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

[ArminVanBuuren]

Hello,

I am highly interested in this light:

**broken link removed**

However, finding CREE's XML T6 specifications, the typical voltage is only 3.35v:

**broken link removed**

The seller states that voltage input is 8.4V & includes such battery to it.
I have searched eBay & a lot of other ("more official") sites and they also offer an 8.4V battery pack for a single CREE XML T6 diode.

Would not it burn immediatelly?

Thank you.

 

[Audioguru]

The four lithium battery cells are each 3.2V to 4.2V. They might be connected in series-parallel.
The headlight has no details so maybe the light is extremely bright when the battery is freshly charged then gets dimmer and dimmer as the battery voltage runs down. Maybe there is a current regulator. Maybe this and maybe that. Who knows?
That is what you get from e-bay.

 

[erikl]

The four lithium battery cells are each 3.2V to 4.2V. They might be connected in series-parallel.

That's right.

Maybe there is a current regulator.

That's true, too.

I got one. Don't worry, it works fine.

 

[andre_luis]

Another possibility we could not check on above CREE datasheet refers the SOA curve, which could give information concerning to pulsed operation, what could explain supply with a voltage much above specified for continuous operation.

 

[ArminVanBuuren]

Thank you all for a fast reply.

I was just wondering. I need a high capacity pack & have contacted several battery-pack sellers. But when I asked for the configuration of eg. 12 Ah 8.4v battery pack, they stated that the actual capacity is a way lower.

If I purchased this housing for batteries:

**broken link removed**

And placed 4 Panasonic 3400 mAh batteries inside, would it provide capacity of 6.8 Ah and voltage 8.4v when fully charged? I contacted the seller & he said that the el. configuration is 2S2P (2 in series, 2 in parallel as mentioned already), so I think it should be like that. And should provide genuine 6800 mAh capacity from high-quality batteries, unlike those fake packs on eBay.

I just need somebody to approve this, as I'm a newbie in such things :-D

Thanks a lot.

The four lithium battery cells are each 3.2V to 4.2V. They might be connected in series-parallel.
The headlight has no details so maybe the light is extremely bright when the battery is freshly charged then gets dimmer and dimmer as the battery voltage runs down. Maybe there is a current regulator. Maybe this and maybe that. Who knows?
That is what you get from e-bay.

Yes, they are definitely connected in series-parallel. One cell is almost always 4.2v when fully charged, so this configuration doubles the voltage of the cell & the capacity two times.

But still, it does not explain how the single LED can withstand such enormous voltage, while there is only 3.35v in the datasheet. One of my friends who deals with electronics said that it would burn. But there are not such reactions on eBay; the lights works everyone, so I am quite confused. And want to figure out what the matter is :-D

 

[erikl]

... placed 4 Panasonic 3400 mAh batteries inside, would it provide capacity of 6.8 Ah and voltage 8.4v when fully charged? I contacted the seller & he said that the el. configuration is 2S2P (2 in series, 2 in parallel as mentioned already), so I think it should be like that. And should provide genuine 6800 mAh capacity from high-quality batteries, unlike those fake packs on eBay.

Yes, they are definitely connected in series-parallel. One cell is almost always 4.2v when fully charged, so this configuration doubles the voltage of the cell & the capacity two times.

I just need somebody to approve this

Yes, that's correct. With the current used in this case (s. below), the usable capacity is only 2*3200=6400mAh, s. the link at the bottom.

But still, it does not explain how the single LED can withstand such enormous voltage, while there is only 3.35v in the datasheet. ... I am quite confused. And want to figure out what the matter is

As mentioned above, the lamp includes a constant current (CC) driver with about 700mA (the recommended, normalized current), which guarantees a working time of more than 9 hours with freshly loaded accus. This also guarantees constant brightness over the full time, and a good exploitation of the accus' Ampère-hours capacity, because it will work with full 700mA current down to 2*2.5V voltage (then the light gets considerably darker, which is good, because the accus shouldn't be emptied any more). This is good news.

The bad news is: It's a DC CC driver, hence the accu voltage is always used with a 700mA load; the LED gets an electric input power of 3V*0.7A≈2W , the CC driver consumes the rest. At a medium accu voltage of 2*3.6V=7.2V this results in a (medium) CC power dissipation of about (7.2-3)*0.7≈3W , hence the driver's electrical efficiency is only 40% -- only 40% of the accus' energy is used

Of course, all this energy is converted into heat. 5W render the lamp's case rather (enjoyably?) warm.

For more info about these accus, see e.g. this posting

 

[ArminVanBuuren]

How do you know the bulb contains a driver that "tells it to consume" only 700 mA? Some other sources claim that the current required can be up to 3 A. So, like this, the 4 Panasonic cells would really provide it energy for 9 hours of constant bright light?

