LM358N power supply requirements / specs

Discussion in 'General Electronics Chat' started by Tonyr1084, Apr 22, 2016.

  1. Tonyr1084

    Thread Starter Active Member

    Sep 24, 2015
    531
    86
    According to Digikey this dual op amp can be supplied from 3 ~ 32 v (or ±1.5 ~ 16 v). I might assume the split rail supply would mean that Vcc is positive and Vee is negative. My question does not revolve around the Vcc and Vee it revolves around "How do I split the power supply if the chip has no midpoint ground? I'm imagining I'm going to use my double isolated secondary transformer / BR (bridge rectifier) for my Vcc / Gnd / Vee but the op amp inputs will be grounded at midpoint level (ground - not Vee). Is my thinking sound or is it muddy and confused?

    Here's where I'm beginning to go with this. I'm not ready for further details just yet, I'm still working up powering this op-amp. Pins 3 & 5 will be grounded inputs (inverting amp in the works)

    OP AMP Power Supply.png
     
  2. cmartinez

    AAC Fanatic!

    Jan 17, 2007
    3,555
    2,517
    The midpoint of your transformer should be connected to ground, and the LM358N does have a ground at its #4 pin.
    Knowing more about your application would be really helpful. I find it rather unusual to use unregulated voltage to feed the OpAmp's power pins.

    EDIT: I think I understand your question now. The LM358 can be used with split power supplies, by using its ground pin connected to the negative side of the supply, just like you've shown in your diagram.
     
    Last edited: Apr 22, 2016
  3. cmartinez

    AAC Fanatic!

    Jan 17, 2007
    3,555
    2,517
    Here's some pretty interesting reading on the subject. Halfway down the page you will find info relevant to your question.
     
  4. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,516
    1,246
    The LM358 does *not* have a dedicated ground pin. Most analog opamps don't, because there is no standard signal relationship with ground when you consider the tens of thousands of possible analog signals, from EKG sensors to boiler thermocouples to audio to whatever. Digital logic circuits of all stripes and certain analog signals such as composit video do have a defined or preferred relationship with ground, and the chips for those applications have ground pins.

    The LM358 is unique as one of the very first IC opamps designed for single supply operation. That does *not* mean it cannot function in a traditional split supply circuit. The single supply stuff is nothing magical. It is a new (at the time) input differential amplifier design using PNP transistors that enabled linear operation even when the input voltage approached ***(here's the fun part) went BELOW the negative supply potential. Very cool now, outstanding in 1971. The output stage also was different, ans could swing close enough to the negative rail that IF the negative rail was a digital logic ground, THEN the output could drive a TTL gate correctly. So the LM358 (and the quad LM324 and the LM339 and 393 comparators) with the first big-time parts to operate near one rail, 10-20 years before today's "rail-to-rail" parts.

    That's the lecture. Actual advice later.

    ak
     
    cmartinez, OBW0549 and GopherT like this.
  5. OBW0549

    Well-Known Member

    Mar 2, 2015
    1,301
    880
    This is a crucial point: no opamp, "single supply" or otherwise, gives a fat rat's derriere where "ground" is in relation to the voltages on its positive and negative supply pins. All it cares about is what voltage(s) its inputs are at relative to the supplies. Opamps (such as the LM358 and LM324) whose inputs can function at voltages all the way down to the voltage at the (-) supply terminal are designated "single supply" opamps simply because this ability makes operation from a single, positive supply a lot more convenient than it would be with an opamp such as the 741.

    That diagram looks good to me, though you might want to consider whether operating the opamp from unregulated supplies will give you the results you need; after all, opamps don't have an infinite ability to ignore their supply voltages.

    Not true. The datasheet shows an opamp pin labeled GND, but that doesn't mean it has to be connected to ground; in fact, it's just the opamp's negative supply connection.
     
    cmartinez likes this.
  6. dannyf

    Well-Known Member

    Sep 13, 2015
    1,771
    359
    Digikey doesn't mean squad in this context.

    Depends on the application. Sometimes, you don't need a "ground", or even fixed potential, for the opamp to work. Without knowing how you intend to use it, it is hard to help you more.

    My personal favorite is to use an opamp to split the rail. It works for small current situations, and with a buffer, even high current applications too.
     
  7. cmartinez

    AAC Fanatic!

    Jan 17, 2007
    3,555
    2,517
    Yeah, I figured that much after I read some more on the subject.
     
  8. Tonyr1084

    Thread Starter Active Member

    Sep 24, 2015
    531
    86
    Not yet ready to go public with a schematic; so perhaps a mere verbal explanation of the goal will do.

    I want to power a Laser beam (cat / dog toy laser) focused on a rotating mirror. The beam bounces off the mirror onto a wall or a projector screen. The laser is pulsed by an audio frequency. Just as you would use an oscilloscope to read X & Y axes, my trick casts a circular beam on a wall (for instance). The difference in my toy is that the beam does not deflect in a wave form but rather pulses (sort of the Z axes) so that you produce dots and dashes on the wall. Depending on the speed of the mirror and the particular frequency you may see dots moving left or right. You may see dots and dashes moving in different directions at the same time. It all depends on the musical composition. Simple music produces the greatest results - I know, back in the 70's I took an LED mounted on a Wayne Newton album I got for my birthday (IN THE SEVENTIES!); mounted it on a motor shaft and made some home made brushes. Varying the speed of the motor you could make the LED appear to be on at all times leaving a solid (looking) ring of light. When strobed from the audio output (speaker output) the LED would flash every time the wave form slewed positive and would extinguish when negative.

