7

In his answer to Will repeatedly turning a light bulb on and off damage it? @Bradicul stated,

An example of inrush current is an LED downlight fitting with 9 W (0.0375 A at 240 V) will have any average inrush current of 7 A for 300 ms.

enter image description here

Figure 1. The 220 - 240 V, 9 W Philips downlighter has an overall diameter of 84 mm.

I thought he was mistaken but he linked to the Philips DN135B LED6S/830 PSR-E II WH datasheet and it states that the inrush for this 9 W lamp is 7 A for up to 300 ms.

Can anyone think of a reason the PSU in these luminaires could draw 1.75 kVA for up to 300 ms when the lamp is only 9 W with a 0.9 power factor? Where is the juice going?

Transistor
  • 168,990
  • 12
  • 186
  • 385
  • 3
    Datasheets have occasionally been found to be apocryphal. –  Sep 22 '18 at 17:10
  • 2
    Into a DC storage cap? – rdtsc Sep 22 '18 at 17:12
  • Worst case scenario + tons of margin? – winny Sep 22 '18 at 17:14
  • 1
    Doesn't make much sense, a reasonable size bulk cap of say 440uF wouldn't hold nearly that much energy. Maybe as Winny said they just picked a number so large that they could be sure they would always be below it. Maybe they measured the absolute peak current at 7A and found that startup takes 300ms, so to be safe they said 7A for the entire startup period. – John D Sep 22 '18 at 17:17
  • That data sheet has been inaccurately quoted. **The word "average" does not appear in the actual data sheet**. So that may very well be a peak current, and an overall duration. – Chris Stratton Sep 22 '18 at 17:32
  • 1
    Well, for one, it may be 7A peak, and 300ms max, but it may be a spike. So calculations based on multiplying 7A*240V*0.3s may be dramatically over-estimating the energy storage. – user57037 Sep 22 '18 at 17:33
  • Apparently @Bradicul misspoke, there is no such thing as "average inrush current", there is "peak inrush current", and inrush current is always "a spike". But there could be "average inrush charge", the quantity which is specified for USB for example. – Ale..chenski Sep 22 '18 at 18:24
  • I am the OP and even though I couldn't believe it myself, I was actually shocked when I walked around numerous sites with designers and engineers from Phillips, Osram, Trend etc. and the finer details were "revealed". As stated in the original post, particularly with automation systems, this can be a bit of an issue over time where most relays have a max loading of 10A "lighting" and can compromise EEPROM interigity when monitored through BMS systems. Thanks for posting your input on this because I dont think it gets looked at as much as it should in modern construction and fit-out stages. – Bradicul Sep 22 '18 at 18:08
  • 1
    Link to Mean Well solution: [*Avoiding LED Driver Inrush Current Problems*](https://www.meanwellaustralia.com.au/news/avoiding-led-driver-inrush-current-problems) – Bradicul Sep 22 '18 at 18:08
  • Ha ha ... this may lead to another question; how many such lamps if switched on simultaneously will trip a 33MW power station. – soosai steven Sep 23 '18 at 00:28
  • 1800Wpeak *worse case* surge for a 9watt bulb seems like they could not filter it out without a justifiable cost benefit. So I would expect a tick of EMI.in the AM band, but that flies under the FCC radar. – Tony Stewart EE75 Sep 23 '18 at 03:27
  • @soosaisteven, "how many such lamps", this is actually a very actual question of big practical importance. The answer is "not many", precisely because of this peak inrush current problem. I did have a recessed lighting fixture in kitchen ceiling with 12 CFL bulbs (similar type of power supply), and the switches arced every time and burned away over time. I had to replace them with rocker type, and I am using now LED bulbs. I am not sure how long the switches will last. – Ale..chenski Nov 22 '18 at 19:49

2 Answers2

3

It's certainly appears not to be a mistake in their datasheet to have inrush currents at 7A.
The website shows the same number for the light, and for many more lights the inrush current is even higher, for example this one at 18A.

The LED drivers have even greater inrush current specified. Here is a 10W driver with inrush at 35A ...though here they specify it is a maximum figure. The inrush current has almost nothing to do with the total power of the light, but is based on the size of the input capacitors used in an APFC design.

The power supply/driver is typically an APFC controller, and in this case I'd posit they don't have any NTC inrush current limiters. APFC supplies (used in many power supplies such as for laptops through to PC power supplies) do typically have high inrush currents.

If you are concerned that the inrush current is going to impact your installation you could consider the following:

  1. Using a zero crossing SSR to reduce the inrush component.

  2. Installing an NTC series resistor to limit the peak current.

PS. This problem of high inrush current is not unique, here is an article about PC power supplies where the topic is raised, and an excellent paper here on APFC supplies that shows a partial schematic.

Jack Creasey
  • 21,428
  • 2
  • 15
  • 29
3

I don't see much contradiction. Inrush current is usually an exponentially decaying spike. The "7A" nameplate value is certainly the peak current. The "300ms" is likely a nameplate for spike duration, probably defined at 10% level, so the total energy is much lower than the bold estimation. Here are "definitions" from Murata:

enter image description here

where the duration looks like is defined at 0% level :-)

There is a line-up of technology and tutorials that deals with inrush current measurement, like Keysight

enter image description here

Let me guess some numbers from the Philips datasheet.

If the input has 7A at 240 V peak, the ESR looks like about 35 Ω. If the 300 ms is defined as RC constant, then the capacitor might be (35 * C = 0.3) C =8,500 uF, which sounds too high. So the 300 ms is likely defined differently in the area of industrial lighting.

Misunderstood
  • 7,287
  • 1
  • 11
  • 24
Ale..chenski
  • 38,845
  • 3
  • 38
  • 103
  • there is always an average in electrical as there are many outlying factors in electrical. If that wasn't the case, we wouldn't need apparent, power and reactive values. Not trying to stir the pot and not trying to say I even come close to knowing the full reason for higher inrush in later model control gears but this is what we are taught or learn through the years. – Bradicul Sep 22 '18 at 18:38
  • 1
    Formalities of comments vs. answers is not a big issue. Bigger issue is that the term "average" is normally applied to frequently repetitive events, like AC rectified waveforms and such. Inrush current is not such event. Although it can be averaged over the duration of pulse/spike, but then its meaning (and practical importance) will be lost. Actually this is your use of the term ("average") that sparkled the 1.75kWA contradiction. – Ale..chenski Sep 22 '18 at 19:40
  • I completely agree. The term average should only be used in those presents but I doubt Phillips would advertise a higher reading then they need to, especially when many other popular brands try to hide it. Osram advertise the high inrush current issue as a "hidden/only to search" issue if you google it. – Bradicul Sep 22 '18 at 20:11
  • Keep in mind the inrush graph is for a power supply to a motor. AC inrush to non-dimmable AC power LED depends on the phase of the start voltage. So the worst case inrush unless limited otherwise well depend on phase, cap ESR etc. – Tony Stewart EE75 Sep 22 '18 at 22:56