Only takes an ammeter that you put inline. Because of A/C current, you
> can't just wrap a sensor around the 2 lines (hot and neutral) to measure
> current as they will cancel each other. You could split the cord, or
> buy a cord splitter, so the ammeter can be placed around just one wire.
> There are also "splitters" that you plug into the outlet, you plug the
> power cord into the splitter, and the ammeter clamps around just one
> half of the splitter.
>
> An analog ammeter (I've had those) will be too slow to measure inrush
> current, and it'll bounce around a lot when the current is suddenly
> high. There are ammeters that will measure peak current, and hold that
> reading, and they don't cost thousands. The ones I recently bought
> (last October at Home Depot) were:

Yeah, my analog clampon is no good for this job.

>
> 2,000 Amp Digital Clamp Meter
> https://www.homedepot.com/p/Klein-Tools-2-000-Amp-Digital-Clamp-Meter-CL900/304423774

Great brand of tools.

>
> Line Splitter
> https://www.homedepot.com/p/Klein-Tools-Line-Splitter-69409/206799961
>
> The ammeter can measure inrush current. The splitter eliminates me
> having somehow separate the power cord wires to clamp around just one of
> them. I used to have a multi-point line splitter: it had multiple
> openings at different magnification allowing to measure small to very
> large currents without a loss in granularity in measurement (reading a
> 10mA current at the 2000A scale isn't going to work).

n

>
> https://www.homedepot.com/p/Combination-AC-Line-Spliter-and-GFCI-Receptacle-Tester-ET200/315142123?MERCH=REC-_-searchViewed-_-NA-_-315142123-_-N
>
> That one has X1 and X10 positions. The old one I had (can't find it)
> had 5 multiplier positions. I can find several at Amazon:
>
> https://www.amazon.com/s?k=ac+line+splitter&crid=2W5X2KFUAGWX0&sprefix=ac+line+splitter%2Caps%2C91
>
> but they only have the X1 and X10 positions. I don't need yet another
> GFI tester making the line splitter longer, so I'll probably get:
>
> https://www.menards.com/main/electrical/electrical-tools-accessories/electrical-hand-tools/sperry-instruments-commercial-line-splitter/e1/p-1444431011716-c-6455.htm?tid=-8194613577563981708&ipos=1
>

I have an X1, X10 splitter for my 400 amp analog clamp on for other jobs.

Rene

[Bob F]

And sometimes, the send button is faster than your eyes.

[Rene Lamontagne]

New light bulb, not on a dimmer circuit.

It only does 1 blip when I turn the computer on in the morning, so not a
problem at all.

Yep, I have one in the kitchen that's about to perish, it flickers for a
second or two when I turn it on.,

One thing I notice is that these LED bulbs are not lasting as long as
touted by the manufacturers.

Rene

[VanguardLH]

Rene Lamontagne <rla...@shaw.ca> wrote:

> VanguardLH wrote:
>
>> Rene Lamontagne <rla...@shaw.ca> wrote:
>>
>>> I am curious about approximately what is the value of this on a cold start.
>>>
>>> What brings this up is seeing my ceiling fixture blink when I power on
>>> the machine in the morning, The ceiling fixture is an LED 100 watt
>>> equivalent drawing about 13 watts.
>>
>> What ELSE is on the same circuit?
>

> Only the computer, monitor, and ceiling light

A circuit could have 15 loads (lights, wall outlets, etc). The maximum
endpoints are determined by building code. Some places put a limit of
15 outlets. Some only go by the following 80% OBCD rule for circuit
breakers. I saw one house that had 30 loads on a circuits, and violated
building code in that area.

Also, circuit breakers can handle a sustained current of only 80% of
their rating. A 15A breaker can handle 12A. A 20A breaker can handle
16A. See:

https://www.ecmweb.com/basics/article/20898940/sizing-a-circuit-breaker

If you flip the breaker for the circuit with your computer and LED
light, nothing else goes off? All the other wall outlets are unused?
It would be very unusual for an entire circuit to get dedicated for just
a single wall outlet used by the computer and a single light fixture (or
1 wall outlet shared by computer and table lamp). I'm not asking how
many devices are plugged into an outlet, but what ELSE is on the same
circuit. I've yet to see an entire circuit dedicated to a single wall
outlet except for high-amperage circuits (e.g., electric range, electric
washer/dryer setup that are 240VAC split across the neutral to provide
higher amps at 120VAC, and why you find, say, ganged 20A breakers for a
40A load).

