Disappointing trend in LED manufacturing

[PhotonWrangler]

The other day I was looking at some LED nightlight bulbs from Feit. They had red, green and blue ones. The clear bulb reveals a wafer with the LEDs mounted on them. The blue LED had just the bare wafer-mounted LEDs, while the red and green ones had the wafer and a blob of phosphor on top of them. The phosphor blobs were glowing dimly under the store's lighting.

So all three colors are using the same blue LED chips, and the phosphor converts blue to either red or green. What bugs me about this is that it would be more efficient (and probably a little brighter) to use native red or green LEDs instead of the phosphor-downconverted blue ones.

I get it that this method reduces the number of components they need to stock at the manufacturer, but it adds one step to the manufacturing process - adding that dollop of phosphor to the non-blue bulbs - but it also introduces another point of failure in addition to the efficiency loss.

I've seen the same thing in a line of larger colored LED filament bulbs from the same manufacturer. Does this bug anyone else or am I just being pedantic?

 

[ampdude]

Why would you do things correctly when you can do them cheaper in some way? Save a penny. It's a psycho world these days.

Who cares about the end product if only you or I and a few others notice the difference. The only saving grace is the interweb.. but nobody uses that anymore, everyone is on their phones watching TikToks. :/

 

[idleprocess]

So all three colors are using the same blue LED chips, and the phosphor converts blue to either red or green. What bugs me about this is that it would be more efficient (and probably a little brighter) to use native red or green LEDs instead of the phosphor-downconverted blue ones.

I recall that Philips invented blue-pumped green precisely because blue die were so much more efficient than green that it was still a significant efficiency boost despite the phosphor losses. Not sure about red, however.

I get it that this method reduces the number of components they need to stock at the manufacturer, but it adds one step to the manufacturing process - adding that dollop of phosphor to the non-blue bulbs - but it also introduces another point of failure in addition to the efficiency loss.

The common LED Vf / driver design for all three channels probably makes for less overall complexity.

 

[LEDphile]

The biggest downside to the phosphor-converted color LEDs is that the color is a bit less saturated than you'd get with a direct-emission color LED. The big reason for using them in a design (assuming you aren't choosing them for their spectral content) is the thermal performance and Vf range are basically the same as for white and blue LEDs. Given that green LEDs are typically higher Vf and red LEDs are typically much lower Vf than blue, this can be significant.

Looking at Osram LEDs (the Duris S5 is available in both direct and phosphor-converted green and red), the PC green is ~188 lm/W vs. 135 lm/W for the direct green (almost 40% more output). The story is a bit different with PC red, as the direct red is approximately 3x as efficient at 25C. But because of the much better thermal droop characteristics of the blue LED die, at 85C (a more typical operating point), the PC red is still outputting 90% of the amount it was at 25C, while the direct red is down to about 50% of its initial output. If you assume a linear driver topology (starting at a constant voltage and using a transistor to regulate current, with the voltage drop difference between the LEDs and power supply being dissipated in that transistor), the wallplug efficiency of the phosphor red is probably about the same as for the direct red due to the Vf differences.

So for comparable or increased performance, why wouldn't you want to go with the option that has greater commonality across variants (and thus a simpler supply chain), especially when the LED manufacturer is the one taking the extra complexity for building the phosphor-converted LED?

(side note - the LED manufacturer may also be seeing some supply chain and/or manufacturing benefits from the PC LEDs vs. the direct colors, but how much this matters depends on the LED manufacturer)

 

[KITROBASKIN]

Very interesting thread. Wife just ordered a multi color, ambience, corded desk light. I will check it out but likely different than this issue.

The observation that supply chain shortcomings could factor in this design decision; interesting, but thinking simplicity of components is probably why?

 

[PhotonWrangler]

This is fascinating. I didn't know that PC green was more efficient than direct green, nor did I know about the thermal droop issue with direct red vs PC red. This makes much more sense now. And I can see how a manufacturer's engineering and procurement teams would want to keep the parts count at a minimum with the supply chain issues still going on.

Thanks everyone. I always come away from this place a little smarter than I was when I logged on.

 

[PhotonWrangler]

It occurred to me that there's one disadvantage of PC red for night lights. The blue content could defeat the purpose of using red in terms of not upsetting one's circadian rithym. The bulbs that I bought were intended for just such a use, so I will need to conduct a check for fluorescence from them, or maybe use a prism to check for any blue energy escaping thru the phosphor.

