Loading ...
Sorry, an error occurred while loading the content.

Re: [SeattleRobotics] Re: Xbee battery drainage

Expand Messages
  • Max Cato
    Kenneth, Great response, as usual. Although not the original poster, you gave me great pause for thought. If it were you, would you instead go with a switching
    Message 1 of 19 , Jul 24 9:30 PM
    • 0 Attachment
      Kenneth,

      Great response, as usual. Although not the original poster, you gave me great pause for thought. If it were you, would you instead go with a switching regulator, or would you go for something else entirely?

      Doug, one other thing you might consider: I know you're doing the design in house, and there's no reason you shouldn't. However, have you considered using the "off the shelf" low power "motes" out there? We used some of these in a wireless sensor nodes class that I took in college. http://bullseye.xbow.com:81/Products/Product_pdf_files/Wireless_pdf/MICAz_Datasheet.pdf

      -Max


      From: K Maxon <k_maxon23@...>
      To: "SeattleRobotics@yahoogroups.com" <SeattleRobotics@yahoogroups.com>
      Sent: Wednesday, July 24, 2013 4:57 PM
      Subject: Re: [SeattleRobotics] Re: Xbee battery drainage

       
       
      Hi Doug,
       
      It sounds like 3qty 1.5v alkaline cells are being used in series to present a nominal (under load) voltage of 4.5V to the input of the LDO.  At the same time the LDO is being asked to  produce a regulated 3.3V based on the part number selected.  The LDO requires a differential of input to output of an absolute minimum (from the data sheet) of 178mV.  This means the each battery can only discharge 341mV before the LDO will begin to become unstable and  start shunting current through the ground terminal as it struggles to maintain regulation during transients. [ 4.5 - (3.3 + 178) ] / 3 ~ 0.341   It is also important  to notice that although the 178mV differential is a gauranteed parameter, it is a MIMUMUM, which means that current rushes through the system return lines will change this potential as the battery discharges due to switching noise in the returns. (as the XBEE turns on / off the transmitter)  This will force the unit to shunt current in pulses as the differential input voltage is approached but still above the 178mV differential, draining the batteries significantly faster.
       
      The description of the circuit above (put together from the various  e-mails in this chain) appear to show that the design is attempting to take advantage of only a fraction of the discharge range of the battery.   In  practice, in many comercial  applications, alkaline batteries are typically drawn down to 800mV and agressive applications will go all the way down to 600mV.  In the application above, the useful range of the battery is only being drawn down to 1.16V, wasting a significant chunk of stored energy.  Combine this with operating the LDO right down to the bottom of the differential range with current swings in the return line and even more energy is being thrown away.  These two steps when  combined are throwing away nearly 60% of the energy in the batteries.
       
      Another issue is the choice to use a linear regulator.  The raw calculation (Vin - Vout) * current delivered = [ ( 4.5 - 3.3 ) * 36mA ] = 43mW are being disipated by the regulator while  the transmitter is active.  Consider that at that same  time 3.3V * 36mV = 118mW is being delivered to the load.  This shows that the choice of linear regulator is throwing away 36% of your battery life right off the top.  Said another way, although they are extremely simple to use and require almost no design expierence, if at all possible linear regulators should not be used in battery driven applications because they are horribly inefficent.  It is worth noting that the 36% mentioned here is best  case and is likely much worse if you have not provided a good amount of copper area to get the  heat flowing out of the die.  It is also  worth noting that the 36% loss mentioned here is compounded on top of the 60% loss mentioned above.
       
      Just something to chew on, but those three pieces, when combined are throwing away most of the energy from the batteries by design.  Hope that  helps take you a bit further down the road in your investigation and suggests a few further aspects to  research....
       
      -Kenneth

      From: Doug Leppard <doug.leppard.list@...>
      To: SeattleRobotics@yahoogroups.com
      Sent: Tuesday, July 23, 2013 5:50 AM
      Subject: RE: [SeattleRobotics] Re: Xbee battery drainage



      Thanks tony,
       
      The XBee works down to 2.1V but I am not sure if it will read the temperature IC correctly once it falls below 3.3v.  So I considered It dead once it falls below the 3.3v.  I watch the current drain coming off the batteries with a scope I see some spiking.  I expected to see variations with transmit but all I saw was it turn on and off with pretty much a square wave for the right period of time.
       
