Could someone explain how to calculate LM723 output voltage?

Discussion in 'Power Electronics' started by psrkallez06, Aug 19, 2016.

  1. psrkallez06

    Thread Starter Member

    Dec 17, 2015
    68
    4
    Datasheet: http://www.ti.com/lit/ds/symlink/lm723.pdf

    I'm having a hard time to figure out how to set the output voltage. I'm even confused where to put the divider resistors. Sometimes it's between pin 6 and ground. other times it is between current sense and ground. I was also wondering what the value of VREF is? for example if i want 12v out, what/where to i put the resistors?

    Or if i want an adjustable 2.8 - 12v, where to put potentiometer etc.

    Cheers!!
     
  2. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,515
    1,246
    The 723 can do many things, and sometimes combining them changes how they are done. What do you need the circuit to do?
    Output voltage range
    Input voltage range
    Output current range
    Output current limiting (yes/no)
    Output current limiting adjustable

    Vref is listed in the datasheet.

    ak
     
  3. psrkallez06

    Thread Starter Member

    Dec 17, 2015
    68
    4
    Going to take a second look in the datasheet then, I wanna use it as a variable power supply like in Figure 19. Positive Voltage Regulator (External NPN Pass Transistor) thanks!
     
  4. psrkallez06

    Thread Starter Member

    Dec 17, 2015
    68
    4
    Update: From datasheet it says [​IMG]

    How come it it doesn't show the "full" range? +2 to +37v?
     
  5. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,515
    1,246
    There are two basic linear voltage regulator circuits, and the 723 can do either one. Figure 19 has one flavor, a fixed voltage reference and a variable feedback network (note that Vref is tied directly to the non-inverting error amplifier input). With this circuit, the output voltage never can be lower than the combination of the reference voltage plus the Vbe drops of the output transistors. Figure 20 has the other flavor, a variable reference voltage and fixed feedback network. This can produce output voltages both above and below the reference voltage value.

    ak
     
  6. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,515
    1,246
    That requires a combination of the two circuits as shown. This might be a non-verbal warning from TI that such a combination might have problems. In theory, it will work.

    ak
     
  7. bertus

    Administrator

    Apr 5, 2008
    15,645
    2,344
  8. crutschow

    Expert

    Mar 14, 2008
    12,981
    3,221
    I see no inherent reason that a combination of the two circuits shouldn't work, using a resistor divider on the reference to reduce it from 7V to 2V and then having the resistor divider on the output to control the voltage over the full range.

    The LTspice simulation of such a circuit is shown below.
    It also is stable in the transient mode, but of course that's not a guarantee that the real circuit will work the same.

    What do you mean "non-verbal warning"?

    upload_2016-8-19_10-57-48.png
     
    Last edited: Aug 19, 2016
  9. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,515
    1,246
    Something learned the hard way with Vicor datasheets - if it isn't in print, it isn't supported. In this 723 case, it might be a quirk in the error amp compensation, instability caused by an internal crosstalk path, or whatever. I agree it should work - I've done it with an LM399 reference and an LM358; one amp as the error amp and one amp for current sense.

    I see you went with the fixed reference/variable gain approach. I tend to prefer the fixed gain/variable reference; fixed gain usually means fixed phase margin.

    ak
     
  10. ian field

    Distinguished Member

    Oct 27, 2012
    4,413
    782
    There are plenty of online tutorials.

    Its worth remembering that the 317 is just the same as any of the 78xx family of regulators - it just regulates at a much lower voltage (1.25V).

    Its as simple as calculating a voltage divider across the output to develop 1.25V across the upper resistor.

    From time to time; I find 78xx regulators used in exactly the same way as the
    The 723 is pretty old - the L200 came out since and is simpler to design around - but I'm not so sure even those are still available.

    The L200 is in a multi lead TO220 style package, so for small jobs you can get away without an external pass transistor.

    If you can use a switcher - the MC34063 has a couple of online calculators that work out the operating frequency and inductance value for you. It may at least be no more difficult than figuring out the 723, its more efficient and you can do step up as well as step down.
     
  11. crutschow

    Expert

    Mar 14, 2008
    12,981
    3,221
    That's certainly a consideration.
    The problem is that I don't see how to readily use one pot on the voltage reference to give the full output range.
    If the NI input has to be a minimum of 2V (which is my understanding of the circuit limits), then the NI reference adjustment range would 2V-7V for an adjustment range factor of 3.5.
    If you adjust the output to give 37V with 7V at the NI input, then then output range is 37v max. to 37/3.5 = 10.6V min.
     
    Last edited: Aug 19, 2016
  12. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,515
    1,246
    Spoiled by the 358.

    ak
     
  13. psrkallez06

    Thread Starter Member

    Dec 17, 2015
    68
    4
    Thank you everyone for your inputs and answers! And i do have a few LM317 laying around too, Right now i need "Crude" power supply with both voltage and current adjustment before i make a good one.

    Even tho LM723 is old, it's seems easier to get both current and voltage adjustment with that regulator, also you can just add a few transistors for higher output current.
     
  14. #12

    Expert

    Nov 30, 2010
    16,252
    6,749
    I cashed some paychecks working on LM723 designs and I'm here to tell you, you can get them to do about anything in the voltage department. The current limiting is not deadly accurate or repeatable. If you want pro quality current limiting, you need to add another op-amp.
     
  15. psrkallez06

    Thread Starter Member

    Dec 17, 2015
    68
    4
    Thanks for your answer. Like i said, i don't need it to be that accurate since I'm going to create a external current limiting circuit later and prolly use the LT3080 regulator. But out of curiosity, do you have an example circuit on where to add that second op-amp?
     
  16. hp1729

    Well-Known Member

    Nov 23, 2015
    1,938
    218
    I haven't tried the design yet, but can we put together both designs and do something like this. External bypass transistor is suggested. With lots of voltage in and little voltage out the LM723 may exceed wattage capability at a pretty low current. No current limiting is shown.
     
  17. #12

    Expert

    Nov 30, 2010
    16,252
    6,749
    No, but I'll make one just for you.
    Here's the basic shape of it.
    A differential amp creates a negative voltage proportional to the current through the sense resistor. Then you make a negative voltage with a pot and compare that to the amplified sense voltage. Then, that amplifier turns on a dumping transistor as much as it takes to defeat the LM723. You might send the dump line back to the zener input if that suits your purposes.
    This is my first try at this, so there might be a better way.

    There is a simpler way if you place the sense resistor between the load and ground. That just requires one op-amp, but you either lose the accuracy of the voltage setting or you connect your voltage reference for the LM723 at the point before the sense resistor. I could work that out if I tried.
     
    Last edited: Aug 20, 2016
  18. hp1729

    Well-Known Member

    Nov 23, 2015
    1,938
    218
    Sweet, but the LM723 does have current limiting built in. I just left it out.
     
  19. crutschow

    Expert

    Mar 14, 2008
    12,981
    3,221
    The current limit is determined by Rsc in this schematic.
    The limit current is approximately 0.65 / Rsc.
    Edit: The limit is shown for Rsc varying from 14Ω to 4Ω.

    upload_2016-8-20_12-15-41.png
     
    Last edited: Aug 20, 2016
  20. #12

    Expert

    Nov 30, 2010
    16,252
    6,749
    And there is the crux of the biscuit. The current limit depends on a base-emitter junction. Not completely repeatable because it changes with temperature, and it requires ~0.65 volts as the one and only voltage difference across the sense resistor. The circuit I drew is as repeatable as the quality of the voltage reference on the last op-amp and completely adjustable with one knob.

    ps, the drawing needs a diode on the last transistor, thus:
     
Loading...