 

[LEDphile]

It occurred to me that there's one disadvantage of PC red for night lights. The blue content could defeat the purpose of using red in terms of not upsetting one's circadian rithym. The bulbs that I bought were intended for just such a use, so I will need to conduct a check for fluorescence from them, or maybe use a prism to check for any blue energy escaping thru the phosphor.

The better PC red LEDs have basically no leakage of the blue pump through the phosphor (of the spectra I've looked at, some have no leakage at all, and the worst I've seen have the pump leaking through at around 2% of the spectral peak). Based on the published nominal spectral sensitivity of the circadian system ( https://www.lrc.rpi.edu/programs/lightHealth/pdf/spectralSensitivity.pdf ), that small blue pump has slightly less effect than the broad red peak. But the other important consideration is that the light level matters, and the amount of output you get from a nightlight bulb is unlikely to have any effect unless you've got it inches from your eyeballs.

 

[PhotonWrangler]

Thanks LEDphile. A pump wavelength leakage of 2% worst case doesn't sound bad. This light will be a good distance away, and depending on brightness I might add some semi-opaque tape over the lens to further reduce it.

I currently have an alarm clock with an adjustable RGB backlight on the LCD display. I've found that the red LEDs don't bother me at all and I can fall asleep while facing them only a couple of feet away. The other colors are a different story.

 

[Lynx_Arc]

I'm of the mind that manufacturers are less interested in efficiency of these very low output bulbs and more interested in profit. I've seen stats that have 40watt EQ LED bulbs more efficient than 7watt EQ ones. I am still using night light LED bulbs (7 watt incan replacement) I bought long ago for really cheap about $1/bulb that have 3x5mm LEDs in them. Newer bulbs have smd type chip LEDs either a single flat mounted one or several on a sort of tower sticking out the middle of the bulb. I run 3 of these lights in other parts of the house so I can see late at night and they cost less than $1 a year to operate. I know when I get home if the power is out at night.

And no, I don't have my main alarm clock close to me, it is by my TV across the room so when it goes off I have to get out of bed to turn it off. I do have a radio by my bed and a small LCD battery clock in case power goes off while I sleep.

 

[JustAnOldFashionedLEDGuy]

The output consistency of the PC products over temperature and even manufacturing makes for much more consistent products especially if you are doing color mixing. As well, the electronics are easier.

 

[PhotonWrangler]

I've just fired up one of those phosphor red nightlight bulbs and I could not see any visible blue leaking out. I also held another one of the phosphor red bulbs next to it and the light from the lit bulb did not produce any fluorescence in the adjacent bulb whatsoever, so I'm satisfied that these will be ok for night light use.

 

[ampdude]

Till it degrades or the phosphor gets jolted off. Seems like another bean counter solution to a problem that didn't really exist, except for making more money for someone.

 

[kaichu dento]

Why would you do things correctly when you can do them cheaper in some way? Save a penny. It's a psycho world these days.

The thread just started when you posted this one, so I could understand the cynicism, but then...

Seems like another beaner counter solution to a problem that didn't really exist, except for making more money for someone.

Lots of posts trying to explain, but your cynicism in the 2nd post wasn't a precursor to finding more out; it appears to be your permanent mindset. Too bad that negativity is all you can add, because there was a lot of interesting stuff here.

 

[ampdude]

No, I did some more research and looked it up before I made my second post. The factory counters want to squeeze every tenth of a penny out of every metric ton. Have you ever worked in a factory? You'll see different parts and sub-parts EVERYDAY from anything they can source the cheapest. One day you'll have nickle sockets, the next day you'll have copper socket, the next day you'll have aluminum sockets, the next day you'll have stainless sockets. Some people are just all about efficiency and cost over reliability and quality, and that's where I differ in my opinions. I think that LED's are designed well for reliability, but when you start putting gunk on them so that you can mimic an effect by cheapening things out, I'm not going to agree with it.

 

[jtr1962]

This data sheet for Cree's new line of 2835 color LEDs is a good source to compare relative efficiencies of direct versus PC LEDs. PC green and amber are far more efficient than direct green/amber. PC red-orange is about equal. They aren't making PC red, but I imagine it would less efficient. LED wall-plug efficiency is highest on both sides of the visible spectrum (blue and red). It drops in between.

Since someone mentioned thermal droop, while PC red would have better thermal droop characteristics than direct red, I noted from the data sheet that photo red (660 nm) still has 88% output at 85C (compared to ~62% for regular red). So you don't necessarily have to use PC red if thermal droop is your only reason for considering PC red. Photo red also has higher WPE. It's not going to get as warm for any given drive current.

