LED voltage vs. battery voltage (a way higher?!)

[ArminVanBuuren]

It disappoints me that even in case of a single CREE LED & the lowest possible current (700 mA), it still has serious overheating issues. I would be interested in knowing how such a beast works, then:

**broken link removed**

Btw what happens when the single CREE is overheating? The current lowers itself on the lowest possible value 700 mA? Not less, even if the overheating is very serious? And then the lifespan of the LED decreases, or what actually happens?

 

[Audioguru]

An overheating LED becomes destroyed then it does not work no more no more.
Heat decreases its forward voltage a little which might increase its current a little making it hotter. Your circuit sets the current, not the LED.

That is a lot of heat in a tiny space so of course a heatsink with a high surface area is needed to provide convection to the surrounding air.

I do not know if the cheap cheap Chinese junk sold on e-bay has proper cooling.

 

[erikl]

I do not know if the cheap cheap Chinese junk sold on e-bay has proper cooling.

22$ including 4 18650 cells, charger, headlight and bike fixtures isn't exactly cheap cheap but - IMHO - a fair price. The black anodized striated Al case looks like 5 K/W thermal resistance. It doesn't get "hotter" than about 20K over ambient in quiet air (case to ambient), correspondingly less moving on a bike. With its 700mA CC, the LED itself gets about 30K over ambient (measured) - there's no serious overheating at all.

I'm very content with it and its really bright light for many hours.

 

[Audioguru]

22$ including 4 18650 cells, charger, headlight and bike fixtures isn't exactly cheap cheap but - IMHO - a fair price. The black anodized striated Al case looks like 5 K/W thermal resistance. It doesn't get "hotter" than about 20K over ambient in quiet air (case to ambient), correspondingly less moving on a bike. With its 700mA CC, the LED itself gets about 30K over ambient (measured) - there's no serious overheating at all.

I'm very content with it and its really bright light for many hours.
View attachment 104042

Then it was probably designed in Germany, not China. But of course it was made in China. I wonder what they use for heatsink compound? Maybe the compound is from Germany.

Were the 18650 cells it has real ones or fakes? https://www.youtube.com/watch?feature=player_embedded&v=eOshOXcSkDA

 

[erikl]

I wonder what they use for heatsink compound? Maybe the compound is from Germany.

I really don't know, Audioguru.

Were the 18650 cells it has real ones or fakes? https://www.youtube.com/watch?feature=player_embedded&v=eOshOXcSkDA

Thank you for this YouTube link! The 4 18650 accus delivered with the lamp were blue (sic!) packaged cells labelled UltraFire SX 18650 3200mAh 3.7V . My measurement with 700mA discharge current from 4.2 down to 2.5V resulted in about 2800mAh.

Not too bad, at least no fake. The charger actually stops charging at 4.15 .. 4.2V.

 

[ArminVanBuuren]

rikl: The temperature measurement were conducted on the 1x CREE light, or 3x LED? If only 1 LED, do you think it also does not have any overheating issues w/ 3x CREE LEDs on it?

Thanks.

 

[Audioguru]

Not too bad, at least no fake. The charger actually stops charging at 4.15 .. 4.2V.

You got a good one.
My electrical utility company gave away compact fluorescent light bulbs so that people would not waste a lot of electricity heating incandescent light bulbs.
They were all recalled and replaced because they caught on fire and dripped burning plastic.
The Chinese manufacturer stole the certification number from a competitor then made and sold his very cheap (cheep, cluck cluck) and unsafe product.

 

[erikl]

rikl: The temperature measurement were conducted on the 1x CREE light, or 3x LED?

rmin ;-) : on the 1x CREE light

If only 1 LED, do you think it also does not have any overheating issues w/ 3x CREE LEDs on it?

Not with my torchlight, because this one has a much larger convection surface (= heatsink). But with a tiny lamp (which corresponds to just the top of my torchlight), I think it could get rather hot, burning your fingers should you touch it, at least in quiescent air. May be good enough cooling on your bike if you run at 30 km/h.

Anyway I think this 3*CREE LED source produces too bright light for biking on normal roads. You'd probably need to wear sunglasses ;-).

 

[Audioguru]

I think this 3*CREE LED source produces too bright light for biking on normal roads. You'd probably need to wear sunglasses ;-).

In Canada cars have daytime running lights. Most cars use dimmed headlights or dimmed turn signal lights. But Chrysler cars and their Jeep trucks use HIGH beams that appear so bright that I doubt they are dimmed. On a cloudy day they are blinding oncoming traffic. They should be recalled.

 

[ArminVanBuuren]

rmin ;-) : on the 1x CREE light


Not with my torchlight, because this one has a much larger convection surface (= heatsink). But with a tiny lamp (which corresponds to just the top of my torchlight), I think it could get rather hot, burning your fingers should you touch it, at least in quiescent air. May be good enough cooling on your bike if you run at 30 km/h.