      It has been an interesting journey.
       
      Doug
       
      From: SeattleRobotics@yahoogroups.com [mailto:SeattleRobotics@yahoogroups.com] On Behalf Of Tony Mactutis
      Sent: Tuesday, July 23, 2013 1:51 AM
      To: SeattleRobotics@yahoogroups.com
      Subject: Re: [SeattleRobotics] Re: Xbee battery drainage
       
       
      A couple quick thoughts Doug.  Looks like the XBee wants to see 3.3v.  The regulator will need the input voltage to be a bit higher, maybe 3.5v minimum.  When your batteries fall below that, the circuit will at some point stop working even though they are not dead by the manufacturer's definition.

      Also, that 40 mA figure is probably an average.  The load from a digital circuit is not steady; it spikes as the transistors switch.  Your circuit will have decoupling capacitors to supply a transient current, but there may still be times (such as on startup) when the circuit demands a much higher current from the power supply.  Also, with a radio the current will be much higher when it is transmitting and much lower when it is receiving.

      If you try to draw a large current from a partially depleted battery, the voltage will droop and the circuit may malfunction, even though it seems ok when you check it with your meter. 
      On 7/20/2013 1:55 PM, Doug Leppard wrote:
       
      Getting closer to what I did wrong.
       
      First in reading  using the micro current reader, I had it on the wrong scale and therefore it was a wrong reading.  Sure enough XBee is about 36ma for about 38ms which is what the documentation said it should be.
       
      Second issue is that AAA Alkaline batteries last about 30 hours at 35ma.  If the current is less they average better but what the battery sees is a 36ma pulse therefore it must be calculated using the higher current.
       
      I am running one of the circuits with a resister to see how long it will last.  That will tell me more precisely how long it should last.
       
      Good news I can “fix” it in software.  I believe I can shorten the time the XBee is awake.  Also instead of taking readings every 15 seconds do it 2 minutes or more which will be ok for my application.  I will play with the parameters until I reach the year goal.
       
      Doug
       
      From: SeattleRobotics@yahoogroups.com [mailto:SeattleRobotics@yahoogroups.com] On Behalf Of jamericanfreddy
      Sent: Saturday, July 20, 2013 12:56 PM
      To: SeattleRobotics@yahoogroups.com
      Subject: [SeattleRobotics] Re: Xbee battery drainage
       
       
      Reading your data i have idea that might out whats wrong

      remove the xbee circuit and put a 40 ma load and 3 fresh new batteries
      and see how long it lasts (LDO and batteries)
      it seems thats it the LDO is the problem,

      SOMEHOW its a internal resistance of the LDO since XBEE is constant.

      ANOTHER idea try a different type of battery like NI-CADS at the same current output.

      --- In mailto:SeattleRobotics%40yahoogroups.com, "Doug Leppard" <doug.leppard.list@...> wrote:
      >
      > I have set up a circuit for remote temperature reading using the Xbee on the
      > zigby mesh. The batteries went dead multiple times faster than I calculated
      > it would. I expected it to be good for 1+year but it was more like weeks.
      >
      >
      >
      > Facts:
      >
      >
      >
      > XBee series 2
      >
      > Setup as an endpoint
      >
      > Sleeps for 15 seconds awakes and reads and transmits for 35-40 ms
      >
      > Uses 3 AAA batteries 1200 mah alkaline energizer E92, had other batteries in
      > other units
      >
      > Uses MCP1700 voltage regulator
      >
      > Uses TMP36 to read temperature
      >
      >
      >
      > Using the eevblog micro current reader
      > http://www.adafruit.com/products/882#Learn going into digit scope to see the
      > voltage and duration
      >
      >
      >
      > When XBee is in sleep mode is uses about 120 micro amps continuously
      >
      >
      >
      > When XBee is awake it uses 1.2 ma for about 38ms. Which is strange it is
      > supposed to use 40ma. I have taken current readings right off the batteries
      > and right off the module and get the same current reading.
      >
      >
      >
      > Either way the batteries to several units died in a matter of weeks not 1+
      > year as calculated.
      >
      >
      >
      > 120 micro amp continuos
      >
      > + 40ma @ 38ms/(15 sec * 1000ms) = 101 micro amps average
      >
      > 221 micro amps average use
      >
      >
      >
      > 1,200,000 micro ah/221 = 5430 hours = 226 days
      >
      > (note over a year by the measurement of 1.2ma instead of 40ma.)
      >
      >
      >
      > But I only got weeks.
      >
      >
      >
      > Anybody with a clue or experience in these areas?
      >
      >
      >
      > Doug
      >
       






