LED lamps vs fluorescent lamps?

[RODALCO]

The LED traffic light have strings of say 50 lots, of 4 LED's in parrallel. and they are in series.
( don't quote me on the exact number , I need to check that )
A powersupply limits the current through the LED's.
My guess in 20 mA per LED. for 4 LED's makes 80 mA.
240 Volts mains is used in NZ.

 

[Flyback]

Thanks, it seems an unusual choice to have 50 "units" in series, where one "unit" is 4-leds-in-parallel......I don't see the current sharing well between the four in parallel....paralleling leds is usually, if at all, done with long series strings placed in parallel......in other words, it seems very strange that they didn't just use 4 parallel strings of 25-leds-in-series instead.
Was there a manufacturer?
I know of a company that was into the "4 leds in parallel" units.

 

[schmitt trigger]

Each lightning technology, heck even a lowly incandescent bulb, have specific strengths and pitfalls, and the selection all depends on the particular application.

For instance, will you ever see an LED or CFL inside a baking oven? Of course not, they would die immediately. But incandescents thrive on heat.

CFLs likewise, have a cold problem. At freezing temps they start with great difficulty or not at all. LEDs actually work very well at cold temperatures.

Museums like halogen lamps, as they usually show the true colors of the objects on display.

There are many scenarios like the ones listed above, I won't go into all details. The bottom line is that there will always be markets, even if they are niche markets, for most lighting sources.

 

[RODALCO]

A couple of pics attached from one of the streetlights I have
2 clusters of 28 LED's
About 80 % of the streetlight casing is Aluminium heatsink.

 

[RODALCO]

This is from another streetlight I have
The controller board was missing.

 

[Flyback]

Thanks RODALCO..great pics!...will take a longer look when home from work

 

[MrAl]

Hi,

Another thing i did not mention yet is that i have had two LED flashlights 'fail' in a manner of speaking. What happened was with both designs the manufacturer overestimated the ability of the LED to survive over current and so the LED's started to fail, one at a time. But they did not fail completely right away, what happened was they started to blink on and off. That's really annoying.

I found that one of those lights had 50ma going through each LED, and those were 20ma LED's !
To top that off, the LED's were mounted in a plastic holder with a plastic lens over them, so the heat was trapped in the LED compartments.

In these cases though it was the manufacturer's fault not the LED.

What i usually ended up doing is ripping out the 'bad' LED with pliers, then replacing it later. Sometimes i add a resistor in series to get the current level down closer to a more acceptable level for that type of LED.

Amazingly, one of these lights was made by a well known manufacturer which really surprised me. But i think they updated their design on this model now.

I think maybe what Flyback was trying to get at was that LED technology might have reached a glass ceiling, or in this case a clear epoxy ceiling
We just have to wait for more significant advances.

 

[audioguru]

All light bulbs in my home are coiled compact fluorescent (except in the oven and fridge). They work fine except the 23W ones in the kitchen take one minute to reach full brightness when the air conditioning keeps the house cool.

I use (Name Brand) Osram Sylvania bulbs guaranteed for 5 years. I have had a few bulbs fail sooner, then a phone call to Osram Sylvania got me a coupon for free ones as replacements. One time they sent me a whole case of bulbs to replace one bulb.

Some people who hate compact fluorescent bulbs use very cheap Chinese no-name-brand bulbs that fail very soon. My electrical utility company gave away compact fluorescent bulbs for free but they were all recalled and replaced because they dripped burning plastic when they failed. The Chinese manufacturer stole the certification number from a competitor so the design was never certified.

A name-brand compact fluorescent bulb costs about $2.00. A no-name-brand LED light bulb costs $35.00 so there is no way I will buy LED bulbs yet, maybe never.

 

[MrAl]

Hi audioguru,

Wow that sounds really nasty. I have seen them break too when trying to remove them, that's nasty also.

Maybe that's the case with my bulbs too, as some have failed and some lasted for a longer time.

Try NewEgg. They have some very good sales on LED bulbs. But if you really like your curlies then maybe stick with that.
I still have to get a hold of some cool white units to try next.
I do find that the curlies work better at slightly lower line voltages than regular tube fluor's which i really can not use at all.

 

[audioguru]

I used to use 3000k pure white CFL bulbs but they are not made anymore, now have 3500k slightly warm pink color.
I do not like 5000k cool white color because it is blue light.

 

[MrAl]

I used to use 3000k pure white CFL bulbs but they are not made anymore, now have 3500k slightly warm pink color.
I do not like 5000k cool white color because it is blue light.

Hi again,

I am not sure i understand your post. Isnt 3000k 'warmer' than 3500k ? So i would think 3500k would be closer to 'pure white'.
I have standard curlies and they appear yellowish on a white ceiling.

 

[Flyback]

MrAl, may I ask if the one that failed was a paralleled led?
Also, please advise Was the company a well known warning light provider?...might have an 'e' somewhere in its name?

 

[audioguru]

I am not sure i understand your post. Isnt 3000k 'warmer' than 3500k ? So i would think 3500k would be closer to 'pure white'.

You are correct, I got the color temperatures backwards.
I tried a 3000k bulb beside a 3500k bulb just now. The 3000k bulb looks like it has a small amount of pink in the white light. The 3500k bulb looks cold and blue, not cool and not white.

