PF for LEDs

texie

Senior Member
Location
Fort Collins, Colorado
Occupation
Electrician, Contractor, Inspector
You guys are thorough. :)

The OEMs may not lie but they may mislead, intentionally or otherwise.
I don't think they are lying or misleading so much as just lousy docs. Poor documentation/instructions in my view is a big problem today with many products. Especially in the area of indicating details required for full NRTL and code compliance.
 

Flicker Index

Member
Location
Pac NW
Occupation
Lights
I looked at that link for the small units and, yes, it says the OUTPUT W and the VA are the same. It seems to me that that PF of .44 is the INPUT PF. That seems explain the discrepancy.
I don't think so. If it was the INPUT PF, then I don't see how it would say "leading or lagging" 0.44. This suggests it is capable handling either or. If the power factor is a characteristic of the UPS front-end, then it would be one or the other, not lead or lag.
 

texie

Senior Member
Location
Fort Collins, Colorado
Occupation
Electrician, Contractor, Inspector
I don't think so. If it was the INPUT PF, then I don't see how it would say "leading or lagging" 0.44. This suggests it is capable handling either or. If the power factor is a characteristic of the UPS front-end, then it would be one or the other, not lead or lag.
Good point.
 

WasGSOHM

Senior Member
Location
Montgomery County MD
Occupation
EE
Poor documentation/instructions in my view is a big problem today with many products. Especially in the area of indicating details required for full NRTL and code compliance.
They're all doing it.
Even if they know the details they leave them out.
People seem to be buy it anyway.

I like to open the package right after I buy it, and I know the seller has to use all available space for display, but I don't see a lot of stores having any space reserved for customers to unpack/inspect stuff.
 

Flicker Index

Member
Location
Pac NW
Occupation
Lights
When it comes to Home Depot, I sometimes open it before I buy it. It's a common practice around here that they put returned items right back on the shelf.
 

paulengr

Senior Member
LEDs are current based not voltage and high power LEDs must be current limited or they burn up with a voltage drop (band gap) of about 6 VDC. So LED power is very different from anything else. Most drivers are switching buck power supplies so on the line side they are capacitive. So leading around 0.7.
 

gadfly56

Senior Member
Location
New Jersey
Occupation
Professional Engineer, Fire & Life Safety
LEDs are current based not voltage and high power LEDs must be current limited or they burn up with a voltage drop (band gap) of about 6 VDC. So LED power is very different from anything else. Most drivers are switching buck power supplies so on the line side they are capacitive. So leading around 0.7.
If true, how do you account for the results at post #12?
 

kwired

Electron manager
Location
NE Nebraska
LEDs are current based not voltage and high power LEDs must be current limited or they burn up with a voltage drop (band gap) of about 6 VDC. So LED power is very different from anything else. Most drivers are switching buck power supplies so on the line side they are capacitive. So leading around 0.7.
I don't have complete grasp of effects of this sort of power factor, but do realize it is a distortion PF and is different from displacement PF like you get from motors. If it typically has a leading effect instead of a lagging effect then that is likely why you don't hear much about trying to correct it as it would have some cancelling effect to the motors that are still causing bigger issues for the POCO's simply because there is so much more motor load on the system.
 

paulengr

Senior Member
I don't have complete grasp of effects of this sort of power factor, but do realize it is a distortion PF and is different from displacement PF like you get from motors. If it typically has a leading effect instead of a lagging effect then that is likely why you don't hear much about trying to correct it as it would have some cancelling effect to the motors that are still causing bigger issues for the POCO's simply because there is so much more motor load on the system.
Switching power supplies are becoming prevalent. You hear complaints about data centers.

In a DC switching power supply the LED will be connected to a capacitor. There will be a sensor too. Hall effect, shunt, or voltage divider. That’s the back end. The voltage/current regulator pulses an SCR or MOSFET on and off to regulate output current. So the AC line side sees a capacitor. Current draw looks like “rabbit ears”. Depending on phase angle it can appear resistive, capacitive, or resistive but highly nonlinear.

Utilities are concerned with transformer load. With true sine waves they are concerned with total amps (kva) because all a transformer sees is amps. There are switching arguments too but amps is the key. So we bill on kw but load is kva.

With harmonics this still applies but in other ways. Not all harmonics are equally bad. VTHD is universally bad. This is high when harmonics are saturating the power delivery system causing “brown outs”. Current is rated on ITDD, current demand distortion. This is measuring current harmonics as a percentage of transformer capacity. If we aren’t near capacity even if current harmonics are horrendous voltage is unaffected and we don’t care. It’s only at capacity where current harmonics temporarily push the transformer to the point that it causes voltage swings or voltage harmonics that it becomes an issue.

Solutions are dilution (as mentioned motor loads offset harmonics), bigger transformers and wiring, reactance, and tuned filters. Line length helps a lot since that is reactance.

With the growth of data centers utilities are used to it. As more motors switch to drives with PMDC even motors are contributing. A bigger local concern is undersized neutrals in 3 phase. Using full size neutrals is an easy fix.

Response to questions about leading/lagging: switching power supplies are like solid state synchronous loads. They can be either but low cost ones are simple nonlinear “rabbit ears” with 20% current %THD (nor ITDD). Thinking about it a little with a pure SCR design since those can only turn on at most firing angles it will be lagging or at least appear that way in the fundamental. It’s a leading load but since the SCR turns on after the voltage zero crossing based on current regulation most of the fundamental harmonic appears lagging and the rest will just reflect what the Fourier math does to such a nonlinear wave. Fourier math just isn’t really suited to PWM. Power factor on switch mode supplies is a square peg/round hole problem.

You can do leading PF easily with FET / IGBT supplies and enough modulation index.
 
Last edited:
Top