    • K Maxon
        Hi Max,   Yes, likely a sepic  configured regulator, with an anlog interrupted operation  controller so  that  it  is  only  operating
      Message 2 of 19 , Jul 25 6:41 PM
      • 0 Attachment
         
        Hi Max,
         
        Yes, likely a sepic  configured regulator, with an anlog interrupted operation  controller so  that  it  is  only  operating intermittently for  a fraction of the  needed time.  This would also allow  for different types / configurations of cells.  I did not make  the recomendation in the original post as high efficency implementations of this type are going to need to be tuned at the  hardware level to the  particular implemenation and are not a buy it and drop it  in sort of implementation.  One of the easiest thing that could be done in Doug's case without taking on the challenge of the afore mentioned implementation would be to simply change out the LDO for one  with a wider input  range and then add one additional cell, allowing each individual cell to be drawn down further.
         
        There are a handful of other options to apply here, but I don't  have visibility on all of the design constraints so it likely isn't a good idea to try and second guess what Doug was trying to  do.
         
        -Kenneth

        From: Max Cato <maxsthekat@...>
        To: "SeattleRobotics@yahoogroups.com" <SeattleRobotics@yahoogroups.com>
        Sent: Wednesday, July 24, 2013 9:30 PM
        Subject: Re: [SeattleRobotics] Re: Xbee battery drainage



        Kenneth,

        Great response, as usual. Although not the original poster, you gave me great pause for thought. If it were you, would you instead go with a switching regulator, or would you go for something else entirely?

        Doug, one other thing you might consider: I know you're doing the design in house, and there's no reason you shouldn't. However, have you considered using the "off the shelf" low power "motes" out there? We used some of these in a wireless sensor nodes class that I took in college. http://bullseye.xbow.com:81/Products/Product_pdf_files/Wireless_pdf/MICAz_Datasheet.pdf

        -Max

        From: K Maxon <k_maxon23@...>
        To: "SeattleRobotics@yahoogroups.com" <SeattleRobotics@yahoogroups.com>
        Sent: Wednesday, July 24, 2013 4:57 PM
        Subject: Re: [SeattleRobotics] Re: Xbee battery drainage

         
         
        Hi Doug,
         
        It sounds like 3qty 1.5v alkaline cells are being used in series to present a nominal (under load) voltage of 4.5V to the input of the LDO.  At the same time the LDO is being asked to  produce a regulated 3.3V based on the part number selected.  The LDO requires a differential of input to output of an absolute minimum (from the data sheet) of 178mV.  This means the each battery can only discharge 341mV before the LDO will begin to become unstable and  start shunting current through the ground terminal as it struggles to maintain regulation during transients. [ 4.5 - (3.3 + 178) ] / 3 ~ 0.341   It is also important  to notice that although the 178mV differential is a gauranteed parameter, it is a MIMUMUM, which means that current rushes through the system return lines will change this potential as the battery discharges due to switching noise in the returns. (as the XBEE turns on / off the transmitter)  This will force the unit to shunt current in pulses as the differential input voltage is approached but still above the 178mV differential, draining the batteries significantly faster.
         
        The description of the circuit above (put together from the various  e-mails in this chain) appear to show that the design is attempting to take advantage of only a fraction of the discharge range of the battery.   In  practice, in many comercial  applications, alkaline batteries are typically drawn down to 800mV and agressive applications will go all the way down to 600mV.  In the application above, the useful range of the battery is only being drawn down to 1.16V, wasting a significant chunk of stored energy.  Combine this with operating the LDO right down to the bottom of the differential range with current swings in the return line and even more energy is being thrown away.  These two steps when  combined are throwing away nearly 60% of the energy in the batteries.
         
