Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations cowski on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

LED vs Fluorescent Lamps

Status
Not open for further replies.

treez

Computer
Jan 10, 2008
87
Hello,

I have been working on Domestic Mains White LED Lighting (for sale to the cheap-end, mass market) for a while.

I have been working at the 15W level as this is equivalent luminous output to a 60W incandescent.

However, there are multiple problems with LED lights here, and in fact fluorescents are far better when all is accounted for.

Fluorescent bulbs driven by a high-frequency electronic ballast have a significantly greater efficiency (110 Lumens per Watt) than LED Lamps (55 Lumens per Watt for affordable LEDs).


--------------------------------------------------------------------------------
The only significant disadvantages of Fluorescent Lamps, compared to LED Lamps is that…….

1) None of the fluorescent lamp components (ferrites, connectors, enclosure, ) can be re-cycled since the fluorescent lamp contains mercury.

2) Also, as stated, Fluorescent Lamps contain mercury and ploughing several hundred million used-up fluorescent bulbs into our landfill sites, year after year, may or may not cause a pollution problem.

-This shouldn’t be under-estimated as a potential fluorescent lamp show-stopper, since newly introduced RoHS regulations have banned mercury from any electronic devices, and here we are introducing billions of CFL’s, -each one containing ~4mg of Mercury, as a replacement for incandescent bulbs.

3) In theory, fluorescent bulbs have other problems, such as dim cold-start, poor repeated on/off ability etc, but these can be quite cheaply mitigated with cheap additional circuitry.

4) The extrusion of a fluorescent lamp’s glass bulb is a high energy process, and so production energy usage is higher for Fluorescents Lamps than LED Lamps.
–LED Lamps still use up significant energy in manufacture, since the metal LED heatsink needs to be made. –however, this component is re-cycle-able.

4) IN THEORY, fluorescent lamps don’t last as long as LED lamps.
-However, the fact is that in truth the life-time of cheap LED Lamps and Cheap Fluorescent Lamps is actually determined by the life-time of the electrolytic capacitors that such cheap lamps end up needing to contain.
In other words. Cheap Fluorescent Lamps and Cheap LED lamps have the SAME life-expectancy. –The life-expectancy of their electrolytics.
--------------------------------------------------------------------------------


It is point (4) that is the “deciding” factor that says that LED lamps just cannot compete with fluorescents…………
LED lamps have a greater need of electrolytics than fluorescent lamps.
Cheap LED lamps produce generally more ripple current in their electrolytics than cheap fluorescent lamps.

Take apart a 15W fluorescent CFL…..

You’ll see that it only needs one small electrolytic capacitor (A 3.3uF 400V Electrolytic with high ripple and temperature rating )

-This cap occurs after the mains diode bridge.
-It does NOT give good Voltage smoothing of the post diode bridge waveform, since this would just mean poorer power factor…..in fact, in UK, the DC Bus ripples from some 200 to 330V in a typical 15W CFL.

Now try and make a CHEAP mains LED Lamp with this small amount of electrolytic capacitance !

I guarantee that you will not manage.

It can be done, -but the solution becomes too expensive for the mass domestic lighting market. Even if you take away the (great) cost of the LEDs……..the electronics (SMPS) is more expensive for a cheap LED Lamp than a cheap fluorescent lamp.


--------------------------------------------------------------------------------
There are certain design constraints which must be imposed on a cheap 15W mains LED lamp…………..

a) Only 1 offline SMPS to be used (not two stages in cascade as its too expensive)
b) DC Bus voltage (post mains diode bridge) cannot be flat, it must have 100Hz voltage ripple on it of magnitude at least 100V (like in the CFL)

-This is because a flat DC Bus would mean too poor power factor, and even though 15W is below the statutory PFC level, there will be several 100 million of these lamps in each country, so an improved power factor is needed in reality.

c) Secondary must be isolated and must have a max voltage of no more than 42V. –Any more than 42V and it would be dangerous.

-Isolation on the secondary is essential, since the LEDs will have to be heatsink mounted, and parts of this heatsink will have to be exposed to free air on the outside of this cheap LED lamp, where human fingers can touch it.

…like the LED lamp heatsinks you can see here:-


d) LED current cannot be very rippley.
-If it is, then the average LED current will be down, and a further LED or two will have to be added to make up the power.
Since power LEDs are very expensive, adding another LED in to the lamp is just not an option.


The need to conform to the above constraints, as well as the fact that it must be cheap…………….points the designer to pretty much just ONE viable SMPS topology for a Cheap Mains LED Lamp…………….

That Topology is a Current-Mode, 1 Transistor Forward Converter with opto isolation.

Schematic of practical SMPS for cheap mains LED lamp:-

1ep1mr.jpg


-The current regulation, has to be done by a current sense resistor in the secondary rectifier loop, -put it any other place and you loose bandwidth and your LED current gets too rippley.

