Some Chinese designed LEDs are now using resistive series ballast

Status
Not open for further replies.

winnie

Senior Member
Location
Springfield, MA, USA
Occupation
Electric motor research
170912-2449 EDT
Then in post 17 you reference an ON datasheet that does not map to the device in your circuit of post 11. The ON chip and circuit appear to be a form of pulse width modulation, far lower power dissipation than a linear circuit.

I only skimmed the datasheet, but it is pretty clear that the PWM seen in the ON seiconductor datasheet is at 120 Hz.

A true switching LED regulator would operate at much higher frequency to permit the use of smaller inductors and capacitors.

In this case they are making lemonade out of lemons: by allowing the LED to flicker at line frequency and shut off when the supply voltage drops during the AC cycle, you get a dimmable lamp. A flickering dimmable lamp.

-Jon
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
170913-1707 EDT

The description in post #1 implies a current regulating chip operating in its current regulating mode and being supplied by a DC source. Post #17 changes this after one does a time constant analysis based on some assumptions.

Pulsing DC from a rectified sine wave is the input to the integrated circuit. The integrated circuit is no more a linear regulator, but only during certain time periods. The combined circuit does not provide a constant output current.

The output is pulses of current clipped to a peak current defined by the regulating chip. The average DC output current will vary as the input sine wave RMS voltage varies. The output pulse width will vary with input voltage changes.

.
 

Electric-Light

Senior Member
170913-1707 EDT
The description in post #1 implies a current regulating chip operating in its current regulating mode and being supplied by a DC source. Post #17 changes this after one does a time constant analysis based on some assumptions.

Pulsing DC from a rectified sine wave is the input to the integrated circuit. The integrated circuit is no more a linear regulator, but only during certain time periods. The combined circuit does not provide a constant output current.

The output is pulses of current clipped to a peak current defined by the regulating chip. The average DC output current will vary as the input sine wave RMS voltage varies. The output pulse width will vary with input voltage changes.

.

gar, my post wasn't to satisfy electronics engineering readers and its not my purpose to discuss unnecessary fine points of LED ballast topology. The On-Semi example circuit doesn't use a storage capacitor. The table in datasheet does a fine job of reducing overall regulator efficiency due from the effect of line voltage. The no-name, non Energy Star lamp and the sample of Energy Star Greenlite lamp that I evaluated utilize an electrolytic capacitor. An LED ballast can not fill the zero crossing gap without energy storage.

for my discussion, reasonable range in this post means stated nominal voltage +5%, -5% and series linear ballast means an in-line device used to provide ballasting action by burning off energy in resistive resistance. Ancillary components that pass current to provide stability to triac dimmer is not part of the primary ballasting scheme.

CEC T20, T24 or Energy Star current do not discuss how line voltage variations are handled. The handling of line voltage regulations within acceptable range should be evaluated as On Semi datasheet shows rise from -5% to +5% clips off LED ballast efficiency by 9% before you consider thermal loading that further drops lamp system efficacy. The upcoming CEC codes place some restrictions on permissible flicker level. The increased ballast heat dissipation raises temperature and accelerates capacitor decay which can lead to premature lamp flicker growth beyond prescribed limits.

I think it would be a good idea to declare lamps EOL when flicker escalates beyond the prescribed limit and set limits on efficacy and life fluctuation when operated within ANSI voltage specs. The design change on the Greenlite lamps were mostly at the ballast. These added requirements would help reduce design practices of tuning for operation on 120.00v laboratory power supply.
 
Last edited:

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
170915-1130 EDT

Electric-Light:

What I am trying to do is analyze from basics.

From Home Depot I bought a Armacost LED Tape Light. The rating is 2700 K, 65000 hours, 250 lumens per foot, 8.2 ft, 2000 lumens, 12 V DC, and 29 W. Calculated current 1.58 A.

Measured data:

..V.........I.. Using an adjustble voltage DC supply, meaning extreme little ripple.
07.6---0.01
08.1---0.12
08.4---0.20
09.0---0.40
09.5---0.59
10.2---0.80
10.6---1.00
11.2---1.20
11.7---1.41
12.0---1.56
12.2---1.60


..Vpeak...I.. Using an adjustble full wave rectified voltage DC supply with no filtering.

08-----0.03 Scope peak voltage, DC average current.
10-----0.24
12-----0.63


Plots:

Source a center tapped secondary feeding two diodes to produce a full wave rectified output. No filter capacitor. A 2 ohm resistor to measure current on the scope. This can be considered part of the internal impedance of the source. The scope voltage is measured across the LED light strip.

Plot 1, a low current level.

DS2_QuickPrint93M.JPG


Plot 2, a higher current level.

DS2_QuickPrint94M.JPG



Plot 3, current at a voltage peak of 12 V.

DS2_QuickPrint95M.JPG



Plot 4, light intensity compared to applied voltage.

DS2_QuickPrint96M.JPG


The 100% ripple is apparent here at a frequency of 120 Hz. Zero light level is at the middle of the screen.
.

.
 

Electric-Light

Senior Member
170915-1130 EDT
What I am trying to do is analyze from basics.