I have already heard about the overheating. That's why I chose only 1 CREE LED, but still 40 % efficiency is enormous waste. However, lots of people also choose the light with 3 CREE LEDs, which I probably find even more waste. Do you know about the current consumption for 3 CREE LEDs? Is it going to be 3x 700 mA?

Moreover, can the overheating seriously damage some parts involved in the circuit, eg the LED itself? When lighting eg. for 3 continuous hours?

Thank you.

 

[D.A.(Tony)Stewart]

If the 8.4V battery has 4 cells, it could be called a 2S2P array. The flashlight must integrate a LiPo charger and Buck regulator to drive the LED at 3.4V nom. White LED's have a threshold, Vth of 2.6V and ESR on this device starts at 0.3 and drops to 0.15 Ohms as I rises from 0.7 to 3A . This results in the range of Vf nom. shown in specs. They strive for the lowest ESR in production but process variation causes higher ESR which results in higher Vf at high currents.

 

[ArminVanBuuren]

So I should not worry about the 3.35v limit in the specs?

Btw why aint they connected in parallel resulting in 4.2v max U and 13.6 Ah capacity? Would not it be more efficient?

 

[andre_luis]

If somebody already have this product in hand, a good way for check if LED is driven in pulsed mode, would be sweeping a large surface to check is illumination occurs strobing.

As mentioned on post #4, this could explain supplying LED array with a voltage greater than specified for continuous operation.

 

[erikl]

How do you know the bulb contains a driver that "tells it to consume" only 700 mA?

Because I opened it and measured the current. The lamp looks exactly like the eBay picture you linked to.

Some other sources claim that the current required can be up to 3 A.

Sure it can. But the luminous flux 1. doesn't increase linearly with the current, and 2. it decreases with rising LED junction temperature, which is unavoidable with higher current. See the corresponding curves in the CREE PDF you linked above.

The main reason, however, probably is: you can't get rid of the dissipation heat at such a small surface: the lamp would get too hot, and this definitely limits efficiency and lifetime of the LED.

So, like this, the 4 Panasonic cells would really provide it energy for 9 hours of constant bright light?

Yes I'd think so. Mine has just two 3000mAh cells in series, only 2200mAh usable at 700mA (down to a 3.2V end-discharge voltage per accu only), and this works fine for 3 hours.

I have already heard about the overheating. That's why I chose only 1 CREE LED, but still 40 % efficiency is enormous waste. However, lots of people also choose the light with 3 CREE LEDs, which I probably find even more waste. Do you know about the current consumption for 3 CREE LEDs? Is it going to be 3x 700 mA?

No, this would be too much of waste. I have a torchlight with 3 CREE LEDs, it uses 3 18650 cells in series, and supplies the 3 LEDs in series connection, too. Hence a medium accu voltage of 3*3.6=10.8V supplies the 3*3=9V for the LEDs with 700mA - so the electric efficiency is much better: >80%. The overall dissipation results in ≈7.5W of heat, which doesn't warm the torchlight so much because of its larger surface and - last not least - its usual shorter on-time.

The disadvantage of this circuit is that you can use only about 80% of the Ah-capacity, because this method works only down to an end-discharge voltage of 3.2V per accu (instead of 2.5V for the full 100% capacity). 9V are needed for the 3 LEDs, 0.6V for the driver circuit. The rest of the accus' charge isn't lost, it's still there, but just can't be used. Decreases charge-time and increases lifetime of the accus.

So I think for a bike lamp it could still be reasonable, good for 3200mAh*0.8/700mA≈3½ hours, warming your fingers in wintertime. Your counter-traffic just won't recognize you as a bike, but as a motor-cycle ;-)

Moreover, can the overheating seriously damage some parts involved in the circuit, eg the LED itself? When lighting eg. for 3 continuous hours?

No, not with a 700mA current.

 

[D.A.(Tony)Stewart]

to examine efficiency, use units of watt-hours.

Buck regulators are more efficient than Buck Boost regulators where the Vout-Vin is small. Thus the 8V arrangement is better, single cell will drop below the maximum Vf of the white LED.

The power capacity of the battery is the same whether it is series or parallel, but in this case converter efficiency is best for 6~9V in and 3 to 3.6V out at various brightness levels.

 

[erikl]

So I should not worry about the 3.35v limit in the specs?

3.5V is the w.c. (max.) voltage at 700mA, 2.9V is typical, s. the CREE spec.

Btw why aint they connected in parallel resulting in 4.2v max U and 13.6 Ah capacity? Would not it be more efficient?

If you use the w.c. value from above, add 0.6V consumption for the driver circuit, you could discharge the accu down to (w.c.) 4.1V. In this case, you wouldn't get anything out of the accu, as the starting voltage of a freshly charged accu at 700mA is just 4.1V, see the Discharge, capacity curves from this test review:



It looks a bit better for typ. LED voltage values: 2.9V+0.6V results in an accu end-discharge voltage of 3.5V, which allows for the extraction of about 2000mAh of the 3400mAh accu, only about 60%.