[Rene Lamontagne]

Each bedroom which is where my computer system is has 3 duplex outlets
and 1 ceiling fixture supplied by one 15 amp 120 volt dedicated breaker.
I am only using 1 duplex outlet, the other 2 are unused. both bedrooms
are the same each has one 15 amp breaker, I believe here in Canada you
are only allwed 4 outlets per breaker.
Likewise the living room has 3 duplex outlets plus 1 ceiling fixture on
1, 15 amp breaker, PLUS 1, 15 amp single dedicated outlet to 1 15 amp
breaker for the through the wall A/C unit

The kitchen has 3 duplex outlets of which 1 is a GFCI outlet and a 2
pole 240 v0lt one for the electric range.
The bathroom with 1 GFC! outlet and 1 over the mirror light array share
a breaker with the 2 hallway lights.

Rene


Rene

[VanguardLH]

Rene Lamontagne <rla...@shaw.ca> wrote:

> One thing I notice is that these LED bulbs are not lasting as long as
> touted by the manufacturers.

Be careful when reading their longevity claim. You'll notice they'll
say the LED bulb lasts for 10 years -- BUT only when used for 3 hours
per day. They make longevity look huge, like saying it has 25,000 hour
maintenance-free lifespan. That's under 3 years of continuous use.
They'll claim their 25,000 hour LED bulb lasts 11 times longer than
incandescents. That means the incandescents were lasting only 3 months,
and that certainly is a lie.

[VanguardLH]

Well, other than going by any published specs on the computer's PSU, if
there some with inrush surge listed, you'll need to get a newer ammeter
and line splitter to test the inrush.

With a motor starting up which looks like a short consisting of lots of
wire, there is a large inrush load. I don't see a computer PSU having
much of an inrush over its maximum rated capacity, and then divided by
whatever is its efficiency; i.e., for your Cooler Master M850 PSU, with
its 90% efficiency (initial rating, not after several years as
efficiency wanes), its 850W capacity would draw 850W/0.9 = 944W, or
about 8A to 9A. On a 15A circuit, that's 54% of the circuit capacity.
As the article mentions a 15A breaker actually handles a 12A load, so
turning on the computer consumes 68% of the circuit capacity.

That's assuming your PSU's output is max'ed out. You could get a
kill-a-watt meter to see what is the actual load by your computer. The
meter won't show inrush current, just the sustained power load.

Switching power supplies usually incorporate inrush current, like using
a resistor (passive) or multiple resistors, switching device, and
control circuit (active). The resistor affect the RC charging curve of
the capacitors to flatten the inrush curve. Instead of a simple
resistor for passive limitation, a negative temperature coefficient
(NTC) resistor, or thermistor is used that has high resistance when
cold, and low resistance when warmed. Inrush current will be low
through the thermistor in series with the power line (after the diode
bridge), and gone by the time the thermistor warms up with the current
passing through it. The inrush is throttled at the start, not as high,
but takes longer to charge the capacitors. A switch can be used in
parallel to zero out (bypass) the thermistor when resistor gets low as
it warms. Active inrush control can consist of resistors, triacs,
thyristors, transistors, and a switching device, and there are way too
many designs to get into here.

The point is current regulation can easily control inrush current. This
is for switchmode PSUs used by computers. They don't use linear PSUs
with huge transformers that look like a short of a very long length of
wire. Whether or not inrush is regulated depends on the quality of
design. I've seen 300W PSUs that have a an inrush rating of <45A, but
that was for a 70% non-activePFC, and inrush is measured peak to peak,
not peak to baseline. That's huge on a 15A circuit, but is very
momentarily, like 3 cycles or 0.05 seconds (check the hold-up time which
is likely around 0.016 seconds), so the breaker doesn't warm up enough
to make it snap open yet I've heard about huge setups with huge PSUs
that tripped a breaker. Maybe that depends on the type of breaker, like
a standard breaker relying on heat to trip versus AFCI breakers designed
to detect arcs in a circuit (which high inrush could emulate). AFCI
(Arc Fault Circuit Interrupter) breakers use electronics to monitor the
line to detect multiple types of faults on the line (whether an actual
arc or something that looks like one), and are often required in new
construction per NEC (National Electrical Code); see
https://www.afcisafety.org/afci/what-is-afci/#1469600217909-520261d7-4cb5.
Due to how LED light bulbs work, that momentary overload could make them
flicker. I suspect the inrush period doesn't last longer than the
hold-up rating. Cooler Master says that PSU has >16ms (so <17ms).
There is no persistence in illumination for LEDs, nor are they designed
to sustain illumination when voltage drops even if only for an instant.