        Another issue is the choice to use a linear regulator.  The raw calculation (Vin - Vout) * current delivered = [ ( 4.5 - 3.3 ) * 36mA ] = 43mW are being disipated by the regulator while  the transmitter is active.  Consider that at that same  time 3.3V * 36mV = 118mW is being delivered to the load.  This shows that the choice of linear regulator is throwing away 36% of your battery life right off the top.  Said another way, although they are extremely simple to use and require almost no design expierence, if at all possible linear regulators should not be used in battery driven applications because they are horribly inefficent.  It is worth noting that the 36% mentioned here is best  case and is likely much worse if you have not provided a good amount of copper area to get the  heat flowing out of the die.  It is also  worth noting that the 36% loss mentioned here is compounded on top of the 60% loss mentioned above.
         
        Just something to chew on, but those three pieces, when combined are throwing away most of the energy from the batteries by design.  Hope that  helps take you a bit further down the road in your investigation and suggests a few further aspects to  research....
         
        -Kenneth

        From: Doug Leppard <doug.leppard.list@...>
        To: SeattleRobotics@yahoogroups.com
        Sent: Tuesday, July 23, 2013 5:50 AM
        Subject: RE: [SeattleRobotics] Re: Xbee battery drainage



        Thanks tony,
         
        The XBee works down to 2.1V but I am not sure if it will read the temperature IC correctly once it falls below 3.3v.  So I considered It dead once it falls below the 3.3v.  I watch the current drain coming off the batteries with a scope I see some spiking.  I expected to see variations with transmit but all I saw was it turn on and off with pretty much a square wave for the right period of time.
         
        It has been an interesting journey.
         
        Doug
         
        From: SeattleRobotics@yahoogroups.com [mailto:SeattleRobotics@yahoogroups.com] On Behalf Of Tony Mactutis
        Sent: Tuesday, July 23, 2013 1:51 AM
        To: SeattleRobotics@yahoogroups.com
        Subject: Re: [SeattleRobotics] Re: Xbee battery drainage
         
         
        A couple quick thoughts Doug.  Looks like the XBee wants to see 3.3v.  The regulator will need the input voltage to be a bit higher, maybe 3.5v minimum.  When your batteries fall below that, the circuit will at some point stop working even though they are not dead by the manufacturer's definition.

        Also, that 40 mA figure is probably an average.  The load from a digital circuit is not steady; it spikes as the transistors switch.  Your circuit will have decoupling capacitors to supply a transient current, but there may still be times (such as on startup) when the circuit demands a much higher current from the power supply.  Also, with a radio the current will be much higher when it is transmitting and much lower when it is receiving.

        If you try to draw a large current from a partially depleted battery, the voltage will droop and the circuit may malfunction, even though it seems ok when you check it with your meter. 
        On 7/20/2013 1:55 PM, Doug Leppard wrote:
         
        Getting closer to what I did wrong.
         
        First in reading  using the micro current reader, I had it on the wrong scale and therefore it was a wrong reading.  Sure enough XBee is about 36ma for about 38ms which is what the documentation said it should be.
         
        Second issue is that AAA Alkaline batteries last about 30 hours at 35ma.  If the current is less they average better but what the battery sees is a 36ma pulse therefore it must be calculated using the higher current.
         
        I am running one of the circuits with a resister to see how long it will last.  That will tell me more precisely how long it should last.
         
        Good news I can “fix” it in software.  I believe I can shorten the time the XBee is awake.  Also instead of taking readings every 15 seconds do it 2 minutes or more which will be ok for my application.  I will play with the parameters until I reach the year goal.
         
        Doug
         
        From: SeattleRobotics@yahoogroups.com [mailto:SeattleRobotics@yahoogroups.com] On Behalf Of jamericanfreddy
        Sent: Saturday, July 20, 2013 12:56 PM
        To: SeattleRobotics@yahoogroups.com
        Subject: [SeattleRobotics] Re: Xbee battery drainage
         
         
        Reading your data i have idea that might out whats wrong

        remove the xbee circuit and put a 40 ma load and 3 fresh new batteries
        and see how long it lasts (LDO and batteries)
        it seems thats it the LDO is the problem,

        SOMEHOW its a internal resistance of the LDO since XBEE is constant.

        ANOTHER idea try a different type of battery like NI-CADS at the same current output.