Yet industrial volume production must consider worst case values, of course!

 

[ArminVanBuuren]

So overall would you recommend the 1x CREE light or with the 3 LEDs?

Also, how many hours would therefore run the single CREE with 2S2P connection of 4x Panasonic 3400 mAh?

Thanks

 

[erikl]

So overall would you recommend the 1x CREE light or with the 3 LEDs?

Depends on your requirements: for biking, the 1x CREE light is actually enough, it creates a really bright light.
For searching in a forest, the 3 LED version spreads a real hellfire, chasing away all animals ;-)

Also, how many hours would therefore run the single CREE with 2S2P connection of 4x Panasonic 3400 mAh?

About 9 hours, as I mentioned above. Makes full use of the 2*3200mAh available with 700mA current, but with an electrical efficiency of about 40% only.

May be in the future there will be buck/boost converters available with this lamp, pushing this efficiency to about 80%.

BTW: Don't forget: the LED's electric-to-light power efficiency is also in the order of 30..40% - nearly the best of all light sources!

 

[ArminVanBuuren]

Thanks for valuable information.

One last thing; it has been mentioned already, but I am not sure. How long would therefore last the 3x CREE light with 4x Panasonic 3400 mAh? You only said the current consumption is NOT 700 mA, but did not mention the value. Is it not 2100 mA then?

 

[erikl]

... I am not sure. How long would therefore last the 3x CREE light with 4x Panasonic 3400 mAh? You only said the current consumption is NOT 700 mA, but did not mention the value. Is it not 2100 mA then?

No, this is a misunderstanding:

... with 3 CREE LEDs, it uses 3 18650 cells in series, and supplies the 3 LEDs in series connection, too. Hence a medium accu voltage of 3*3.6=10.8V supplies the 3*3=9V for the LEDs with 700mA - so the electric efficiency is much better: >80%.

Means: the same current of 700mA runs through all 3 LEDs in series connection, and the power supply consists of 3*18650 cells, also in series connection.

The electrical efficiency of this configuration is >80%, however the charge capacity yield is only about 80%, because in this connection you can't discharge the accus more than this percentage. I.e. of the 3200mAh extractable charge at 700mA (from a 3400mAh accu), you can use only about 2500mAh, because after this amount the discharge voltage of the 3 cells gets too low for further operation. This increases the lifetime of the cells, however.

Runtime with this configuration (with 3 cells) is about 3½ hours. 4 cells doesn't make much sense in this configuration. Of course you could double the runtime with a 3S2P (6 cells) configuration.

It's all in the aforementioned answer, pls. read it thoroughly!

 

[ArminVanBuuren]

Thank you for your reply. Most of the 3x CREE lights are supplied with a 4 cell battery pack, though!

Btw. how does 1 CREE & 3 CREE cope with the overheating (supposedly higher @ 3x CREE)? Is the current lowered, or the temp remains, destroying the LED?

I still dont get the current consumption much. The 700 mA value is the least possible? Eg when it is almost completely discharged? And 3 A is the highest when the battery is charged completely?

Thanks a lot.

 

[Audioguru]

In an LED, the higher is the current then the higher is the brightness and the higher is the heat.
The absolute maximum allowed current in the Cree LED is 3A only if you are an expert and know how to cool it properly.
It will still get very hot at 700mA but you can calculate and select an available heatsink that will allow it to survive.

If you calculate a heatsink for 1 LED then a heatsink for 3 LEDs will need 3 times the surface area.

An LED does not have a single operating voltage like a light bulb that are all made the same. LEDs cannot be made the same, their forward voltage is a range of voltages. You cannot pick a certain voltage, you get whatever they have available so look at the maximum forward voltage on the datasheet, multiply it by 3 then add a few volts for the current-limiting resistor. The total is the minimum battery voltage you need.

Lithium battery cells are 3.2V when almost dead and are 4.2V when fully charged. 3 x 3.2V= 9.6V and 4 x 3.2V= 12.8V which is the lowest voltage you want the LEDs to work at without being too dim. Then you can select 3 or 4 cells.

 

[D.A.(Tony)Stewart]

Good torches have great parabolic reflectors that act as a heatsink connector to a heavy-duty water tight aluminum battery holder and heat spreader that contains the LiPo Battery. THis keeps them cool and gives long life.

I have one such single LED unit with one 18650 LiPo cell that is exceptionally bright that uses PWM to regulate 3 ON states.


Note the top curve will shift left as junction temperature rises and lower one will drop 10~20% if jcn. is hot. These arelways done at 25'C junction on a cold plate.