Powering T3.2 or T3.6 from a single LiFePo4 battery

Status
Not open for further replies.

christoph

Well-known member
Powering T3.2 or T3.6 from a single LiFePO4 battery

Hi all,

I'm looking at the schematics, trying to figure out how to supply a Teensy 3.2 or 3.6 from a single LiFePO4 battery. Some random things to start with:
  • The battery will have its own charger circuit with an extra USB connector (see https://hackaday.io/project/18041-lifepo4wered18650). I can get ground, D-, D+ and battery voltage from a header on the charger circuit. There is a pair of pads to connect a non-USB source.
  • Since D+ and D- are required for communiation with a host, there will be a USB cable plugged into the Teensy. D+ and D- are not broken out for the Teensy's device port (the bottom D+ and D- pads are for the host port) so I can't connect them to the battery charger's header.
  • LiFePo4 batteries are charged to about 3.6 V and discharged down to some 2.5 V or less.
  • Depending on how I connect the battery, I'll have to subtract voltage drop for Teensy's regulator (330 mV typ. @ 500 mA) and at least one diode (probably another 0.3 V) on the way.
  • The main microcontroller will operate down to 1.71 V.
  • I'll need an SD card. I think they will be happy with as low as 2.7 V.
  • I don't know the boot loader chip's voltage range.
  • Whatever other devices I decide to connect later has to cope with the resulting supply voltage situation.

Now there are a few ways to connect the battery:
  • VBAT on the Teensy is only for the RTC, right? So I'd ignore that pin right away.
  • VIN pin, cut VUSB: There will be up to 0.6 V voltage drop (regulator plus diode). That doesn't sound good.
  • 3.3V pin, cut VUSB:voltage might be too high when the battery is charged to 3.6 V since there's zero added voltage drop.
  • 3.3V pin + series diode, cut VUSB: With a series diode I could get about 3.3V from a battery that's charged to 3.6 V. Might be unsafe with low currents, though, since diode drop will be lower in that case.
  • 3.3V pin with up/down converter, cut VUSB: seems like a lot of effort for this.
  • 3.3V pin with another low dropout linear regulator, cut VUSB: might be better than just a diode
  • VIN pin, cut VUSB, remove and short on-board diode: might fail/damage the teensy while removing the existing diode. Creating a short there is probably also hard.

The simplest solution might be: Cut VUSB and connect the Teensy USB connector's 5V (VUSB) to the battery charger's input, so that I don't need the charger's USB port at all. Then I can connect the battery output to the 3.3V pin and hope for the best. A series diode might make things safer, but it limits the usable battery voltage range and usable capacity.

Any thoughts on this? What's a simple, reliable way of supplying the Teensy from a LiFePO4 cell?

Regards

Christoph
 
Last edited:
OneHorse's charger would have been a good choice for a LiPo battery, but this is a different chemistry. The MAX1551/5 chips stop charging at 4.2 V which is far above what a LiFePo4 battery would like (3.6 V).
 
Hello,

- 3.3V pin, cut VUSB:voltage might be too high when the battery is charged to 3.6 V since there's zero added voltage drop. ----> add a step up regulator 3,3V between
- VIN pin, cut VUSB: There will be up to 0.6 V voltage drop (regulator plus diode). That doesn't sound good. ---> add a step up regulator 5V between, e.g.: https://eckstein-shop.de/Pololu-5V-Step-Up-Spannungsregler-U1V11F5

Hint: teensy 3.2 has no serial diode before LDO like teensy 3.5/3.6 (see reference on PJRC.COM)


best regards
und gutes Gelingen
 
I ordered the board mentioned in the first post and hooked it up to a Teensy 3.6 with VUSB cut. No regulator.

My test sketch calculates Vdd by measuring the internal 1.2V reference and then extrapolating. So the measurement is not very accurate, but should give me a rough impression of what Vdd is. The sketch outputs 2000 mV dummy data to set the serial plotter's scale. Then the calculated Vdd is reported while the charger is not supplied with power. After timestep 110 I applied power to the charger and removed it again at about step 350.

First I connected it to VIN, this is the result:
Screenshot from 2017-10-02 21:26:38.png
A bit on the low side since the voltage drops over the Teensy's on-board diode and voltage regulator. Also not very smooth.

Now when connected to the 3.3 V pin:
Screenshot from 2017-10-02 21:45:23.png
Now it shows the actual battery voltage and the plot is a lot smoother (note: it's also scaled differently).

I will probably give this a go again with regulators as suggested by larry_berlin.
 
Here's the schematic for what I use to run a Teensy and a bunch of other chips off of a single 3.7V LiFePo4. It's the TPS61202, a Texas Instruments boost converter. With it, the battery can go as low as a .3 volts (yes, really), and it puts out a rock solid 5V. It also has a programmable undervoltage lockout (UVLO) where it stops regulating, in order to protect the battery. You can also get a 3.3V output version: the TPS61201. The only downside is that the chip package is about 3x3mm, with 10 pins and a thermal pad, so break out your stereo microscope.

If this chip does not work for you, there are literally hundreds of battery managment IC's out there.

TPC61202.jpeg
Edit: Sparkfun has one of these on a 'dev board'.
 
Last edited:
Status
Not open for further replies.
Back
Top