Anyway I think this 3*CREE LED source produces too bright light for biking on normal roads. You'd probably need to wear sunglasses ;-).

I would use it for forest rides (with no other light source around), for traffic it is definitely too bright.

So do you think I should worry about overheating running the 3x CREE LEDs on 100 % for longer time, or not?

Thanks a lot

 

[erikl]

I would use it for forest rides (with no other light source around), for traffic it is definitely too bright.

So do you think I should worry about overheating running the 3x CREE LEDs on 100 % for longer time, or not?

I think it could burn your fingers, so be careful when touching the lamp after a long time shining. But I don't think it will damage the LEDs; they just will shine a bit less bright after a while because of decreasing light yield with increasing temperature. This occurs so slowly, however, that you probably won't notice it.

 

[ArminVanBuuren]

I think it could burn your fingers, so be careful when touching the lamp after a long time shining. But I don't think it will damage the LEDs; they just will shine a bit less bright after a while because of decreasing light yield with increasing temperature. This occurs so slowly, however, that you probably won't notice it.

And do you know approximate value of the current that the light (3x cree) uses? So that I can know how many hours would it work with 4x Panasonic 3400 mAh connected in 2S2P. I am thinking of buying the 3x CREE light for forests where is no other light around ...

 

[erikl]

And do you know approximate value of the current that the light (3x cree) uses? So that I can know how many hours would it work with 4x Panasonic 3400 mAh connected in 2S2P.

This 3x cree lamp doesn't work with a 2S2P power supply, it needs 3S1P (or 3S2P), if used in a DC CC circuit. I've already told you the operation time with a 3S1P supply in my previous answer #11 (at the end), pls. (re)read it!

 

[ArminVanBuuren]

Are you sure it is not going to work? They sell a pack like this here:

**broken link removed**

"Battery configuration: 4 x 18650 2S2P Batteries pack(included), Capacity:2800mAh"

... and I have seen it on bunch of other websites. Is it possible that the driver gets these values and then sends it to the LEDs themselves differently?

// after further research, I found out that almost ALL of the 3x CREE bike lamps require an 8.4V battery pack. I would conclude that the el. configuration is therefore 2S2P for it as well. But your argument is strong, could you please revise it once more?

 

[erikl]

You're right, Armin: these 3x CREE bike lamps are supplied by 3 single parallel currents, see: Circuitry 6-4-2A in your above link, which means 2-1.33-0.67A per LED, depending on the selected brightness.

The runtime is stated as 1.5 hours, which is a middle value. Calculation for a fresh 2S2P 18650 pack with 2800mAh nom. capacity and about 90% yield (at those discharge currents) results in a total capacity of 5100mAh (@ nom. 7.4V), which should allow for runtimes of 50'-1¼-2½ hours, respectively.

The bad news, again, is: only about 40% of the cells' power are used (from the medium discharge voltage of about 7V , the LEDs use just ≈2.9V), 60% are directly converted into heat by the current control circuit (if that is a pure DC CC controller, not an SMPS circuit. They don't state this, so I think it's a simple DC CC controller).

At the largest current (6A), a total power of about 40W is converted into heat. I wonder how hot this relatively small lamp case (=heatsink) will get ... I'm sure you can't use the middle and high brightness adjustment for a long time period. This probably would not only burn your fingers :-(

I'd suggest you also throw a glimpse on this **broken link removed** which I bought 2 years ago: it can also be installed on a bike of course, has a better heat convection, and can still be used as a hand torchlight. For efficiency and runtime data (700mA serial current), s. my answer #11.

And for comparison, see this **broken link removed**, stating to use a booster circuit (SMPS), which increases the efficiency to 80..90%.

 

[ArminVanBuuren]

And if you take a look here:

**broken link removed**

They also state there is the booster circuit, does it also imply 80 to 90 % efficiency? Also, when the efficiency is this high, does it therefore produce LESS heat <=> could work for longer on the mid / high settings?

I wonder if 6 A I out would not actually ruin the batteries! Especially those Chinese ones. I am planning to buy a pack of 4 Panasonic 3400 mAh if I manage to sort out all these troubles. But really, 6 A current output? For 2S2P, it's 3 A for a cell, right? This must damage them quite a lot :-(

 

[erikl]

**broken link removed**

They also state there is the booster circuit, does it also imply 80 to 90 % efficiency?

Yes, that's right.

Also, when the efficiency is this high, does it therefore produce LESS heat <=> could work for longer on the mid / high settings?

Sure. But here a short estimation, following the XM-L-T6-LED data sheet: 3A per LED means 10W per LED. With the cool white XM-L-T6-LEDs they put out a max. of 900 lm * 3 = 2700 lm, and not 4000 lm as is stated.