https://www.superiorlighting.com/lighting-resources/light-bulb-learning-center/led-light-bulbs/led-lamp-components-explained/
"An LED driver - regulates the current flowing through the LED, similar
to a ballast in a compact fluorescent lights.LED drivers can be internal
or external. LED Light output is proportional to its current; any slight
variation in the current can result in unacceptable changes in light
output."
("unacceptable" could be viewed as "perceptible", but a flickering LED
has a shorter life.)

For your PSU:

https://www.coolermaster.com/catalog/power-supplies/mwe-series/mwe-gold-850-v2-full-modular/#specifications

That doesn't show the inrush current for that PSU. Not really
surprising since Cooler Master doesn't manufacture anything. Same for
Corsair, Crucial, and many brands. They contract with someone else to
do the manufacture accord to their specs, and slap their name on the
product. Think of wanting to have a cabinet built to your specs, but
you give the blueprints to a carpenter.

This PSU has active PFC (Power Factor Correction). Because the PSU is
changing its frequency, it's possible that feedback into the line hits
the same used in the regulator inside the LED light bulb.

https://www.nuvation.com/resources/article/power-supply-design-basics-active-power-factor-correction
"Requires better filtering to prevent high frequency hash from getting
to the line." For the <17ms hold-up time (and probably close to the
inrush time), and because LED light bulbs are nothing more than a bridge
diode with an inline capacitor on input, and a capacitor, resistor, and
LED bulb(s) on the output, the RC time is very short, so 16ms could be
more than sufficient a drop in voltage to affect the LED bulbs. The
capacitors are noise filters, not for capacity.

Use the Kill-a-Watt meter to see what is the sustained load consumed by
your computer's PSU. No, don't test when idle. Obviously the computer
is not idle when you do a cold boot. Keep the hardware as busy as
possible, like running Prime95, during which you measure the power or
amps consumed by the PSU. A Kill-A-Watt meter won't measure inrush
current, but you'll get a better idea of what is the actual load drawn
by your PSU from the line. As worst case, multiple by 2 for the inrush
current with the PSU under full load (or whatever maximum load you can
put on the PSU during testing). It won't cost $20,000 USD to measure
inrush current, but more like $160 USD by getting a better ammeter and
line splitter. If you don't get the tools, you may never know the
inrush current your PSU puts on the line.

You mentioned using a fridge light bulb, but didn't mention whether it
was incandescent or LED. There are LED appliance light bulbs. You
might've replaced an LED with another LED. While getting more rare, you
can still find an incandescent to buy to test when powering up your
computer from cold.

[Rene Lamontagne]

First, thanks for taking the time for this very good explanatory reply,
Answers a lot of questions.

With my Kill-A-WattIiread
!59 watts at full 100% Prime 95 load
52 watts at idle

Yes, it was an incandescent filament type bulb from the fridge.

At 87 years old I probably will not invest in any more test equipment as
I live in an apartment and really wouldn't have much further use for
such gear.

Thanks again.

Rene

[VanguardLH]

Rene Lamontagne <rla...@shaw.ca> wrote:

> I live in an apartment.

The electric service entry isn't inside your apartment. It's probably
in a utility room on the bottom floor of the apartment complex. I'm
guessing your suite is on a feeder: main breaker panel for building via
a feeder line to a sub-panel breaker box in your suite which then has
branch circuits (lines from the sub-panel in your suite with its
overcurrent protection to the wall outlets, fridge, electric range, etc)
within your suite.

https://www.jadelearning.com/wp-content/uploads/2018/02/WHAT.IS_.FEEDER.BLOG_.RVSD_.IMAGE_.JD_-e1632480897841.png

