• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 79 subscribers
  • Views 972 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

msp430G2232: Low Power Scenario

Blake Carpenter
Expert 6455 points
Part Number: MSP430G2232
Other Parts Discussed in Thread: MSP430WARE, MSP430F2274, MSP430FR2311

Hi, 

I have to sample an ADC at 1 KHz and then do some processing based on the ADC reading.

I have a power supply that can source 400uA max for the 430.

The whole design is to monitor 3 voltage rails, 1 with the ADC, the other two are simply digital on/off.  The ADC readings are for a scaled AC, so it will need to reconstruct the 60hz AC waveform based on the 1K Hz signal and estimate the peak.  It then sends a "you are drooping - better shutdown" GPIO to the big host processor.

For a G series (G2232), I can use a big cap on the micro to give it some holdup on the 3.3v so the 400uA can be supplemented by the capacitor if I wake, run at full speed, then go into sleep mode.

My question is based on the 1K Hz ADC sampling rate, should I run very slow and stay on all the time (and how slow do I have to run to meet a 400uA steady state worsts case), or run very fast and sleep most of the time waking 1000 times a second to sample the ADC?  This question centers around my 400uA max average power for the power supply.

It is a telecom application, so I have to meet worst case spec's for everything.

Best Regards, 

Blake

  • 0
    TI__Guru 56685 points
    Blake,

    Whats the PtP voltage of the signal being measured? It will need to be conditioned before getting to the MSP430 in order to sample including adding a DC offset so no negative voltage reaches the MSP430. If you can get the signal conditioned to a DC value, then you can use a simple ADC Compare functionality with an Xtal and get an average current of 5uA (max 40uA at sampling). This can be shown in the following ADC sample from MSP430Ware: msp430g2x32_adc10_12.c . Otherwise, you'll need to either A) do reconstruct signal in active mode to try to see if the peaks are lowering over time (high current cost), or B) implement an external resetting peck detector circuit that you cant hen do the LP ADC compare against.

    Now with this conditioning or the peak detect circuit, it mean some Op-Amps are needed. With that being the case, a MSP430F2274 or MSP430FR2311 might be a better fit here.
  • 0 in reply to Jace H
    TI__Expert 6455 points

    Hi Jace,

    No worries on the incoming signal - rectified and rescaled AC so the IO will only see 0-~2.5 VDC.

    Without an external crystal do we have a problem?  I can recommend the external crystal but would like to use the internal clock for cost.

    We will definitely use scenario A) where we do some basic signal reconstruction at the expense of some current.

    I guess my big question is do I need a crystal or not to hit the 400uA (in theory - since the code isn't written yet :).  I anticipate no more than 500-1000 clock cycles to analyze the ADC data.

    Best Regards, Blake

  • 0 in reply to Blake Carpenter
    TI__Expert 6455 points
    Also - I suspect a 32.768 kHz crystal would be what you recommend if the on-board is inadequate for power concerns?
  • 0 in reply to Blake Carpenter
    TI__Guru 56685 points
    Blake,

    Option A is going to be tough to do with your current budget. You would be stuck to using 1 MHz MCLK which means decreased response time. you may not have enough clock cycles to reconstruct signal and do an evaluation before having to get the next set of data. I'm not sure if you will be able to be at 1MHz plus your ADC current and be under 400uA as the max current at 1MHz, 3V is 400 uA (worst case). Now that max current could be a little lower at a different DVCC voltage, say 2.6V, to allow margin of your incoming ADC signal, but we don't have a current specification at that voltage.

    Since you have the signal already conditioned, instead of actively reconstructing the signal, you can get around some of the challenges above by just taking an average voltage. If for say you stay in LPM as long as possible, and fill a buffer of ADC signals to take an extended average. Maybe enough samples for 1-2 periods of your incoming signal. The average should be ~1.25V since the conditioned signal goes from 0-2.5v. If the average dips significantly, then you know you are losing power.

    I would recommend a 32738 LFXtal for power savings in LPMs. Mind you, you will be in LPMs most of the time.your average current for the MSP430 is going to be really low of the 400uA limit, but you could get some peaks during the averaging that's getting to the 400uA limit. Especially at 3.3V as 3V is getting close to that limit.

    Does that make sense?
  • 0 in reply to Jace H
    TI__Expert 6455 points
    Perfect! Just what I was looking for. Thx!

    Blake

**Attention** This is a public forum