On 11/15/2017 10:44 AM, Ian Field wrote:


"mike" wrote in message
news
On 11/14/2017 1:44 PM, Ian Field wrote:


"mike" wrote in message
news On 11/14/2017 11:22 AM, Ian Field wrote:
Just curious about the failure mode - the cells really weren't worked
that hard.

Thanks.
You're gonna have to be a LOT more descriptive about the 'pack'.
passive/active/protected?
What did it power?

Just a bare 2P 18650 pack powering an e-cig with a 1 Ohm coil.

The only electronics is a power MOSFET where there used to be a
microswitch for the fire button.

Charging is controlled by a shunt regulator, there is a feed SB diode so
the shunt failing short can't harm the battery.

That seems to contradict the statement the the only electronics is the
MOSFET. I don't know what to make of that... And something has to
control the MOSFET


Yes - a tact switch salvaged from the front panel of a Sky box.

You're saying that the shunt regulator failing open can't harm the
battery?

It helps if you read before replying.

A shunt regulator failing short circuit (the most likely fault) wouldn't
do the battery any good at all - A SB diode in the feed makes that
accident unlikely.

I wouldn't put a lot of faith in the assumption that your regulator
won't fail open. I suggest that a bad solder joint might be more likely
than an actual component failure.

The diode also solves the problem of disconnecting the power source while
the battery is connected.


But it adds another failure mode. A shorted diode will cause the battery
voltage to exceed the safe max voltage. I'd also worry about the forward
voltage vs forward current curve for the exact diode you're using.
Depending on the charging current and the capability of your diode
heat sink, the junction temperature might get well above ambient
and need to be considered.
A diode conducts all the way to zero volts. It's easy to assume that
the current is small enough to be ignored, but VERIFY that's the
case in your application. Look at the curves for your diode over
temperature. What's the maximum battery voltage if you left it
on charge for a week.

If you want to charge it as quickly as possible to the highest
capacity, typically 4.2V depending on chemistry, you might find
that it's not possible to do safely with a diode inline. Backing
the max voltage off 200mV to be safer seriously impacts run time,
but it's probably a good idea.

Another alternative is to set your shunt regulator to 4.2V and put
a resistor across the diode. You get quickly to 3.9V or so. Takes
longer for full charge but can be much safer.

Lithium batteries are usually charged right up near the destruction
limit. There isn't much margin for error.
If you have a limited supply of paranoia, charging lithium batteries
is a good place to allocate some.