I tried to think about this a little bit.

One easy point, for us it's ok to charge the battery to 4.2 V. We
shouldn't be storing our batteries for long at elevated temperature.
AFAIK a day on the pad in the hot sun at 4.1 V vs 4.2 V is negligible in
terms of capacity change. Although it's true that higher CV voltage may
reduce cycle life slightly, this should be a non-issue for us since the
HAP battery shouldn't cycle much and we can theoretically monitor and
replace it on a schedule.  Also some newer batteries are much better in
this respect than the older cells were, so i would expect our batteries to
last a couple of years.


I think either of the chips you mention can be got to work.

On the plus side the LTC4067 has a timer, a thermistor interface, and an
"ideal diode" control output. I like that it allows monitoring of the
input current and the battery charge current separately.

It's interesting that a separate current limit is applied to the input
pin. I don't think we need that, but it still seems useful.

On the bad side, the overvoltage protection is only good to 13 V, which
is too low to be useful for us. It also lacks an enable pin, though this
can be worked around with only one transistor on the program pin.

Something i really don't like is that there doesn't seem to be a thermal
resistance specification. That seems bad for a linear regulator. --Wait,
i take it back, the theta-JA is given as 43 C/W. It's in only one
place in fine print next to the package drawing on datasheet p2.


The LTC4085 (i'd use the 4.2 V version) is a bit fancier chip. It
supports dual power inputs, intended for wall-adapter and USB power. For
our application it seems like we'd pick one and ignore the other.

The LTC4085 seems to have all the good pin functions of the '4067 plus
some extras, plus slightly better thermal performance (40 C/W) probably
due to the slightly larger package.

The '4085 is only good to 7V, but that's not really worse than the '4067
because either one will blow up at our bus voltage.


To get integrated overvoltage protection you'd have to go to something
like the TI bq25040, which is good to 30 V. The up-sides of the '25040
are built-in overvoltage protection, programmable charge termination
current, and actual thermal resistance ratings on the datasheet.

To be fair, linear probably has thermal resistance ratings on it's
packages somewhere on it's website, but not on the datasheet. An
unattributed footnote under a package drawing doesn't cut it.

The big downsides of the '25040 are it's lack of thermistor or timer
circuits. It also doesn't have power path stuff, but i don't see that as
a big deal since we can push enough current through the charger to power
everything downstream without a separate power path.

The missing thermistor function could be implemented on the
microcontroller, but previously we hadn't wanted to do that. Since the
'25040 has an enable pin, thermistor function is a pretty easy add-on
using something like the LT6700 to build a window comparator. The lack
of a timer is annoying, but in principle can be handled on the micro and
should not present an excessive safety issue for us.


I also considered the LTC4090, which is a synchronizable switching
charger capable operating with input voltages as high as 40 V. It's not
a super efficient switcher, but it presents the interesting possibility
that a HAP circuit could be built without the SPS buck converter. The
LTC4090 would replace the SPS and convert rocket bus power directly to
battery charge current.

The '4090 includes circuits for thermistor, timer, and power path

I think a '4090 solution would be more complicated to design since it
involves yet another switcher, but it might save both space and power
assuming it would work at all.


As to preference, there's more than one way to do it ;) Forced to pick at
this moment, i think the LTC4085 is the cleanest solution.

(2009.04.26) kenz...@yahoo.com:
> I think I have the new charger IC for the HAP narrowed down to two linear 
> ICs, datasheets can be found here:
> http://www.linear.com/pc/downloadDocument.do?name=4067f.pdf
> http://www.linear.com/pc/downloadDocument.do?name=40851f.pdf
> Any thoughts regarding a preference??I like these better than the other 
> choices I've looked at because they: 
> 1) have functional NTC inputs.
> 2) have built in ideal diodes and/or controllers for?power path swtiching (no 
> need for external ideal diode controllers).
> 3) have good sized thermal pads (which is important because they are linear 
> battery chargers not switchers).
> 4) have nice complete datasheets.
> So far I prefer the 4.1V float of the LTC4085-1 since it will provide for 
> longer shelf life on the battery. On the other hand?I like the built in OVP 
> of the LTC4067, though it comes shy of the 30volts I would like in that part 
> of the circuit.
> I'll build the package for both now, then choose which to put into the 
> circuit tonight, probably?doing the circuit design on Monday..

psas-avionics mailing list

Reply via email to