    Back then the effect was enhanced with a little mind altering smoke. I've been messing with this for a little while, mostly in my mind, but the other day I managed to take a 386 and get it to trigger a laser against the motor mounted mirror. Looked real cool - for a short while. The 386 didn't like how I had it hooked up so I'm now working out a different amp arrangement.

    One issue we had was that you had to dedicate a speaker output to lighting the LED. I want to amplify an audio pre-amp output so I don't have to mess with the volume to get the laser to fire. In fact, using the pre-amp output means I can set it and forget it. I can turn the music up or down without affecting the apparent light produced. AND I didn't have to sacrifice a speaker output. The Beatles typically had most of their instrumental arrangement on one side of the stereo track while the other track contained mostly vocals. Rather unspectacular to hear someone dramatically reading the lines while elsewhere the music (that would be playing on the left channel - or right) went mostly unheard.

    The reason for using a dual op-amp - left and right channels. Two beams doing their own thing. Perhaps I could just use a single op-amp and set it up as a summing amp - or a mixing amp and just use the single beam, but I like to go the extra mile and make something better than just good enough.

    In the days (weeks - maybe months) to come I'll keep you apprised of my progress. Much of what has to be done is fabrication of a project box for this little projector. I'll post a video when I get that far. One more thing to be accomplished is a PWM motor speed control. Easy I know, but without planning failure is imminent.
     
    Last edited: Apr 22, 2016
  9. Tonyr1084

    Thread Starter Active Member

    Sep 24, 2015
    531
    86
    The transformer - one I have laying around in a box of spare txfmrs. The posted readings are what was measured. I'm looking to use that as a power source as opposed to a 12 v gel cell battery or a 13.8 v power supply. The LM358N - seems it's not fussy about the supply, so no real need for regulation.

    My shop is still young and awaiting many projects to increase my capabilities.
     
  10. OBW0549

    Well-Known Member

    Mar 2, 2015
    1,301
    880
    Given the application, I'd agree. If you were making some sort of precision signal-processing thingie, or something where you wanted absolutely no detectable feedthru from the supplies, my comment would have been pertinent; but for what you're making, unregulated supplies should be fine.
     
  11. dannyf

    Well-Known Member

    Sep 13, 2015
    1,771
    359
    For that kind of application (ac amplification), pretty much any thing will do. I would try a bjt or even a tl431 as an amplifier.
     
  12. cmartinez

    AAC Fanatic!

    Jan 17, 2007
    3,555
    2,517
    Sounds like a cool project. Maybe when you finish it you may like to post it in the finished projects forum.
     
  13. Tonyr1084

    Thread Starter Active Member

    Sep 24, 2015
    531
    86
    I thought about a BJT but because my pre-amp voltages are lower than I know off hand what they may be - I've chosen the op-amp. Now: I'm not married to that idea. You mentioned a TI431. What's that please?
     
  14. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,516
    1,246
    Sounds like an interesting project, and the LM358 should be perfect for it. And, there is nothing in your description that demands either single or power supplies.

    Calling pin 4 GND is a stunning error. But in Fairchild's defense, the Nation Semiconductor datasheet has the same thing. Still, Fairchild basically invented the labels Vcc and Vee, and should know better.

    And, the Digi-Key Vcc description mentioned above is correct. The fact that DK lists both uni- and bipolar power supply voltages is important, because they do that only for parts where the manufacturer's datasheet is explicitly clear about the capability. Once you get used to how it "thinks", the DK search engine is a great way to search hundreds of manufacturer's catalogs in just minutes when fishing around for a part.

    ak
     
    Last edited: Apr 23, 2016
  15. cmartinez

    AAC Fanatic!

    Jan 17, 2007
    3,555
    2,517
    and for a noobie like me... being misguided by that label can negatively affect the learning process quite considerably...
     
  16. #12

    Expert

    Nov 30, 2010
    16,257
    6,757
    That seems the best description on the page. Divorce your mind from the idea that our planet is ground. The op-amp only cares whether it's inputs are between the supply voltages, and sometimes even that is not the limit because some inputs will work when their voltage is more negative than the lowest power supply voltage.

    Look at this power supply where the op-amp might be nearly a thousand volts more positive than Earth ground and you will see one method where the usual beliefs about "ground" are irrelevant, but the circuit works. Learning how to, "float" an op-amp opens up new worlds to explore.
     
    Last edited: Apr 23, 2016
    OBW0549 and cmartinez like this.
  17. OBW0549

    Well-Known Member

    Mar 2, 2015
    1,301
    880
    And there are some opamps that will work with their inputs above the positive supply, as well. Linear Tech's LT1490A is one such.
     
  18. dannyf

    Well-Known Member

    Sep 13, 2015
    1,771
    359
    TL431. You can think of it as a high gain npn with Vbe of 2.5v.
     
  19. Tonyr1084

    Thread Starter Active Member

    Sep 24, 2015
    531
    86
    Thanks Dan. But since I am unsure, let me ask this: Is the Vbe 2.5v the voltage to send the transistor "Full ON?" Or is that the bias value? I'm OK with gates and tinker with transistors from time to time, I have a basic understanding of the mechanics but often run into problems when it comes to biasing them. And is this TL431 a BJT or a FET? I've not been able to find any information on just the TL431 nomenclature.
     
  20. Tonyr1084

    Thread Starter Active Member

    Sep 24, 2015
    531
    86
    Dan: Just found some info on the TL431. I don't understand why you recommend using that. Can you explain your application of this part to what I might be after? If need be I can post some sort of schematic on where I'm going with my project. I just don't see where I'd use it.
     
Loading...