The current sense resistor voltage has to be put into a compensated op-amp error amplifier, whose other input is a reference voltage.
The output of this error amplifier feeds the opto-coupler diode. The opto-coupler transistor acts on the PWM comparator to control the main primary FET.

The use of a forward converter means you need less secondary electrolytic capacitance because of the secondary inductor.

-Even so, a fairly significant amount of electrolytic capacitance is still needed because of the need to de-rate it so that it lasts a long time, whilst it hangs up there on the ceiling, -the hottest place in the house –also alongside those hot power LEDs.

Try to use just a 3.3uF, 400V Electrolytic following the mains diode bridge (like in the 15W CFL) and you run into instant problems.
–your primary ripple current is way too much in such a forward converter……….

-You can’t solve this problem by derating the input capacitance, since then you end up with a too_poor conduction angle on the mains diode bridge diodes, and a poor power factor.

-Your only option is then to use a passive “Valley Fill” Power Factor Corrector, which ups your component count but does allow you to sufficiently derate the input capacitance.

-Unfortunately, the “snap-off” (every 10ms) of the centre diode in the “Valley Filler” suddenly breaks the current in the inductance (ESL) of the “Valley Fill” electrolytics and causes a voltage spike and a worsening of Common-Mode-Noise……which you have to mitigate by keeping the capacitor charge current down with a resistor in series with the “Valley Fill ” electrolytics….this makes your efficiency worse.

-----------------------------------------
-Sounds bad, but Flybacks (CCM or DCM) are even more hopeless in this situation.

Flybacks need much more secondary electrolytic capacitance than forwards, which has to be well derated to keep the life-time of the LED lamp relatively high….which means using even more electrolytic capacitance………..and thus the size and expense of the SMPS becomes impractical.

Another unfortunate point about Power LEDs, is that even though they can be made with an illumination angle of 180 degrees…….-that is, sufficient angle of illumination to light-up an entire room………………the luminous intensity cannot be made to be equal over every ste-radian……………..and the LED’s illumination intensity falls to
about half the axial intensity at an angle of some 60 degrees from the axis.

-eg see page 18 of this white power LED datasheet…


This means that if your power LEDs point directly downwards from the ceiling, your room will not be evenly illuminated, and if you have just enough brightness in the axis, then it will be too dim on the periphery.

-You can get round this by having complicated, angled , mounting fixtures, but these will add significantly to the expense of the LED lamp.

-You can also get round this by having a diffuser, but then you need higher power LEDs to get the same luminous intensity, and your costs increase too much.
-----------------------------------------------

I believe the above shows that for the mass, cheap, domestic, white lighting market, Fluorescent Lamps are clear winners over LED lamps.

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

However, Professor Colin Humphreys, of Cambridge University, UK, claims to be able to provide us all with cheap White, GaN, Power LEDs of some 250+ Lumens Per Watt in about 2014.

-This would mean LED Lamps win over Fluorescents.

This kind of efficiency improvement would be too significant to ignore.

Until that hopeful day, Fluorescents come first…….At least for the Cheap, Mass Market, Domestic Mains White Lighting World.

I wonder if others agree?
 
Replies continue below

Recommended for you

If you wanting to be an advocate or promote one technology over another, this is NOT the place to do it. You obviously have a lot of interest in this - maybe put up your own website or blog.

If you have specific questions of a non-philosophical nature, I'm sure someone will be happy to answer.

"The more the universe seems comprehensible, the more it also seems pointless." -- Steven Weinberg
 
trees; Never-the-less LEDs will be the lighting of the future. You are hinging too much on 'cheap'. In my town, (arguably filled with crazies), there is a store that stays in business selling LED lights.

I looked at a LED driver last night that specifically ran with no electrolytics. It may not have been off-line I don't recall.

LEDs will get substantially less expensive as time goes by.

Besides the mercury issue is just too horrible to ignore.

Keith Cress
kcress -
 
I might point out the now standard argument used for the Hg question with CFLs. The fact is, the amount of Hg released into the atmosphere by fossil fuel power plants to power an equivalent amount of incandescent lamps far exceeds the amount of Hg that ends up in the environment from discarded CFLs, by something like a 100:1 ratio (don't quote me on that ratio though, I'm just too lazy to find that article again). And the CFLs can be legislated into being recycled as well, which, although not perfect of course, will result in even less of the Hg getting into the envionment.

Still, the only viable replacement for CFLs in the future will be LEDs IMHO, as long as they they can be made to work out reliably and inexpensively. I bought one to try it out and I am still a bit disappointed in the lumen output and light quality; it's a bit too blueish for me. The damned thing died in a week as well so I had to send it back for a refund. I think LED lighting is improving at a logarithmic rate, but it's not there yet.