You may not be overloading the branch circuit for the room with your
computer and light, but all your suite loads might be close to
overloading your feeder line. I've seen DIY'ers really fuck up stuff.
The feeder should go back to its own breaker back in the main panel, but
it's possible to tap off a feeder for another apartment. Just because
an electrician did the layout doesn't mean they got a permit and their
job got inspected, and too often the inspectors never check the work, or
are too dumb to understand. I've seen boobs put the luminaires (lights)
on the same breaker as the wall outlets, so if someone blows the circuit
the wall outlets go dead, but you're also left in the dark. Plug
something into a wall outlet, zap goes the breaker, and now you're in
the dark, too. Having lights and outlets on the same circuit is how my
60-year old home is wired. It won't stay that way when I get the place
rewired. Building codes change, and in area it became against code to
put lights and wall outlets on the same circuit. I've seen sump pumps
put on the same circuit as wall outlets, so if the wall circuit blows
due to a wall outlet overload then so does the sump pump go dead which
means your basement floods during heavy rains when you're away on
vacation (unless you add a backup power supply to the sump pump, but
those last maybe 3 days while you're on vacation for 2 weeks). The only
control you have over the loads in your suite is to the sub-panel. I
don't know enough code to know if a feeder can be tapped to supply
multiple apartments.

I've seen discussions where folks have asked if tapping multiple units
from a common feeder is against code. Apparently it is not illegal in
many areas. A 60A feeder could pass through a 1st floor unit to
terminate in a 2nd floor unit, like when converting apartments to
condos. This means multiple units are sharing the feeder and, to me,
that's like changing a feeder into a branch where the rooms are units
(apartments). With tapping or sharing of a feeder, someone else could
be heavily loading the same circuit you are one, like running an
electric dryer, dishwasher, hair dryer, and so on.

If the lights flicker just once when you cold boot the computer(how
often is that?), and just in that one room, I wouldn't bother worrying
unless you have other gear sensitive to power fluctuations (which you
say there is nothing other than the computer and light). The flickering
will diminish longevity of the LED light bulb, but those last nowhere
the claimed lifespan, anyway.

Since there's nothing you can do, and you don't want to buy better
electrical tools, just ignore it. Replace the LED bulb when it dies in
2 years instead of 3.

[Mark Lloyd]

and I've had self-clicking touchpads, that would "click" when I just
moved the pointer over "send" (maybe trying to get to the 'file' menu).

--
Mark Lloyd
http://notstupid.us/

"Why is it when we talk to God we're said to be praying, but when God
talks to us we're schizophrenic?" -- Lily Tomlin

[Mark Lloyd]

On 1/2/22 16:56, Rene Lamontagne wrote:

[snip]

> One thing I notice is that these LED bulbs are not lasting as long as
> touted by the manufacturers.
>
> Rene

There seems to be a big variation in bulbs. About 4 years ago I replaced
the 6 light bulbs in my refrigerator/freezer with LEDs (all same brand,
replaced at the same time). 1 went out in less than a month and the
others are still working.

[Rene Lamontagne]

On 2022-01-03 2:44 a.m., VanguardLH wrote:
> Rene Lamontagne <rla...@shaw.ca> wrote:
>
>> I live in an apartment.
>
> The electric service entry isn't inside your apartment. It's probably
> in a utility room on the bottom floor of the apartment complex. I'm
> guessing your suite is on a feeder: main breaker panel for building via
> a feeder line to a sub-panel breaker box in your suite which then has
> branch circuits (lines from the sub-panel in your suite with its
> overcurrent protection to the wall outlets, fridge, electric range, etc)
> within your suite.
>

Just so you know what I am about, till I retired in 2000 I was a Power
Engineer 3rd class and held an electricians ticket for up to 600 volts.

Our apartment building is 11 stories with 10 suites per floor
The electric service is in the basement electrical room consisting of 4
x 1200 amp air breakers, each feeding 4 distribution panels on floors 3,
5, 8, and 11, each panel has 30 odd 100 amp breakers one for each suite
plus some for common areas.

In each suite is a 16 circuit, 100 amp load center panel, of which 8
slots are occupied., In my case 1 for each room plus 1 for the A/C and1
x 2 pole for the range.

I just took inventory of my running loads and here is what I have at
this moment

68 watts 2 x 34 F8 fluorescent tubes kitchen
7 watts, Range standby kitchen
37 watts, octopus lamp, 5 LED's
19 watts -, TV , set top box, DVD player phone answer machine standby.
6 watts, Weather station
18 watts, 2 hall lights
13 watts, bedroom light
13 watts, bedroom light
90 watts, my computer
82 watts, Son's computer
21 watts, router

for a total of 374 watts or roughly 3.11 which is peanuts on a 100 amp
panel.
My monthly consumption is usually 400 kWh or less

>

> Since there's nothing you can do, and you don't want to buy better
> electrical tools, just ignore it. Replace the LED bulb when it dies in
> 2 years instead of 3.
>

Of course it is not a worry, only an attempt to learn a little more
about a common occurrence within my environment. :-)

Rene

[Rene Lamontagne]

Should be 4 x 4000 amp air breakers
should read 3.11 amps total
I should proofread more.

