Best power good circuit?

Discussion in 'General Electronics Chat' started by oleoleo2, Sep 18, 2015.

  1. oleoleo2

    Thread Starter Active Member

    Jan 3, 2009
    33
    0
    Hey Guys.

    I am working with some various designs incorporating several microprocessors. They have be able to run from a 9V battery as well. What is the best way to build a power good function that shuts down when the battery can't deliver the voltage to make the voltage regulator stay at 5V?

    I have tried with a couple of different voltage regulators with build in power good function, but by my experience they often don't seem to have a "sharp enough" limit. Does anyone know of good voltage regulators for that purpose?

    - It has to be very silent and of cause not draining the battery all to fast, but the first one is very critical.
     
  2. pwdixon

    Member

    Oct 11, 2012
    488
    56
    Many processors can detect power problems and provide a signal to turn off the power and shutdown the PSU, you could start there I guess.
     
  3. GopherT

    AAC Fanatic!

    Nov 23, 2012
    5,997
    3,755
    Check if your processor has power off for Brownout feature so you can shutdown in software.

    Alternatively, if your Microcontroller has ADC, you can measure the battery's voltage (using voltage divider).

    Or, Microchip and several other companies make "voltage supervisor" chips. These can be ordered by 0.1V increments (if I remember correctly).
     
  4. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,520
    1,247
    Oh, please, keep us in suspense! Since you have measured the voltage monitor limits of some devices and compared them to the performance parameter you need, what is that parameter? What is "sharp enough"? And, when the power monitor "shuts down", what powers the monitor to keep the power off?

    Let me rephrase your question:

    Guys, I bought some stuff (part numbers are secret), measured them (method is secret), didn't like the results (results are secret) because they did not meet my needs (the BIG secret) and I need something (circuit, ebay Chinese modules, ???) that does what I want (method, environment, production quantity, for work, school, home - ALL secret) and fits within my skill set and budget (you guessed it - secrets). Ready,,,,go.

    Not to be grouchy or anything, but, dude, wtf?

    By definition, we are here to help people who know less than we do; that's the essential function of a question-answer forum. But you need to think about how to ask a question of a professional engineer who is ready and willing to help for free, but doesn't know *anything* about you or your application. If you were getting the question cold as we are, what would you want to know to provide an answer?

    ak
     
    Last edited: Sep 18, 2015
    Sinus23, KJ6EAD, #12 and 2 others like this.
  5. Papabravo

    Expert

    Feb 24, 2006
    10,137
    1,786
    The circuits on the silicon of the microprocessors are as good as it gets. Most modern processors will continue to run down to very low values of Vcc, much lower than +5V. There are special purpose voltage monitor circuits (MAX809, MAX810) that will place the processor in RESET when the voltage drops below some threshold and I think you will have a very difficult time doing better than that.

    http://www.nxp.com/documents/data_sheet/MAX809_MAX810_3.pdf
    http://datasheets.maximintegrated.com/en/ds/MAX803-MAX810Z.pdf
     
  6. ISB123

    Well-Known Member

    May 21, 2014
    1,239
    527
    Modern CPU runs at voltages between 1.2-1.5V that is VCore but I think that IO works at 3.3V.
     
  7. BR-549

    Well-Known Member

    Sep 22, 2013
    1,981
    388
    I am studying a chip that runs on 3.3 to 3.7 volts. It’s 32 bit, but handles 16 bit and 8 bit hardware and programming quite nicely.

    This chip monitors both chip core voltage and the peripheral voltage. Normally it resets with the min on data sheet.

    But one can control the voltage point monitored, and generate an interrupt instead. The interrupt can change the chip mode the operate in the 2.4 volt range.

    You can drop from the 2.4 volt range to a 1.7 volt range in the same way.

    After all that, when it goes to sleep, it only uses nano amps. If you lose power completely..........................

    The chip also has a internal Real Time Clock circuit, powered by an external button battery. This RTC has some memory in it.

    On the last interrupt, before sleep, areas of the chip and hardware memory can be stored in the RTC memory.

    This data can be restored during the wakeup or startup routine.

    I think that’s pretty damn good.

    With all that said, there are other strategies to your problem.

    I paid 10 bucks for the study board.
     
  8. oleoleo2

    Thread Starter Active Member

    Jan 3, 2009
    33
    0
    Sorry guys I'm so slow! Thought I would have email reply, telling me about the posts! Thank you very much for the inputs I will go through right away!
     
  9. oleoleo2

    Thread Starter Active Member

    Jan 3, 2009
    33
    0
    I am working with AVR MCU's and other digital electronic circuits. I haven't found a way with these processors to do it well enough (?)
     
  10. oleoleo2

    Thread Starter Active Member

    Jan 3, 2009
    33
    0
    They have ADC's but the MCU is running at eg. 5V and I wan't to shut down the power before I reach the limit that makes the MCU unstable, then I would not rely on the internal ADC. And an "unregulated" voltage divider would just follow the voltage as it decreases. I could use a 2.5V regulated reference and I might, but that would consume some extra energy from the battery. I will check out Michrochip (I think I have been around these and tried them out, but will give it an extra shot.
     
  11. oleoleo2

    Thread Starter Active Member

    Jan 3, 2009
    33
    0
    I also have other circuits running like Relays at 5V, that won't work properly under 5V. So, I would like a circuit that shuts down the power of all devices at the same time at a specific Voltage (5V in these cases first of all). Ultimately a circuit that has a higher "turn-on" limit than "turn-off". I can imagine some ways of doing that. I have already something that works, but is cumbersome and not as good as I want it. I will look at the MAX'es!
     
  12. oleoleo2

    Thread Starter Active Member

    Jan 3, 2009
    33
    0
    I am looking for inspiration and "hands-on experience" in general, of this wide area. You don't HAVE to contribute if you don't want to..
     
  13. oleoleo2

    Thread Starter Active Member

    Jan 3, 2009
    33
    0
    Update: I am using a LD29150PT50R LDO in a particular design right now and use the INH input as on/off. I started out by using a APX803 (DI), as a reset circuit for the LDO, but I discovered that using a passive voltage divider on the INH input actually gave me a sharper limit (I think it's around 300mV). I would like it to go down to at least 100mV, but the narrower the better, and as it get narrower the chance of some noise making it flip on and off, increases. If it it possible to make a circuit (or find an IC) that latches the "off" function until a certain voltage is present again I guess it would be the best solution.
     
  14. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,520
    1,247
    A comparator circuit with hysteresys can act as a bistable latch. You said you want to protect 5 V regulator output, so what you need to know is the regulator headroom, or input-output differential. This the minimum voltage drop across the regulator. Let's assume that your LDO requires 0.5 V min. A hysteresis circuit could have a Vlow trip point of 5.5 V and a Vhigh trip of 7 V. So the output would be latched off at 5.5 V, and would not be re-enabled until the input rost to at least 7 V. Many single supply opamps will function with Vcc down to 3 V.

    ak
     
Loading...