        --- In mailto:SeattleRobotics%40yahoogroups.com, "Doug Leppard" <doug.leppard.list@...> wrote:
        >
        > I have set up a circuit for remote temperature reading using the Xbee on the
        > zigby mesh. The batteries went dead multiple times faster than I calculated
        > it would. I expected it to be good for 1+year but it was more like weeks.
        >
        >
        >
        > Facts:
        >
        >
        >
        > XBee series 2
        >
        > Setup as an endpoint
        >
        > Sleeps for 15 seconds awakes and reads and transmits for 35-40 ms
        >
        > Uses 3 AAA batteries 1200 mah alkaline energizer E92, had other batteries in
        > other units
        >
        > Uses MCP1700 voltage regulator
        >
        > Uses TMP36 to read temperature
        >
        >
        >
        > Using the eevblog micro current reader
        > http://www.adafruit.com/products/882#Learn going into digit scope to see the
        > voltage and duration
        >
        >
        >
        > When XBee is in sleep mode is uses about 120 micro amps continuously
        >
        >
        >
        > When XBee is awake it uses 1.2 ma for about 38ms. Which is strange it is
        > supposed to use 40ma. I have taken current readings right off the batteries
        > and right off the module and get the same current reading.
        >
        >
        >
        > Either way the batteries to several units died in a matter of weeks not 1+
        > year as calculated.
        >
        >
        >
        > 120 micro amp continuos
        >
        > + 40ma @ 38ms/(15 sec * 1000ms) = 101 micro amps average
        >
        > 221 micro amps average use
        >
        >
        >
        > 1,200,000 micro ah/221 = 5430 hours = 226 days
        >
        > (note over a year by the measurement of 1.2ma instead of 40ma.)
        >
        >
        >
        > But I only got weeks.
        >
        >
        >
        > Anybody with a clue or experience in these areas?
        >
        >
        >
        > Doug
        >
         










      • Doug Leppard
        David, I do like the thought of using Supercaps. From the specs on the batteries they do better when in low current mode, even pulsing brings down their
        Message 3 of 19 , Jul 29 9:41 AM
        • 0 Attachment

          David,

           

          I do like the thought of using Supercaps.  From the specs on the batteries they do better when in low current mode, even pulsing brings down their ability to deliver.

           

          Doug

           

          From: SeattleRobotics@yahoogroups.com [mailto:SeattleRobotics@yahoogroups.com] On Behalf Of David Buckley
          Sent: Tuesday, July 23, 2013 3:57 PM
          To: SeattleRobotics@yahoogroups.com
          Subject: Re: [SeattleRobotics] Re: Xbee battery drainage

           

           

          Wondering what the result of low level current pulses might be on AAA cells I asked Google, nothing in the first three pages but I did come across this

          It might be .co.uk but the English in places leaves me mystified!

          However it does talk about adding SuperCaps in parallel with the cells so the SCs handle the current pulses and the cells just trickle out current.

          It may help, at least doing that you can eliminate the 36mA current pulses from the AAA cells and set their maximum current and thus check for bad cells or bad Data sheets.

          DAvid

           

           

           

          ----- Original Message -----

          Sent: Saturday, July 20, 2013 9:55 PM

          Subject: RE: [SeattleRobotics] Re: Xbee battery drainage

           

           

          Getting closer to what I did wrong.

          First in reading  using the micro current reader, I had it on the wrong scale and therefore it was a wrong reading.  Sure enough XBee is about 36ma for about 38ms which is what the documentation said it should be.

          Second issue is that AAA Alkaline batteries last about 30 hours at 35ma.  If the current is less they average better but what the battery sees is a 36ma pulse therefore it must be calculated using the higher current.

          I am running one of the circuits with a resister to see how long it will last.  That will tell me more precisely how long it should last.

          Good news I can “fix” it in software.  I believe I can shorten the time the XBee is awake.  Also instead of taking readings every 15 seconds do it 2 minutes or more which will be ok for my application.  I will play with the parameters until I reach the year goal.

          Doug

          From: SeattleRobotics@yahoogroups.com [mailto:SeattleRobotics@yahoogroups.com] On Behalf Of jamericanfreddy
          Sent: Saturday, July 20, 2013 12:56 PM
          To: SeattleRobotics@yahoogroups.com
          Subject: [SeattleRobotics] Re: Xbee battery drainage

           

          Reading your data i have idea that might out whats wrong

          remove the xbee circuit and put a 40 ma load and 3 fresh new batteries
          and see how long it lasts (LDO and batteries)
          it seems thats it the LDO is the problem,

          SOMEHOW its a internal resistance of the LDO since XBEE is constant.

          ANOTHER idea try a different type of battery like NI-CADS at the same current output.

          --- In SeattleRobotics@yahoogroups.com, "Doug Leppard" <doug.leppard.list@...> wrote:
          >
          > I have set up a circuit for remote temperature reading using the Xbee on the
          > zigby mesh. The batteries went dead multiple times faster than I calculated
          > it would. I expected it to be good for 1+year but it was more like weeks.
          >
          >
          >
          > Facts:
          >
          >
          >
          > XBee series 2
          >
          > Setup as an endpoint
          >
          > Sleeps for 15 seconds awakes and reads and transmits for 35-40 ms
          >
          > Uses 3 AAA batteries 1200 mah alkaline energizer E92, had other batteries in
          > other units
          >
          > Uses MCP1700 voltage regulator
          >
          > Uses TMP36 to read temperature
          >
          >
          >
          > Using the eevblog micro current reader
          > http://www.adafruit.com/products/882#Learn going into digit scope to see the
          > voltage and duration
          >
          >
          >
          > When XBee is in sleep mode is uses about 120 micro amps continuously
          >
          >
          >
          > When XBee is awake it uses 1.2 ma for about 38ms. Which is strange it is
          > supposed to use 40ma. I have taken current readings right off the batteries
          > and right off the module and get the same current reading.
          >
          >
          >
          > Either way the batteries to several units died in a matter of weeks not 1+
          > year as calculated.
          >
          >
          >
          > 120 micro amp continuos
          >
          > + 40ma @ 38ms/(15 sec * 1000ms) = 101 micro amps average
          >
          > 221 micro amps average use
          >
          >
          >
          > 1,200,000 micro ah/221 = 5430 hours = 226 days
          >
          > (note over a year by the measurement of 1.2ma instead of 40ma.)
          >
          >
          >
          > But I only got weeks.
          >
          >
          >
          > Anybody with a clue or experience in these areas?
          >
          >
          >
          > Doug
          >

        • Doug Leppard
          Max and Kenneth, Thank you so much for your input. I had to work on other projects and had ignored this list until now. On the motes idea. I am using the
          Message 4 of 19 , Jul 29 11:26 AM
          • 0 Attachment

            Max and Kenneth,

             

            Thank you so much for your input.  I had to work on other projects and had ignored this list until now.

             

            On the "motes" idea.  I am using the XBees because they are low power and cost effective.  I like the zigbee mesh network is working well for the most part for my project.

             

            Kenneth your summary of what I said was right on.  Yes you are right I am not using a portion of the battery abilities by only allowing the voltage to go to 1.1 instead or .8 or even lower.

             

            In future releases I might go with the four battery idea or change regulator.  I like the regulator I use because of its low power drain and cost.

             

            If you have other ideas let me know.

             

            Doug

             

             

             

             

             

            From: SeattleRobotics@yahoogroups.com [mailto:SeattleRobotics@yahoogroups.com] On Behalf Of K Maxon
            Sent: Thursday, July 25, 2013 9:42 PM
            To: SeattleRobotics@yahoogroups.com
            Subject: Re: [SeattleRobotics] Re: Xbee battery drainage

             

             

             

            Hi Max,

             

            Yes, likely a sepic  configured regulator, with an anlog interrupted operation  controller so  that  it  is  only  operating intermittently for  a fraction of the  needed time.  This would also allow  for different types / configurations of cells.  I did not make  the recomendation in the original post as high efficency implementations of this type are going to need to be tuned at the  hardware level to the  particular implemenation and are not a buy it and drop it  in sort of implementation.  One of the easiest thing that could be done in Doug's case without taking on the challenge of the afore mentioned implementation would be to simply change out the LDO for one  with a wider input  range and then add one additional cell, allowing each individual cell to be drawn down further.

             

            There are a handful of other options to apply here, but I don't  have visibility on all of the design constraints so it likely isn't a good idea to try and second guess what Doug was trying to  do.

             

            -Kenneth

             

            From: Max Cato <maxsthekat@...>
            To: "SeattleRobotics@yahoogroups.com" <SeattleRobotics@yahoogroups.com>
            Sent: Wednesday, July 24, 2013 9:30 PM
            Subject: Re: [SeattleRobotics] Re: Xbee battery drainage

             

             

            Kenneth,

             

            Great response, as usual. Although not the original poster, you gave me great pause for thought. If it were you, would you instead go with a switching regulator, or would you go for something else entirely?

             

            Doug, one other thing you might consider: I know you're doing the design in house, and there's no reason you shouldn't. However, have you considered using the "off the shelf" low power "motes" out there? We used some of these in a wireless sensor nodes class that I took in college. http://bullseye.xbow.com:81/Products/Product_pdf_files/Wireless_pdf/MICAz_Datasheet.pdf

             

            -Max

             

            From: K Maxon <k_maxon23@...>
            To: "SeattleRobotics@yahoogroups.com" <SeattleRobotics@yahoogroups.com>
            Sent: Wednesday, July 24, 2013 4:57 PM
            Subject: Re: [SeattleRobotics] Re: Xbee battery drainage

             

             

             

            Hi Doug,

             

            It sounds like 3qty 1.5v alkaline cells are being used in series to present a nominal (under load) voltage of 4.5V to the input of the LDO.  At the same time the LDO is being asked to  produce a regulated 3.3V based on the part number selected.  The LDO requires a differential of input to output of an absolute minimum (from the data sheet) of 178mV.  This means the each battery can only discharge 341mV before the LDO will begin to become unstable and  start shunting current through the ground terminal as it struggles to maintain regulation during transients. [ 4.5 - (3.3 + 178) ] / 3 ~ 0.341   It is also important  to notice that although the 178mV differential is a gauranteed parameter, it is a MIMUMUM, which means that current rushes through the system return lines will change this potential as the battery discharges due to switching noise in the returns. (as the XBEE turns on / off the transmitter)  This will force the unit to shunt current in pulses as the differential input voltage is approached but still above the 178mV differential, draining the batteries significantly faster.

             

            The description of the circuit above (put together from the various  e-mails in this chain) appear to show that the design is attempting to take advantage of only a fraction of the discharge range of the battery.   In  practice, in many comercial  applications, alkaline batteries are typically drawn down to 800mV and agressive applications will go all the way down to 600mV.  In the application above, the useful range of the battery is only being drawn down to 1.16V, wasting a significant chunk of stored energy.  Combine this with operating the LDO right down to the bottom of the differential range with current swings in the return line and even more energy is being thrown away.  These two steps when  combined are throwing away nearly 60% of the energy in the batteries.

             

            Another issue is the choice to use a linear regulator.  The raw calculation (Vin - Vout) * current delivered = [ ( 4.5 - 3.3 ) * 36mA ] = 43mW are being disipated by the regulator while  the transmitter is active.  Consider that at that same  time 3.3V * 36mV = 118mW is being delivered to the load.  This shows that the choice of linear regulator is throwing away 36% of your battery life right off the top.  Said another way, although they are extremely simple to use and require almost no design expierence, if at all possible linear regulators should not be used in battery driven applications because they are horribly inefficent.  It is worth noting that the 36% mentioned here is best  case and is likely much worse if you have not provided a good amount of copper area to get the  heat flowing out of the die.  It is also  worth noting that the 36% loss mentioned here is compounded on top of the 60% loss mentioned above.

             

            Just something to chew on, but those three pieces, when combined are throwing away most of the energy from the batteries by design.  Hope that  helps take you a bit further down the road in your investigation and suggests a few further aspects to  research....

             

            -Kenneth

             

            From: Doug Leppard <doug.leppard.list@...>
            To: SeattleRobotics@yahoogroups.com
            Sent: Tuesday, July 23, 2013 5:50 AM
            Subject: RE: [SeattleRobotics] Re: Xbee battery drainage

             

             

            Thanks tony,

             

            The XBee works down to 2.1V but I am not sure if it will read the temperature IC correctly once it falls below 3.3v.  So I considered It dead once it falls below the 3.3v.  I watch the current drain coming off the batteries with a scope I see some spiking.  I expected to see variations with transmit but all I saw was it turn on and off with pretty much a square wave for the right period of time.

             

            It has been an interesting journey.

             

            Doug

             

            From: SeattleRobotics@yahoogroups.com [mailto:SeattleRobotics@yahoogroups.com] On Behalf Of Tony Mactutis
            Sent: Tuesday, July 23, 2013 1:51 AM
            To: SeattleRobotics@yahoogroups.com
            Subject: Re: [SeattleRobotics] Re: Xbee battery drainage

             

             

            A couple quick thoughts Doug.  Looks like the XBee wants to see 3.3v.  The regulator will need the input voltage to be a bit higher, maybe 3.5v minimum.  When your batteries fall below that, the circuit will at some point stop working even though they are not dead by the manufacturer's definition.