"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376
 
My company (not my site though so I don't know much) is involved in making equipment to manufacture LEDs. We had a presentation a year or two back showing the time line for when they thing LED will be able to generally replace CFL.

I can't remember the dates now but I know it was a little way out. Also I don't know if the economic turmoil of the last year or so is affecting this.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies: What is Engineering anyway: faq1088-1484
 
Interesting point Jraef. BTW that was a Popular Mechanics article.

KENAT; The 'turmoil' does affect this as it gets people's focus off of 'green' and onto 'money'.

Keith Cress
kcress -
 
So it's either the National Green Front or the Universal Greenback?

What a choice!
 
Do LED's solve some of the application problems of current CFL's?

For instance, CFL's aren't always well suited to 'enclosed' fixtures originally designed for incandescent bulbs. They tend to overheat if I recall correctly and certainly I’ve had a couple go in a short time frame in such applications and I’ve switched back to incandescent bulbs in at least one fixture. I’m not sure if this is the CFL as such or the circuitry the OP mentions. Then again, the LED solutions I’ve seen so far aren’t drop in replacements for bulbs, they seem to require new fixtures for the most part, though I haven’t looked in a while.

Another issue is their use in moist/humid areas. I tried replacing the bulbs in our bathroom but they seem to have a fairly short life (one blew with a nice puff of smoke while I was in there), I’m assuming because of humidity as they aren’t enclosed fixtures. Again, not sure if this is the CFL as such or the circuitry.

Itsmoked, I pretty much figured the answer would be a yes, although with all the green money in the ‘stimulus plan’ wasn’t sure if it would be for better or worse;-).

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies: What is Engineering anyway: faq1088-1484
 
As I understand it, Hg is a by product of some mineing processes, sort of like floride. The question is if we don't use it for something, then how do we dispose of the large amounts of it?

It's from the earth. If the worst is it is eaten by fish, then the fish die, and become part of the muck at the bottom of the sea. Then it will be returned in a few million years in a volcano.
You can't get rid of it, only deal with it.

LED's are great, 5W of LED keeps my plants growing very well. At least until it gets warm enough to put them outside.
 
cranky108 said:
If the worst is it is eaten by fish, then the fish die, and become part of the muck at the bottom of the sea. Then it will be returned in a few million years in a volcano. You can't get rid of it, only deal with it.
Unless we catch and eat those fish... then it's a bigger deal.



treez... I stopped reading after the first page as your post seemed to be more of a philosophical quest than an actual engineering question. I'm with Keith... your bent seems to be towards cheap cheap cheap. Give me two products, a more expensive one made with QC and a cheaper one made without, and the cheaper piece is going to fail more quickly. You can make a proper judgment on LEDs (or CFLs, for that matter) by picking up the cheapest you can find at the local drug store and seeing how long two bulbs last.

Dan - Owner
Footwell%20Animation%20Tiny.gif
 
I can confirm the 'cheap' cfls (subsidized by the local electrical company) don't appear to last well, not even as long as an incandescant. Though, as I mention a couple of my applications may not have been well suited to CFLs.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies: What is Engineering anyway: faq1088-1484
 
Certainly, the earlier CFLs, even with the slightly higher quality of Costco, were less than long-lived.

TTFN

FAQ731-376
 
Which is strange because one of the common claims for CFLs is they last longer.

It's certainly true for some as my parents got one in the early 90's (when I still lived there) and it was still going 10+ years later.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies: What is Engineering anyway: faq1088-1484
 
The first twenty I bought.. Not a single one lasted more than a month. Now the $1 ones I've bought, all have lasted more than a year. Go figure.

Keith Cress
kcress -
 
My point is Hg is here to stay, not that feeding it to fish is a good idea. It will come back in time.

But hey we aren't using that place for spent nucular fuel, so I guess we can store Hg there.
 
Dunno, Somalia is looking pretty good to me as a substitute for Yucca Mountain.

TTFN

FAQ731-376
 
Interesting tidbit:

After extensive testing the Pentagon will be switching all 4,200 lighting fixtures in Wedge 5 to LEDs.

LEDs resulted in a 22% reduction in energy over fluorescents and substantially improved the light quality.

Payback will be less than 4 years.

HVAC power requirements dropped. Maintenance was reduced. It will drop CO2 emissions 140 tons per year for that wedge.

"The hazardous waste for disposal of mercury laden fluorescent bulbs was eliminated".

Keith Cress
kcress -
 
Aren't most of the active elements in LEDs fairly harmful? I know they're encapsulated but in this green world that encapsulation will be recycled and the nasties inside will be on the loose.


----------------------------------
image.php

If we learn from our mistakes I'm getting a great education!
 
I'll stick to something more conventional like sugar... [smile]


----------------------------------
image.php

If we learn from our mistakes I'm getting a great education!
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor