SPY-BI-WIRE and battery power conflict

I do not see any reason to bother about voltage drops in prototyping system because you can always compensate voltage drop by increasing supply voltage, so low drop diode(s) will be fine. If you are building battery-powered mass-product for developers like EZ430-Chronos and you want flawless dual-supply solution then look for ideal diode (controller).

Alternate option would be to use msp430 as an ideal diode controller. You need low drop diode with parallel P-MOS that's controlled by msp430 pin (low state = P-MOS on). Gate of P-MOS shall be externally pulled-up by weak pull-up resistor. When processor detect that it runs on battery alone, it turns P-MOS on to short diode, thus eliminating voltage drop of diode.

Hi Kris,

Is there a reason that you want to provide power from the programmer even if the battery is connected? If you use the MSP-FET430UIF tool, and connect Vcc to pin 4 of the JTAG header (instead of pin 2), this is a voltage sense pin, so it will detect the voltage coming off of the battery and adjust all of its signals accordingly while the debugger is attached. Then you can just always be powering off of battery, even while debugging.

Regards,

Katie

Hi Katie,

  I am not using 4-wire JTAG since the tool is more expensive compared to a SPY-BI-WIRE tool. SPY-BI-WIRE does not have voltage sense functionality.

Kris

Hi illmars,

 My customers will also be programming the board (it is a msp430 + cc2520 design). I want to make the board as power efficient as possible. I don't want to waste energy in the diode if it can be easily avoided. Another requirement is that I want the power source selection to be automatic (no jumpers etc). 

I will also be putting a dc-to-dc converter between the battery supply and the board so that I can run the two chips at around 1.8 volts.

I like your alternate option. I am using the MSP430G2533 which has an internal Vcc measurement channel. I can use it to detect the power source and the use a GPIO to turn on the P-MOS. Only problem is I loose one more precious I/O pin.

Thx

Kris

rk39300 said:

I don't want to waste energy in the diode

But is it ok to waste power in:

rk39300 said:

dc-to-dc converter

Any converter means some power loss. Also consider that radio achieve max output power only on max supply voltage. You definitely shall run power efficiency calculations.

rk39300 said:

I loose one more precious I/O pin.

Which chip and it's package do you use?

rk39300 said:

  I am not using 4-wire JTAG since the tool is more expensive compared to a SPY-BI-WIRE tool. SPY-BI-WIRE does not have voltage sense functionality.

These are independent things. The FET430UIF supports both, 4-Wire JTAG and SBW, and independently of which one is used, it can adjust its output signals to an external supply or use its internal one.
Well, that's the reason why it is 'a bit' more expensive (not to mention the cool case, the diodes - and the fact that it supports all MSPs and free updates are provided as soon as a new MSP is released)

Hi Jens,

  My customers will be using the value line MSP430 Launchpad to program the boards. This Launchpad has spy-bi-ire only so I need a way to avoid conflict between Launchpad supply and coin cell.

Thx

Kris

ram krishnan said:

  My customers will be using the value line MSP430 Launchpad

Launchpad is customer-unfriendly device that does not have case, no ESD protection. Are you sure that this is right solution? Customer-friendly software update is no-brainer, no-extra-tools update.

Thou it's good question: what your customers are. If they are just few friends interested in electronics then perhaps LaunchPad is ok :D If they will be thousands of random buyers, then you will get problems with such a kind of solution.

Hi illmars,

  I am using the MSP430G2553 (28 pin).

 

  If my board is continuously running (no duty cycle) then I can save power by running the board at 1.8 volts through a dc-dc converter as explained in this doc (www.ti.com/lit/slyt356).

  I agree that if the board (msp430 + cc2520) is mostly sleeping, then the quiescent current will almost get doubled (couple of microamps) and that will halve the battery life.

 

  I will run the board at around 2.0 volts to get better rf output. Sensitivity remains same at low voltages.

Thx

Kris

 

Hi illmars,

 I agree with your comments.

At this stage, I am thinking that the Launchpad is the cheapest option. My customers will mostly be developers in India who will be using the board as a platform to develop their applications. I will be putting the Launchpad in an enclosure. I have been developing mesh software on top of 802.15.4 for the past 4 months and all my boards are still working fine.

In the end it is a compromise between cost and functionality.

Thx

Kris

ram krishnan said:

At this stage, I am thinking that the Launchpad is the cheapest option.

Disagree here. Because you easily can add fraction of LaunchPad cost to your board and upgrade msp430 to one having USB and much more memory.

ram krishnan said:

At this stage, I am thinking that the Launchpad is the cheapest option.

Officially, the LaunchPad is no option at all. It is nto meant to be a programmer for custom boards, it is meant as a standalone evaluation system. As such, it isn't suited for the application you want it to use for. Especially the mixed supply simply isn't supported.

The official way to program MSPs is the FET430UIF. And if your customers want to do serious development beyond experimenting with the LaunchPad itself, the FET430UIF is the tool to buy.