            Also, that 40 mA figure is probably an average.  The load from a digital circuit is not steady; it spikes as the transistors switch.  Your circuit will have decoupling capacitors to supply a transient current, but there may still be times (such as on startup) when the circuit demands a much higher current from the power supply.  Also, with a radio the current will be much higher when it is transmitting and much lower when it is receiving.

            If you try to draw a large current from a partially depleted battery, the voltage will droop and the circuit may malfunction, even though it seems ok when you check it with your meter. 

            On 7/20/2013 1:55 PM, Doug Leppard wrote:

             

            Getting closer to what I did wrong.

             

            First in reading  using the micro current reader, I had it on the wrong scale and therefore it was a wrong reading.  Sure enough XBee is about 36ma for about 38ms which is what the documentation said it should be.

             

            Second issue is that AAA Alkaline batteries last about 30 hours at 35ma.  If the current is less they average better but what the battery sees is a 36ma pulse therefore it must be calculated using the higher current.

             

            I am running one of the circuits with a resister to see how long it will last.  That will tell me more precisely how long it should last.

             

            Good news I can “fix” it in software.  I believe I can shorten the time the XBee is awake.  Also instead of taking readings every 15 seconds do it 2 minutes or more which will be ok for my application.  I will play with the parameters until I reach the year goal.

             

            Doug

             

            From: SeattleRobotics@yahoogroups.com [mailto:SeattleRobotics@yahoogroups.com] On Behalf Of jamericanfreddy
            Sent: Saturday, July 20, 2013 12:56 PM
            To: SeattleRobotics@yahoogroups.com
            Subject: [SeattleRobotics] Re: Xbee battery drainage

             

             

            Reading your data i have idea that might out whats wrong

            remove the xbee circuit and put a 40 ma load and 3 fresh new batteries
            and see how long it lasts (LDO and batteries)
            it seems thats it the LDO is the problem,

            SOMEHOW its a internal resistance of the LDO since XBEE is constant.

            ANOTHER idea try a different type of battery like NI-CADS at the same current output.

            --- In mailto:SeattleRobotics%40yahoogroups.com, "Doug Leppard" <doug.leppard.list@...> wrote:
            >
            > I have set up a circuit for remote temperature reading using the Xbee on the
            > zigby mesh. The batteries went dead multiple times faster than I calculated
            > it would. I expected it to be good for 1+year but it was more like weeks.
            >
            >
            >
            > Facts:
            >
            >
            >
            > XBee series 2
            >
            > Setup as an endpoint
            >
            > Sleeps for 15 seconds awakes and reads and transmits for 35-40 ms
            >
            > Uses 3 AAA batteries 1200 mah alkaline energizer E92, had other batteries in
            > other units
            >
            > Uses MCP1700 voltage regulator
            >
            > Uses TMP36 to read temperature
            >
            >
            >
            > Using the eevblog micro current reader
            > http://www.adafruit.com/products/882#Learn going into digit scope to see the
            > voltage and duration
            >
            >
            >
            > When XBee is in sleep mode is uses about 120 micro amps continuously
            >
            >
            >
            > When XBee is awake it uses 1.2 ma for about 38ms. Which is strange it is
            > supposed to use 40ma. I have taken current readings right off the batteries
            > and right off the module and get the same current reading.
            >
            >
            >
            > Either way the batteries to several units died in a matter of weeks not 1+
            > year as calculated.
            >
            >
            >
            > 120 micro amp continuos
            >
            > + 40ma @ 38ms/(15 sec * 1000ms) = 101 micro amps average
            >
            > 221 micro amps average use
            >
            >
            >
            > 1,200,000 micro ah/221 = 5430 hours = 226 days
            >
            > (note over a year by the measurement of 1.2ma instead of 40ma.)
            >
            >
            >
            > But I only got weeks.
            >
            >
            >
            > Anybody with a clue or experience in these areas?
            >
            >
            >
            > Doug
            >

             

             

             

             

             

             

          Your message has been successfully submitted and would be delivered to recipients shortly.