I want to use a LiFePo4 cell as a supply for the msp430. For this purpose TI developed the TPS61221. I'm wondering if that's really needed if I can guarantee that the cell voltage is between 2 and 3.6V? Is it a problem if the supply changes its voltage? I know the frequency probably changes linearly with the voltage, but apart from that, are there any other problems?

I'm aware that it's probably a good idea to use this IC but I also want a very cheap solution for my msp430. Could you elaborate a bit on this topic? (For example: I want to measure the voltage of a cell using the same cell also as a supply, is there a ground problem if I don't have this IC?)



If you're confident that the battery voltage remains within the MSP430's specs you don't really need the regulator, but I would recommend to use one to have a clean power supply voltage. In any case, be sure to have a look at other components the MSP430 connects to. For instance when you drive a MOSFET the gate voltage may be below pinch-off and the FET won't do its job.
(I don't quite remember the specs, but I once tested how low the MSP430 would go, and found that it still worked at 1.3V. Don't take this value as a reference, though. Always refer to the datasheet!)

If you want to measure the battery's voltage you'll need an ADC reference voltage that's independent of that. For instance use an LDO to create a fixed ADC reference voltage of 2V, and divide the battery voltage by two before applying it to the ADC (to make it lower than the reference)

  • \$\begingroup\$ hmm, right.. but then it's probably better to use a proper supply and use the internal reference if I have to use an LDO anyway? \$\endgroup\$ – duedl0r Jul 6 '11 at 15:13
  • \$\begingroup\$ @duedlor - Yeah, it probably is :-) \$\endgroup\$ – stevenvh Jul 6 '11 at 15:17
  • \$\begingroup\$ @duedlor - In previous answers I already said I'm quite a fan of the Seiko S-812C LDO. Low-power-wise it would perfectly fit the MSP430. \$\endgroup\$ – stevenvh Jul 6 '11 at 15:22

In the following I'm using an MSP430F20.. as an example. Results should be identical or similar for most MSP430s, BUT check the data sheet for the part you are using!!! YMMV.

Most (not all) MSP430's have a 1.8 - 3.6 Volt operating Vdd range. Vdd is the absolute maximum operating voltage. Absolute maximum non operating voltage is typically 4.1 V.

If you want full rated clock operation you'll need a voltage near to 3.6V. As an example only, figure 1 page 20 in http://focus.ti.com/lit/ds/symlink/msp430f2011.pdf shows that you require Vdd >= 3.3 V and <= 3.6 V for full speed operation. Below 3.3V, max allowed clock speed drops linearly with supply voltage down to 2.2V min for programming or 1.8V min for operation.

If a LiFePO4 cell is moderately heavily loaded it will deliver most of its capacity above 3.0V. A MSP430 processor can be run at about 70% of rated maximum clock speed at this voltage.

Maximum voltage: A LiFePO4 has a maximum safe fully charged voltage of 3.6 Volt. It is possible but immensely unwise to charge them above this. Notionally a LiFePO4 may thus have a terminal voltage in excess of the MSP430 max operating voltage but this is extremely unlikely to occur or to matter in practice. If you wanted to be completely certain a shunt regulator could be used to bleed the small "tail" of voltage above 3.6V after the battery had been fully charged.

Minimum voltage: As above, the processor needs >= 3.3V for full speed operation. A lightly loaded liFePO4 (say loaded at no more than 10 hour discharge rate = C/10) will deliver almost all its energy at >= 3.3V. As loading increases terminal voltage falls for a given % of discharge. eg at C (1 hour rate) it would typically drop below 3.3V with around 80% of its capacity remaining.

A LiFePO4 will still be well above 1.8V or 2.2V when safely fully discharged.


  • if a LiFePO4 cell is used to directly power an MSP430, maximum voltage will be safe for most purposes and protection could be provided against battery overcharging.

  • If the battery is lightly loaded (C/10 or less) then the processor can be run at full speed for the whole of the battery life.

  • If the processor shares the battery with a device that loads it heavily (eg a motor or heater or ...) then reduced maximum processor speed will be required.

  • If both full speed operation and greater than C/10 battery discharge rate is required then a boost converter is required to maintain Vdd voltage.

If the controller is crystal controlled and if a supply independent ADC reference is available then operation directly from a LiFePO4 cell seems entirely acceptable.

  • \$\begingroup\$ Thanks for your answer. I think for me it's best to use more than one cell as supply using an LDO with 3.6V. Having this, I might not need a divider and can read the cell voltage from 0-3.6V \$\endgroup\$ – duedl0r Jul 7 '11 at 10:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.