 

[jtr1962]

The factory counters want to squeeze every tenth of a penny out of every metric ton. Have you ever worked in a factory? You'll see different parts and sub-parts EVERYDAY from anything they can source the cheapest. One day you'll have nickle sockets, the next day you'll have copper socket, the next day you'll have aluminum sockets, the next day you'll have stainless sockets. Some people are just all about efficiency and cost over reliability and quality, and that's where I differ in my opinions. I think that LED's are designed well for reliability, but when you start putting gunk on them so that you can mimic an effect by cheapening things out, I'm not going to agree with it.

And this drives me crazy because I often end up fixing things where they cut corners to save tenths of a cent. Ask anyone who buys something if they would be willing to pay a few cents more for something much more reliable. The answer I'm sure would be yes. If the factory has to pay a nickel more for decent parts, so what? Charge the end customer $10.04 instead of $9.99 for the product. You make the same amount of profit on it.

The worst part about all this is the fact you don't even know what's good any more. Buying name brands, or paying a lot more, doesn't always guarantee better quality. More often than not the name brand of something is the exact same as the generic version, but it has a fancy sticker slapped on it. It used to be that more costly things were actually better made.

Too bad we don't have the same philosophy as manufacturers did a century ago. Design it to last a lifetime. And design it to be easily fixed if it does break. Now we intentionally design things to fail after x years or number of uses.

 

[fulee9999]

The worst part about all this is the fact you don't even know what's good any more. Buying name brands, or paying a lot more, doesn't always guarantee better quality. More often than not the name brand of something is the exact same as the generic version, but it has a fancy sticker slapped on it. It used to be that more costly things were actually better made.

for me this is the most infuriating part about today's manufacturing... I have to do extensive research sometimes to figure out if the $16 part is the same or worse than the $99 one. And sadly a lot of the times it's the same on the inside, just have a fancier package.

 

[PhotonWrangler]

The trend in general is a "race to the bottom" as others have referred to it. The early name brand household LEDs were pretty over engineered with lots of potting compound in the base for heatsinking as well as decent component choices like higher-temp electrolytic capacitors. This was when they were touting 50,000 to 100,000 hours of lifetime for the bulb. Look at the lifetime claims on the current crop of inexpensive bulbs and they're much shorter. The goal is to get the cost of the bulbs down to the point that they were cost competitive with incandescents. It makes sense from the standpoint of getting your product to move off the shelves, but ultimately the customer pays by having to replace them more often.

That nightlight bulb appears to be an exception in terms of output vs cost and I have to admit that I was pleasantly surprised when I learned of the efficiency gains of PC LEDs vs native-wavelength LEDs. I always come away from this forum better educated than when I came, and I am humbled by the breadh and depth of knowledge and experience (and even healthy skepticism) shown by the membership here.

 

[Lynx_Arc]

The trend in general is a "race to the bottom" as others have referred to it. The early name brand household LEDs were pretty over engineered with lots of potting compound in the base for heatsinking as well as decent component choices like higher-temp electrolytic capacitors. This was when they were touting 50,000 to 100,000 hours of lifetime for the bulb. Look at the lifetime claims on the current crop of inexpensive bulbs and they're much shorter. The goal is to get the cost of the bulbs down to the point that they were cost competitive with incandescents. It makes sense from the standpoint of getting your product to move off the shelves, but ultimately the customer pays by having to replace them more often.

That nightlight bulb appears to be an exception in terms of output vs cost and I have to admit that I was pleasantly surprised when I learned of the efficiency gains of PC LEDs vs native-wavelength LEDs. I always come away from this forum better educated than when I came, and I am humbled by the breadh and depth of knowledge and experience (and even healthy skepticism) shown by the membership here.

Now that they have pretty much turned their backs on incan and CFL bulbs they have to keep the cost down and the longevity claims were preposterous as the electronics in bulbs often give out long before the LEDs. I still have 20 and 60 watt CFLs I paid about 25 cents each for that I will use. I don't run these for hours at a time mostly the total amount of all is about 1-2 hours a day which would take a long time to pay for the difference in cost/efficiency of an LED bulb.

As for efficiency of colored vs doped blue/white LEDs I've seen that for a long time. For me red LEDs vs firefly/candle mode white LEDs is a no brainer. I don't need to use red and the efficiency of RED LEDs as the same current/output is worse which makes them less attractive than ever. The two things I like about LED vs CFL is start up time and temperature in use. LEDs are better for garages