Thermal estimation: The XM-L-T6-LED's heat resistance between its light producing junction and its case R_th,j-c is 2.5 K/W, so at 10W the LED's junction is 25K hotter than its case. For 3*10W of heat convection, you would need a heatsink of at least the same (low) R_th,c-c. Let's suppose it actually has such low thermal resistance, then you'd get a temperature difference of 3*10W*2.5 K/W=75K=75°C , so you have a temperature difference of 75°C between the LEDs' cases and the outer surface of the lamp.

Now let's suppose you can keep the thermal resistance between the bike's case and ambient at only R_th,c-a=1K/W (needs high wind velocity!), at an ambient temperature T_a=25°C your bike lamp's outer surface temperature is 25°C+(30W*1°C/W)=55°C - burns your fingers when you touch it. The 10..20% power losses of the driver here aren't yet considered.

The LEDs' case temperature then is 55+75=130°C, the LEDs' junction temperature 130+25=155°C, 5°C more than the max. allowed 150°C. This means: only 70% light efficiency (compared to 700mA operation), means you just get about 1900 lm instead of the expected 2700 lm, and reduced life expectancy of the LEDs. I.e. you should only switch on this highest brightness condition for short time intervals. Similar, but for longer time periods the medium stage is usable. For long time, I'd use only the lowest brightness stage. That's ok. for long term usage and slow forest runs.

I wonder if 6 A I out would not actually ruin the batteries! Especially those Chinese ones. I am planning to buy a pack of 4 Panasonic 3400 mAh if I manage to sort out all these troubles. But really, 6 A current output? For 2S2P, it's 3 A for a cell, right? This must damage them quite a lot :-(

No, 3A per cell isn't a problem at all - they're good for even 20A.

 

[ArminVanBuuren]

This post is so enriching! I appreciate it much, thanks a lot.

One last thing. If I therefore run the 3x CREE light on the lowest settings, how would you compare it with running a single CREE LED on its highest setting? In the previous post, they state Circuitry 6-4-2A, does it mean 2A current for the lowest possible settings on 3x CREE? And is it the effective valule, or the maximal one (only before it gets heated heavily, then drops to 700 mA as in 1x CREE)? So how about the brightness, would it be 3x bigger, or more or less the same?

Thanks

 

[erikl]

Re: 1 LED @ 2A versus 3 LEDs @ 700mA

One last thing.

That's your 3^rd time "One last thing", I guess! ;-)

If I therefore run the 3x CREE light on the lowest settings, how would you compare it with running a single CREE LED on its highest setting? In the previous post, they state Circuitry 6-4-2A, does it mean 2A current for the lowest possible settings on 3x CREE? And is it the effective valule, or the maximal one (only before it gets heated heavily, then drops to 700 mA as in 1x CREE)? So how about the brightness, would it be 3x bigger, or more or less the same?

Ok; one more (last) estimation:

A single cool white XM-L-T6 CREE LED on its highest setting (2A) uses a voltage of about 3V (at 100..125°C junction temperature - yes, it will get so hot!), i.e. it uses 6W , resulting in a bike's case surface over temperature of 60K, i.e. 85°C @ 25°C ambient temperature at quiet air condition (no wind). Driver losses not yet implied. Under this condition, the luminous flux will be 692lm * 85% ≈ 590 lm (85% is the light yield at about 100°C LED junction temperature).

A single such CREE LED @ 700mA needs a voltage of ≈2.8V (@ ≈100°C junction temperature), i.e. an input power of 2W. Hence 3 of them need 6W, so we have similar thermal conditions as above, the single LED just getting 2W*2.5 K/W=5°C hotter than its inner case contact (instead of 15°C difference for the above single LED @ 2A condition). Also the outer case will be 10°C colder than in the above condition, 75°C at still air. Much higher life expectancy for the LEDs. Luminous flux (=brightness) for a single LED @ 700mA is 280 lm * 90% ≈ 250 lm, or 750 lm for all 3 of them.

Now I think, the decision is clear.

In the previous post, they state Circuitry 6-4-2A, does it mean 2A current for the lowest possible settings on 3x CREE?

2A for all 3 of them, i.e. ≈700mA per LED.

 

[D.A.(Tony)Stewart]

The benchmark for luminous efficacy is Lumens per Watt. The challenge to meet is the degrees per Watt in cooling capacity in the heat spreader design to limit Tjcn=85'C

This Chip comes in a wide range of efficacy by bin sorting (reflected in P.N.) so you can choose what you can afford from an estimated 60 to 150 LPW using 8.5W dissipation with their design guides and Lumen output attached below.


Your solutions appear to be in the nominal range <100 LPW (lumen per watt)

Operating at 125'C is a poor design option.