Tape and rope LEDs for 12v use parallel strings of three LED elements + resistor in series spread over a large physical distance. Each string can be made of three LED packages and a SMD resistor or it could be a three chip package and less obvious resistance trace. LEDs possess an NTC characteristics and parallel driven elements that are not thermally bonded together causes current to accumulate towards the hottest string hence ballasting of individual strong.

The commutation behavior as seen from the line can be different from a long, single series string. For 120v mains circuit, there can be 40-48 in series, and for 240v, 80 to 96. As you get closer to edging Vf to Vpk, things change. There is a general region where linear ballasted LEDs operate in constant BF, variable BEF mode where ballast dissipation is a function of line voltage.

This post does not take side in favor of or against the lamp type used for converting electrical current into light, although i am discussing the effect on light quality and luminous efficacy from ballast implementation that are designed around manufacturing cost interest.

Linear ballast works by burning off energy in the process
vwzlg5.png


2poy7uq.png


A design implementation like this primarily benefits to widen the gap between consumer price and cents of variable cost per final product. www.onsemi.com/pub_link/Collateral/DN05013-D.PDF

There are few references in there pitching superiority to fluorescent. it omits notably inferior light quality due to presence of extreme flicker. how convenient right? This is not LED vs fluorescent. It comes to LEDs being able to run from a crude ballasting circuit that can have abysmal efficiency and terrible light quality.

To manage flicker, it is necessary to store energy in the phosphor, thermally or in the drive circuit. Incandecents and HIDs do so thermally. Fluorescent do so in the phosphor and drive circuit. An entirely solid state circuit together with LEDs is a recipe for extreme flicker.

A decent LED ballast is based on the same design as a good fluorescent ballast and provides the same benefit like good line maintained efficiency over line variation and low flicker. The ballast is a significant part of fluorescent system cost. A crude LED ballast can present an appearance that LEDs are economical without providing the same performance. Omitting capacitor and letting a lot of flicker spew out is an easy way out of having to deal with thermal vulnerability of capacitor in integral ballast.

CEC T20 and T24 are leaping ahead in recognizing flicker issues, but for those outside of California Energy Star basically have no restrictions on flicker. Influence of line voltage variation, flicker index and flicker percentage are those that requires particular attention.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
170915-1727 EDT

I have looked at a number of different light sources (LEDs, CFL, fluorescent, and incandescent) and qualitatively (no absolute calibration and linearity) measured 120 Hz ripple. If there is interest I will present some scope plots.

.
 

Electric-Light

Senior Member
170915-1727 EDT

I have looked at a number of different light sources (LEDs, CFL, fluorescent, and incandescent) and qualitatively (no absolute calibration and linearity) measured 120 Hz ripple. If there is interest I will present some scope plots.

.
Sounds about right. It's a question of how much rather than reducing it below threshold of detection.

Though no government mandate, some ballasts have a internal specification limit on flicker. This is a matter of ballast design and it is something that can be applied to both fluorescent and LED light.

https://unvlt.com/pdf/literature/navigator/pdfs/Sect_1 Electronic.pdf "Ballasts shall operate lamps with no visible flicker (<3% flicker index)"

What you observed is documented here:
http://apps1.eere.energy.gov/buildings/publications/pdfs/ssl/flicker_fact-sheet.pdf

Regarding narrow optical performance LED products, "tuning" is a legitimate concern that reduce the overall purpose. Products with fairly wide input voltage range, but a narrow optimal range inevitably means optimizing somewhere, or widening optimal range. Kept unchecked, OEM supplier may tune for certification benchmark criteria for lucrative market that becomes inefficient in the real world or when the same spec items are also recycled for lower volume products or less restrictive regions.

For example, 277 to 347v rated linear ballast polished for efficacy at 277v can become terribly inefficient at 365v (347 +5%) and designed to export inefficiency into Canada which is something not desirable from energy conservation view.
We might be able to do something statewide by adding to CEC or E.S. criteria disallowing the recycling of US market design into export market to benefit economy of scale of mass production in the US market.
 
Last edited:

b1miller

Member
Location
Washington
Occupation
Electrician/Electrical Administrator/SCADA Engineer
LED bulbs with garage door openers

LED bulbs with garage door openers

While my experience with LED bulbs inside the house has been very good, I encountered a weird situation in my garage & shop. I figured I would save some money and install LED bulbs in all 3 of my garage door openers. Then my wife complained that the remote Acura Link in her car would no longer close any doors. I also noticed that my remote in my PU would not work either. The only thing that they would do is raise the door IF the opener lights were off. Went back to fluorescent lights and everything works. I have a bank of 4' LED fixtures about 3 feet away from the openers with no ill effect. These bulbs must put out some hellacious EMI. Go figure.
5 year old Liftmaster openers BTW
 

GoldDigger

Moderator
Staff member
Location
Placerville, CA, USA
Occupation
Retired PV System Designer
You could try different model LEC bulbs. There are relatively RFI ratings available in this Forum and through an internet search.

Sent from my XT1585 using Tapatalk
 
Status
Not open for further replies.
Top