Power path control

Discussion in 'The Projects Forum' started by men8th, Jul 7, 2011.

  1. men8th

    Thread Starter New Member

    Jul 7, 2011
    11
    0
    Hi all,

    I have an application using the LTC4011 battery charger IC from Linear Technologies. I will be using this IC to charge 3 NiMH from a 24V external input. The batteries power a wireless chip (3.6V max), motor and some other stuff.

    The LTC4011 has "Power path control" (excuse me for using Linear's marketing terminology) which switches the power supply presented to the system load. If the external 24V is available the system will run off that whilst the batteries are charging. If the external 24V is not available then the batteries are used.

    My question is this: What is the best way to regulate the power supply presented to the system load. The wireless can only take 3.6V max therefore putting 24V across it will fry it. As this is a battery powered device the solution must draw a minimal current.

    Should I use a 3.6V linear regulator and allow it to drop out of regulation when the external power is removed? Will there be a power penalty for that?

    Schematic below[​IMG]

    Thanks in advance,

    Tim
     
  2. StayatHomeElectronics

    Well-Known Member

    Sep 25, 2008
    864
    40
    Linear regulators are very inefficient. Regulating 24v down to 3.6V will give you an efficiency max. of about 15%. Switching regulators can provide 80-90% efficiency.

    Look through the power management areas of linear tech, national semiconductor, and others to find parts and designs to meet your needs.
     
  3. men8th

    Thread Starter New Member

    Jul 7, 2011
    11
    0
    Hi,

    Thanks for your reply. I'm not bothered about the efficiency when running from 24v as that will come from the mains supply. I'm more concerned that if I put, say, a 3.6V regulator on the line then when the 24V supply is removed and the 3.6V battery kicks in the regulator will still consume current even though it has dropped out of regulation. Is this correct?
     
  4. StayatHomeElectronics

    Well-Known Member

    Sep 25, 2008
    864
    40
    I don't know what will happen if you disconnect the input to the linear regulator, especially when connected to the charging circuit. You could disconnect the regulator as well by using a 24v relay, or similar... The DCIN fets may provide enough isolation to be fine.

    A linear regulator will, however, dissapate a tremendous amount of power in the 1A or 2A charging schemes.
     
  5. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    It would be helpful if you could tell us how much current your 3.6v rated load will require.

    Whatever that current is, multiply that by 3.6v to get the power consumed in your load.
    Then, multiply the voltage source less the load voltage by the current:
    (24v-3.6)*load_current = regulator wattage.

    For example, let's say your load current is 100mA.
    3.6*100mA = 360mW
    (24v-3.6v)*100mA = 2.04 Watts of power; plus whatever the regulators' overhead current is.
    Your efficiency will be roughly 360mW/2.04 <= 17.6%.

    At any rate, you really need some kind of switching regulator to drive your load when your 24v is available - however, it will be difficult to figure out what would be suitable and cost-effective without knowing your load current.

    By the way, your Q3 MOSFET does not have a low enough threshold voltage; Rds(on) is not given below 4.5v. You will need a MOSFET with a lower threshold and better Rds(on) rating.

    You might consider a DMP2215L. Both Mouser and Digikey stock them. It's an SMT P-ch MOSFET, Vds=-20, Id=.216 Ohms max when Vgs=-2.5v, typically 165 though.

    And, the plumbing for Q3's gate will need to be changed, as the way it is now would fry the MOSFET due to exceeding Vdss.
     
  6. men8th

    Thread Starter New Member

    Jul 7, 2011
    11
    0
    Hi,

    The 3.6V load will require <100uA most of the time (the wireless chip will be asleep) but ~200mA for a couple of seconds each hour (when the motor is running).

    Thanks for suggesting a suitable mosfet. The ones that are in there now are place holders so you've saved me some work there.
     
  7. men8th

    Thread Starter New Member

    Jul 7, 2011
    11
    0
    OK, I've been reading data sheets and generally researching and come up with a new concept based on Linear App Note 68 Fig 10. Schematic below.

    The idea is as follows:

    When the 24V external supply is present D1 will conduct and Q2 and Q3 will be turned off by the external supply voltage applied to their gates. Two p-chan mosfets are required in series. If only Q2 was used then the body diode would conduct during battery charging when the voltage can be up to 4.5V.

    When the external supply is removed the gates of Q2 and Q3 go low and they start to conduct. D2 conducts during the transition to improve turn on time.

    Any comments?

    [​IMG]
     
  8. jmpraveen

    New Member

    Sep 22, 2011
    1
    0
    Good evening ,

    in my project i am using VDD & VCC as voltage vaiables. In this VCC is battery backed, When main power is present VDD is Supply, at same time when main power cut off means the power is supply thru VCC.But the Vcc supply is for some selected device .How can i make this
     
  9. ke0ff

    New Member

    Sep 13, 2011
    4
    0
    Why don't you connect your load directly to the batteries? The 200mA that your load requires shouldn't be a problem for the charging circuit to keep up with, and the batteries will essentially regulate the output voltage in concert with the charger (plus or minus a bit during the charging operation). A brute force limiter would be recommended (like a tranzorb or high power zener with a small limiting resistor) to protect the load in case the batteries are removed while the 24V is present. This would allow you to remove R5, D1, Q3, and C3 from your original schematic. If you try this, I'd recommend a kelvin style connection of the load to the battery (in other words, connect the load as close to the battery terminal as possible, preferably along a conduction path that avoids the charger current path(s)).

    Your second schematic is reminiscent of a switching scheme I tried recently with an LTC4100. I found that simpler is better.

    BTW, why does your AC sourced input have to be 24V? Just curious.
     
Loading...