 

[MrAl]

Hi again,

Fly:
This one light which was a big problem has separate resistors for each LED, but they were far too small for the type of LED being used and the heat dissipation factor was not worked out very well for the head design. I since then replaced LEDs with a high power unit and cut the lens off for better cooling.
The other light im not sure of because i did not attempt to repair it yet. I'll have to check tomorrow.

audioguru:
Wow that's weird. I thought you had to get up near 5000k before it started getting 'cool' or more white.

 

[Flyback]

RODALCO the street light driver looks interesting....As you say its 56 LEDs in total. It looks like theres 6 DCDC LED
driver PCBs there. It seems unusual to use so many separate boards but maybe they want that if one fails, the remaining
ones can keep going.
I can see the PFC stage, and then the main Switching converter, interestingly I don't see any resonant capacitors, but I
do see a secondary side big inductor, which suggests either Half Bridge, Full Bridge, or two transistor forward
converter..........
-probably say not half bridge as I cant see a current transformer.
-probably say not full bridge because I cant see enough transistors, and also the powers not likely to be high enough.
-So probably Two Transistor forward converter.....but I can't see the high side gate drive transformer, so that's also in
doubt.
-possibly a one transistor forward converter....but that's a devil of a converter to put after a PFC stage because of the
high FET OFF drain voltage, also it needs the third reset winding.
-Could be an active clamp forward converter though.
If we knew the controller IC then that may give it away.

 

[audioguru]

audioguru:
Wow that's weird. I thought you had to get up near 5000k before it started getting 'cool' or more white.

Scientifically, "pure white light" is produced from a clear blue sky with the sun is shaded. Obviously it is blue, not white. Or from a very hot tungsten filament.

Antique color charts show that red and orange light have low color temperatures. Yellow light has medium temperature and blue light has high color temperature.
I do not believe it since pure white contains all colors, not just blue.

2700k "warm white" looks very pink.
5000k "cool white" looks very blue.
As seen with the new lenses in my eyes (recent cataracts surgery).

Maybe people with original eye lenses that are old and that tint all light with yellow, see blue light as being white light? Then ask a young kid (or an oldie with cataracts surgery) if blue light looks white.

 

[RODALCO]

RODALCO the street light driver looks interesting....As you say its 56 LEDs in total. It looks like theres 6 DCDC LED
driver PCBs there. It seems unusual to use so many separate boards but maybe they want that if one fails, the remaining
ones can keep going.
I can see the PFC stage, and then the main Switching converter, interestingly I don't see any resonant capacitors, but I
do see a secondary side big inductor, which suggests either Half Bridge, Full Bridge, or two transistor forward
converter..........
-probably say not half bridge as I cant see a current transformer.
-probably say not full bridge because I cant see enough transistors, and also the powers not likely to be high enough.
-So probably Two Transistor forward converter.....but I can't see the high side gate drive transformer, so that's also in
doubt.
-possibly a one transistor forward converter....but that's a devil of a converter to put after a PFC stage because of the
high FET OFF drain voltage, also it needs the third reset winding.
-Could be an active clamp forward converter though.
If we knew the controller IC then that may give it away.

Flyback, I will make a few more close pics of the driver.
I am not sure what the photo limit is on this forum.
Found the traffic lights too , info will follow.

 

[RODALCO]

I have made some photo's from the Red LED cluster from which the LED's are in 49 lots in series of 4 LED's in parrallel.
Total 196 LED's are used.
LED Current is 124 mA

 

[Flyback]

Thanks Rodalco, that's interesting, a traffic light is only 15.4 Watts at the lamp....(124.4V * 0.124A)..that certainly is lower power than I thought.
The LEDs pass 31mA each (unless they are PWM dimmed), as long as the 4-in-parallel groups share the current with each other.
It seems bizarre that they didn't just put 4 parallel strings of 49-leds-in-series...and reap the benefits of statistical spreading of Vf variation, in order to assure (to a degree) equal currents in each led....they could have put a resistor in each of the 4 strings to stop any string from hogging too much of the overall current.

The back of the LED PCB shows that most of the bottom layer copper has been etched away.....maybe they recycle it to save money(?)....if say one of the leds in the 4-in-parallel 'unit' hogs most of the current, then it will overheat, and without a decent size copper 'pcb heatsink' to dissipate the heat, the led will likely overheat....so it seems unusual why they etched away the bottom copper.....also, parallel leds work better with good thermal coupling, so its odd that they missed the chance to better thermally couple the leds by removing the bottom copper.
(but maybe these are special leds which incorporate in-built resistance to aid parallel operation).

I suppose using many small leds in theory means that you don't need to glue the led pcb to a aluminium heatsink, because the heat is in theory spread out. ..so you save money on a heatsink.
When they produce such a product, I wonder how many led pcbs fail at the switch-on-and-test stage because some of the leds are much dimmer than others due to current hogging.(?)

I wonder if they have to leave them on soak test for several hours to try and precipitate any such "current hogging" failures(?)
-in the Far East, they have many staff who work for cheap, so maybe they can afford to have such "extended soak testing".

The 'all-through'hole' assembly suggests the Far Eastern manufacture.

So this looks like a cheap product from the Far East, but are the West mugs?, buying cheap products, that just fail prematurely? (or do they last a long time?)....some posters here have said they've seen that some leds in traffic lights they've seen have failed

 

[audioguru]

LED traffic lights in Canada are about 6 years old and you can clearly see that about 5 LEDs do not light in each light.
The dark LEDs are not in rows, they are random.