Rene

[VanguardLH]

Rene Lamontagne <rla...@shaw.ca> wrote:

> for a total of 374 watts or roughly 3.11 which is peanuts on a 100 amp
> panel.
> My monthly consumption is usually 400 kWh or less

Having a 100-amp sub-panel or 100-amp master breaker doesn't mean you
have 100A service to your suite, only that you could. They're running
3-gauge wire (~0.25") to your sub-panel (assuming 75 C is acceptable)?
Maybe. Ampacity is usually by unit size; e.g., 40A to a 1-bedroom unit,
80A to a 3-bedroom unit. It's also possible to up ampacity to one unit
while shorting the max amps on another unit. There's also the 80% rule
on breakers and panels: a 100A panel supplies 80A, a 20A breaker handles
16A.

11 stories X 10 suites/story X 100-amp/suite = 11,000 amps, and more for
the common area lighting, like hallways and entrances, and common rooms,
like laundry room, and outside, like parking, and grounds. That's if
they were all independent or isolated. Breakers don't provide
isolation, just over-current protection. I haven't worked on electrical
for apartment complexes. Is 11,000 amps considered low, normal, or high
for those buildings? In the USA for large and medium buildings, a
utility-owned step-down transformer provides 13.8kV down to 480/277 VAC
to stepgear to a feeder to more transformers down to 120VAC and then to
branch circuits. I'm not sure that's the setup for a 110-unit apartment
complex. I doubt you have a step-down transformer alongside your
sub-panel in your suite.

Getting away from the spikes in current distribution in your suite when
you turn on your computer, there are other more local possibilities.
Some LEDs have PFC (power factor correction) to up their efficiency.
Some don't. PFC can induce noise on the line which can affect other
devices. I've seen this interference when using powerline remote
control. If the powerline remote lets you change its frequency (both
transmitter and receiver), they can operate at a different frequency to
avoid the interference with other PFC gear. The PSU in your computer
uses PFC. You probably don't have an oscilloscope, and one that logs,
to watch the harmonics or noise induced by your computer's PSU on the
power line when it is cold booted. LEDs with PFC put more noise on the
line. Without sufficient EMI filtering, switch-mode power supplies
induce noise that affect other electronics: computer PSUs, battery
chargers, electronic fluorescent ballasts, etc. Similar happens for
microwaves that interfere with cordless phones or wifi devices, because
they operate at the same 2.4GHz unlicensed band.

I remember a long time ago I needed an EMI/RFI filter that plugged into
the wall, but it's been too long to remember between which devices I
needed to eliminate the interference. Some power strips have an EMI
filter. You wouldn't both the computer and LED light (like for a table
lamp) on the same EMI power strip. You want them on opposite sides of
the filter. If you have the LED bulb in a table lamp, or in something
that plugs into the wall outlet, use the EMI power/surge strip on that,
and leave your computer plugged into the wall outlet. If the LED bulb
is in a permanent fixture, you can't add an EMI power/surge strip to it,
so try the strip on the computer's power.

[Rene Lamontagne]

12;000 amps would be normal for a 110 unit building, the utility
underground transformer in this building supplies 120/240 volts directly
to the 4 sub distributions.
I mostly worked in large commercial buildings such as the Concert
hall, Museum, Planetarium and Convention Centre, I was never really
involved in residential apartment buildings.

When I sold my old 12 room home some 7 years ago I had to downsize
massively, Unfortunately most of my electronic test gear including all
my various test generators, bench meters of all types and 2
oscilloscopes had to go. :-(

Rene

[VanguardLH]

Rene Lamontagne <rla...@shaw.ca> wrote:

> When I sold my old 12 room home some 7 years ago I had to downsize
> massively, Unfortunately most of my electronic test gear including all
> my various test generators, bench meters of all types and 2
> oscilloscopes had to go. :-(

I'd try the surge strip with EMI/RFI filtering. They aren't expensive.
Just be sure everything for the computer, and connected to the computer,
is upstream of the one surge protector, not across wall outlets (bare or
on separate surge protectors on the same of different wall outlets).
Even if the power strip didn't help (which reflects on poor EMI
filtering), you could use it for something else, or wrap it up as a
present given to someone else (with lots of electronics).