Single Cell LiFePO4 supply for Teensy 4

Status
Not open for further replies.
Hi all,

I am working on a project that requires power by single-cell LiFePO4. I have selected a number of chips for the supply/charging circuitry, but I'm having trouble making sense of what kind of power the Teensy 4 needs to operate properly.

Currently I have designed the circuit with the following:
-Battery Charger: MCP73123
-TPS65130: Dual supply DC-DC converter to supply some OP-amps in the circuit
-1-cell Fuel Gauge/Protector: DS2784

It is my understanding that I will need to cut the trace connecting Vin and power from the USB if I intend to charge the board externally. This poses two questions for me:

1. Do I need to have a boost to 5V to supply USB_VBUS if I intend to use the USB? The power to USB looks like it is hardwired to Vin, but I also see a little VUSB powerflag that makes me think maybe I can power that bus from USB and power the board externally from the battery. I'm sure this is plain as day to anyone reading this, but I nosed around on the forums and I wasn't completely clear about this.

2. What is the lowest Vin that the Teensy would run off if we forget about USB? The iMXRT1060 datasheet makes it look much lower than I think it really is. Let's assume for the sake of argument that there is nothing else externally in my circuit that relies on exact 3.3V logic. For example I have a differential input ADC that can do I2C with logic high essentially as low as 2.7V. If I can power the board with a lower voltage like this, I get rid of a bunch of power supply junk on my board.

I see from the i.MXRT1060 electrical specifications datasheet (pg. 20):

Table 7. Absolute maximum ratings:
---------------------------------------------------------------------------
Parameter Description Symbol Min Max Unit
----------------------------------------------------------------------------
Core supplies input voltage VDD_SOC_IN -0.3 -> 1.6 V
VDD_HIGH_IN supply voltage VDD_HIGH_IN -0.3 -> 3.7 V
Power for DCDC DCDC_IN -0.3 -> 3.6 V
Supply input RTC/storage VDD_SNVS_IN -0.3 -> 3.6 V
USB VBUS supply USB_OTG1/2_VBUS ---- -> 5.5 V
Supply for 12-bit ADC VDDA_ADC -0.3 -> 3.6 V
 
Looking at the schematic for Teensy 4.0 /4.1 you could power the 3V3 rail directly. I had a look at Chapter 16 (Power Management Unit PMU) in the i.MX RT1060 Processor Reference Manual. According to Figure 16-1 (Power system overview), you can power 3V3 directly. If you want to connect a USB cable for debugging, you'll have to cut the trace to VIN/USB 5V underneath the board.

Just to be sure, I tried this on a Teensy 4.1 I had waiting for me and powering directly to the 3.3V pin works - just as expected. I then cut the trace from USB 5V and I could program it as usual, but powered from a 3.3V source.

So to your questions:
1. Not unless you want to. Avoid VIN and USB_VBUS and power directly to the 3.3V pin instead.
2. Do not use VIN, but provide a clean 3.3V to the 3.3V pin. For stable operation, you should keep the voltage above 3.0V.

According to the discharge curve for LiFePO4, that should be no problem to maintain above 3.0V, but note that going above 3.6V on the 3.3V pin will cause trouble. If a LiFePO4 cell is fully charged, it's at 4V so you'll need an LDO to solve that.
 
If you power teensy 4.0 from 3.3v only, then it is quite likely that it will NOT power up.
Supplying VBAT with a backup battery may allow the Teensy to wake up when 3.3v used to power it.
 
@BriComp I don't know where you have that information from? If you look at the schematic, you'll easily see that VDD_SNVS_IN is powered from either VBAT or 3.3V. Applying either will wake up the SNVS LDO that powers the real time clock and SNVS blocks.

3.3V alone is thus all you need unless I missed something?
 
That information came from Paul on this Forum. The power needs to come up in a controlled manner. This is accomplished when powered by 5v but not 3.3v. Paul suggested having vbat supplied with voltage before 3.3v is up may be Ok
 
That information came from Paul on this Forum.
Got a link? I'm sure that could answer @drewhamiltonasdf question then :)

There is a section on power sequencing in the i.MX RT1060 Processor Reference Manual, but looking at the schematic, it looked like this was solved using capacitors.
 
No I don't have a link but you could look at Paul's posts. Just click on his name and select posts or whatever it is listed as.
 
Cool. So all that is needed is a small bodge-wire to that test point and it'll always work? Not too difficult.
 
Status
Not open for further replies.
Back
Top