On 2/6/2018 8:48 PM, Jeff Liebermann wrote:
On Tue, 6 Feb 2018 10:52:50 -0800, sms
wrote:

My wife's Lezyne Deca 1500XXL stopped taking a charge, at all. Taking it
apart, I saw that the batteries were made in July 2015. Not too good for
it to stop working that soon. These lights don't have user-replaceable
batteries, but by removing two screws I was able to open it, and the
battery pack does have a connector on it so at least they didn't solder
it directly to the printed circuit board.

I don't see a problem. If your wife used the light every day for
about 2 years, that would be 730 charge cycles. That's about the
correct lifetime for a 60% DoD (depth of discharge). See table 2:
http://batteryuniversity.com/learn/article/how_to_prolong_lithium_based_batteries

Not nearly every day. I'd estimate about 200 total charge/discharge
cycles. Something failed in the battery pack because it wasn't like the
operating time was less than when new, the pack would not charge at all.

Charging the batteries in parallel with a 2A USB charger is also about
right. For two alleged 2800mA-hr cells in parallel, that would be
about 0.35C charge which is quite safe.

I've found that the cells that die quickly are usually helped along by
a charge circuit that overcharges the battery. 4.2v should be the
absolute maximum. LiIon loses about 10% of it's capacity during the
initial rapid discharge from 4.2 to 4.0V. While I don't like losing
the 10%, the battery will last much longer if only charged to 4.0v.
Charging to 3.92v yields the best compromise between two failure
mechanisms (electrolyte oxidation and growing crud on the anode). See
"summary" section:
http://batteryuniversity.com/learn/article/bu_808b_what_causes_li_ion_to_die

The charge voltage was correct. And of course the protection circuit in
he pack ensures that the charge voltage doesn't exceed 4.2V, regardless
of what the charging circuit voltage actually is.

It's a 2 cell 18650 pack with the batteries in parallel, and a
protection circuit board shared between the two cells. The cells are
allegedly 2800mAH, for a total of 5600mAH.

Seems rather high. Note that batteries are tested at a 0.2C
discharge, which yields larger numbers than the usual headlight
discharge rate. The Lezyne Deca 1500XXL claims 1500 lumens, which
also seems a bit high. Assuming 120 lumens/watt at a nominal
3.7V/battery, that's:
1500 / 120 / 3.7 = 3.4A
to run the headlight at full brightness, or 1.7A per cell.
Meanwhile, the cell capacity is tested at:
0.2 * 2800 = 0.56A

This is the battery in the original pack:
http://www.gebc-energy.com/Uploadfile/pdf/ICR18650/ICR18650H3.pdf

I disassembled the pack. The batteries are completely discharged, 0V. To
me this indicates a failure of the protection circuit which normally
would not allow discharge below 2.8V.

The closest I could find
on-line was a 2x2600mAH parallel pack
https://www.amazon.com/dp/product/B003SH4BV6.

Notice that his version:
https://www.amazon.com/Tenergy-Li-ion-2200mAh-Rechargeable-Battery/dp/B002Y2LJW0/
is series connected, but claims only 2200ma-hr. Kinda looks like the
same cells, but wired differently. Looks like the one you purchased
might be lying about the battery capacity. You'll find out soon
enough if the battery is discharged faster than expected.

Do any manufacturers not lie? It's a matter of degree. Tenergy is an
industrial supplier of batteries with a real building in Fremont, and
not like buying no-name batteries on Aliexpress.

Ready to buy a battery discharge tester
http://www.westmountainradio.com/cba.php
and a lux meter?
https://www.ebay.com/itm/391973339920
https://www.ebay.com/itm/401324425264

I moved the connector from the old pack to the new pack, plugged it in,
and closed it up. Seems to work fine now.

Yeah, for how long will it work? Dig out your digital volts-guesser
and measure the EoC (end of charge) voltage. If it's over 4.2v, that
may be why it died early. Also, see how long it will run. You don't
want it going out prematurely on the initial test run. Maybe carry a
2nd headlight for the initial testing.

The charge voltage from the plug that connects to the battery was
4.19V. I checked that before I ordered a new battery because I thought
that the problem with the light might have been with the charging circuitry.

The measured lumens by mtbr.com of the Deca Drive XXL was 1390, so they
were not using chilumens
https://www.urbandictionary.com/define.php?term=chilumen, and there
was minimal lunmenflation
https://www.urbandictionary.com/define.php?term=lumenflation.

This morning I put the light on the maximum setting and it ran for just
about exactly two hours before it stepped down the intensity.

5200 mAH [rated]
3.7 V [rated]
19.24 Watt-Hours [calculated]
2.0 Hours [measured]
9.62 Watts [calculated]
2.6 Amps [calculated]
1390 lumens [measured by mtbr.com]
144.5 lumens/watt [calculated]
3 LEDs
48.2 lumens/LED [calculated]

Of course the reality is that the batteries were not fully discharged
when it dropped the output, so the lumens per watt was likely quite a
bit higher.

I like the design of the Lezyne and how they use 3 LEDs to mitigate
thermal issues, as well as to give a more usable beam pattern that
includes sufficient side illumination. Definitely NOT StVZO
https://www.urbandictionary.com/define.php?term=stvzo. What some
people don't understand is that optics are very important in bicycle
lights where you are dealing with a limited amount of battery or dynamo
power.

The Lezyne also shows you whether you're connected to a low-power USB
port (500mA) or a 2A USB port. The charging LED is green for low power,
blue for high power. You don't need their 2A charger, any 2A charger
will do. This is the one I use: https://www.frys.com/product/8335977.