Creating negative rail supply - 2

Discussion in 'General Electronics Chat' started by AnalogKid, Nov 14, 2017.

  1. AnalogKid

    Thread Starter AAC Fanatic!

    Aug 1, 2013
    5,726
    1,626
    In this thread:
    https://forum.allaboutcircuits.com/threads/creating-negative-rail-supply-with-op-amps.142170/

    a. I missed all the fun.

    b. The TS is partly not at fault because of an "error" in the LM358 datasheet. The package pinout drawings have pin 4 (the negative rail pin) labeled as GND. This has created confusion in several threads over the years.

    The language and attitude (all of which I can only infer) notwithstanding, it doesn't surprise me that an inexperienced person would think that one of the characteristics of a single-supply opamp is that they take care of the negative rail bias internally. Especially since some parts actually do that.

    The LM358 has a non-traditional input stage (for its time) and runs with only one power rail better than many of its contemporaries. BUT it is a traditional opamp in most respects, and will run just fine with bipolar power supplies.

    ak
     
    JohnInTX likes this.
  2. OBW0549

    Well-Known Member

    Mar 2, 2015
    1,975
    1,641
    Which parts? It's an interesting idea and I can see situations where it would be useful, but I've never encountered any op amps that do that.
     
  3. ian field

    AAC Fanatic!

    Oct 27, 2012
    5,634
    993
    Pretty sure I've seen at least 1 data sheet - but AFAICR: never met one in person.
     
  4. RichardO

    Well-Known Member

    May 4, 2013
    1,748
    617
    Conceptually simple for a micropower op-amp. All that is needed is a charge pump. I would be interested in seeing such a device as well.
     
  5. OBW0549

    Well-Known Member

    Mar 2, 2015
    1,975
    1,641
    I've used one op amp, the LTC1152, which has a charge pump to create an internal positive supply rail above the +V supply input (as a way of achieving rail-to-rail input capability); but I've never seen one that generated its own negative rail.
     
  6. RichardO

    Well-Known Member

    May 4, 2013
    1,748
    617
    I need to get the data sheet on that.

    Somewhat related. I wish I was the guy at Maxim that came up with the idea of putting a charge pump on an RS-232 converter. The MAX232 was briliant at the time. :D
     
  7. MrChips

    Moderator

    Oct 2, 2009
    14,707
    4,363
  8. OBW0549

    Well-Known Member

    Mar 2, 2015
    1,975
    1,641
    Maxim has also been selling the +5V-to-±10V charge pump from the MAX232 separately, as the MAX680/681. There's a newer version, the MAX864, which operates at higher frequency and can supply more output current.
     
  9. MrChips

    Moderator

    Oct 2, 2009
    14,707
    4,363
  10. OBW0549

    Well-Known Member

    Mar 2, 2015
    1,975
    1,641
    Interesting parts. Thanks.
     
  11. JohnInTX

    Moderator

    Jun 26, 2012
    2,979
    1,448
    I linked the closed thread to here. Hopefully that TS will be able to benefit from all of your help when he returns.
    Thanks.
     
  12. AnalogSwitch1986

    New Member

    Wednesday
    3
    0
    The OP thought that single supply Op Amps generated negative voltages inside since they only require one supply input rather than 2. That's what caused the confusion.

    I'd like to know what happens if you attach say 12V to the Vcc pin of the LM358 and say -12 to the GND pin. Will the LM358 take this as a single supply from 0 to 24V, or will it take it as a dual supply ? It seems to me it will not see it as a dual supply ? I think it will see it as a single supply from 0 to 24V, and if you try to get a negative output from it it will crash like a CPU ?
     
  13. MrChips

    Moderator

    Oct 2, 2009
    14,707
    4,363
    Opamps don't crash, they saturate. In other words, the output reaches the limits of the supply rails and can go no further.

    Opamps don't know or care if you are using a single or dual supply. What it sees is a lower supply voltage at V- and a higher supply voltage at V+. What is more important is the voltage applied to the inverting and non-inverting inputs.

    I believe what is tripping the TS is his mathematical superiority which suggests that if the formula for an inverting amplifier is:

    Vout = -Av x Vin

    where Av is the voltage gain,

    and if Av = Rf/Rin = 2
    and Vin = 3V
    then Vout = -6V

    The reality is the output cannot go below V-.

    btw: Check the data sheet for LM358. You will find that pin-4 is specified as GND/V-. It is the negative supply for dual supply configurations.
     
    philba and OBW0549 like this.
  14. OBW0549

    Well-Known Member

    Mar 2, 2015
    1,975
    1,641
    No, it will just behave like any other op amp that is given +12V and -12V supply voltages. There's nothing special about the LM358's GND pin: as on all other op amps, it's simply the negative supply pin. The part can be operated on a single supply or a dual supply, and the GND pin does not have to be connected to ground.
     
  15. MrChips

    Moderator

    Oct 2, 2009
    14,707
    4,363
    The important part that needs to be discussed is what you choose as a COMMON reference point at which all inputs and output voltages are referenced. It does not have to be ground.

    Here are three ways to create a reference beside GND.

    1. Use a pair of resistors to create a voltage divider. Follow this with capacitor filtering.

    [​IMG]


    2. Use the voltage divider and an op-amp to create a low impedance voltage source.

    [​IMG]


    3. Use an IC virtual ground such as TLE2426.

    [​IMG]



    Reference: https://tangentsoft.net/elec/vgrounds.html
     
    OBW0549 likes this.
  16. AnalogKid

    Thread Starter AAC Fanatic!

    Aug 1, 2013
    5,726
    1,626
    The MAX232 RS-232 interface chip probably is the most popular/familiar part with internal voltage generation. But dynamic rams were doing this in the early 70's, and most EEPROM-based parts (CPLD, FPGA, etc.) do this. As above, a ton of Linear Tech parts have boost charge pumps for n-channel FET gate drive. MAXIM has video mux and driver parts that make ground-clamped video with only one supply.

    ak
     
  17. dl324

    AAC Fanatic!

    Mar 30, 2015
    5,254
    1,155
    If you look at a sampling of opamp datasheets, you'll see that they reference input and output voltage range to the V+ and V- supplies. Exceptions would be opamps, such as the LM358, that were designed to operate from a single supply.

    LM358 will operate from split supplies but, because it was optimized for single supply operation, the output stage is different than dual supply opamps. To minimize crossover distortion for split supply operation, you need to have a constant load to the negative supply.
     
  18. bertus

    Administrator

    Apr 5, 2008
    17,288
    3,079
    Hello,

    Also a LTC3265 can be used to create the dual powersupply:

    LTC3265_typical application.png

    Bertus
     
  19. DickCappels

    Moderator

    Aug 21, 2008
    4,050
    1,164
    You can get the output of an LM358 / LM324 to swing below ground with a single supply by cheating a little.

    upload_2017-11-15_23-23-22.png
    Test circuit

    upload_2017-11-15_23-23-35.png
    With triangle wave input

    It is only good for a few hundred mv but it allows handling of a signal at ground without noise clipping and the attendant rectification effects.
     
  20. hrs

    Member

    Jun 13, 2014
    134
    23
    I'm still a bit fuzzy about point 1 and 2. When do you bypass the virtual ground to both rails and when do you only bypass to the most negative rail? Does it depend on whether you reference the signal to virtual ground or to true ground?
     
Loading...