Powering op amps: dual-supply versus single-supply

Thread Starter

xox

Joined Sep 8, 2017
838
Learning about op amps and their various power options:
  1. With dual supply, are we talking Vcc+/Vcc- being AC signals, or just constant DC voltages (positive/negative)?
  2. If the latter, then how is the negative voltage typically generated? Does it require an AC signal, or can it be done using a positive DC voltage as well?
  3. Sometimes I see this variation where either Vcc+ and/or Vcc- are run through capacitors or other components. Are there any major advantages/disadvantages of incorporating that sort of thing into a circuit design?
  4. Are there any "universal" methods for converting a dual-supply circuit to a single-supply design, and vice-versa?
 

Marley

Joined Apr 4, 2016
502
Op-amps are used with dual supplies (positive and negative) because it allows the input and output signals to be positive and negative and be truly 0V. Generating the negative supply is quite easy: DC - DC converters, switched capacitor converters or for mains-powered equipment a simple centre-tapped transformer and rectifier.

Don't understand what you mean by "run through capacitors". There will usually be bypass capacitors from the power supply rails to ground. This is to bypass noise and ac signals to ground. Usually very important to have these bypass capacitors for stable operation of the circuit.

With single-supply designs there can be a pseudo-ground reference, often at half the supply voltage. Signals can swing above and below this voltage instead of 0V. For applications where the exact DC voltage level is important (test and measurement) the exact voltage of this pseudo-ground is very important and must be stable. Often easier to generate a negative supply rail (which does not have to be stable).
 

Thread Starter

xox

Joined Sep 8, 2017
838
Op-amps are used with dual supplies (positive and negative) because it allows the input and output signals to be positive and negative and be truly 0V. Generating the negative supply is quite easy: DC - DC converters, switched capacitor converters or for mains-powered equipment a simple centre-tapped transformer and rectifier.
Okay, so typically a positive and negative DC voltage. But at least in some situations an AC signal can be used to supply an op amp (with Vcc- being an inverted version of Vcc+), or is that just never done?

There will usually be bypass capacitors from the power supply rails to ground. This is to bypass noise and ac signals to ground. Usually very important to have these bypass capacitors for stable operation of the circuit.
Good to know, thanks!

With single-supply designs there can be a pseudo-ground reference, often at half the supply voltage. Signals can swing above and below this voltage instead of 0V.
So a simple voltage divider would be sufficient?
 

Thread Starter

xox

Joined Sep 8, 2017
838
When I first started tinkering with electronics (ah, such fond memories of just a few months ago), I had this naive idea that a small amount of voltage couldn't possibly harm a component like an LED - "if it can't shock you then what's the worst that could happen if I did this" sort of thing. I had just gotten a brand new package of LED's and just wanted to see all the different colors, honestly. So I took my new power supply and cranked it up to "only" 9V or something (didn't use the current-limiting feature of course)...and proceeded to burn out one LED after another, the whole time thinking I'd been ripped off by a supplier who'd sent me a bunch of duds. After trying maybe one of every color it finally occurred to me that it might be a good idea to do a little research. It was a valuable lesson, actually. I realized that one really can't assume anything when dealing electrical circuits. Nowadays, I try to use a "check list" type of approach when breadboarding, just to make sure I don't fry a component (or myself) in the process.
 

OBW0549

Joined Mar 2, 2015
3,566
I had just gotten a brand new package of LED's and just wanted to see all the different colors, honestly. So I took my new power supply and cranked it up to "only" 9V or something (didn't use the current-limiting feature of course)...and proceeded to burn out one LED after another, the whole time thinking I'd been ripped off by a supplier who'd sent me a bunch of duds.
It's pretty rare anymore to get a defective electronic component (other than from a shady supplier)-- and even more rare to get a whole batch that are bad.

If something malfunctions, you can reckon there's about a 99.9% chance it's because you did something wrong-- in which case the thing to do is stop immediately and figure out what it is.
 

Thread Starter

xox

Joined Sep 8, 2017
838
It's pretty rare anymore to get a defective electronic component (other than from a shady supplier)-- and even more rare to get a whole batch that are bad.

If something malfunctions, you can reckon there's about a 99.9% chance it's because you did something wrong-- in which case the thing to do is stop immediately and figure out what it is.
I've also noticed that it isn't entirely uncommon for components to be "out of spec", so it pays to do some tests and measurements before plugging stuff into a circuit too. I've run into things like potentiometers that literally dropped out to zero ohms randomly along the dial, and mislabeled capacitors that ended up causing me several hours of frustration before I thought to verify their values. You don't necessarily have to check each and every component, just a few out of each package at least. And you know, I'm actually starting to enjoy that little task. Guess it makes me feel like I have a better grasp on what to expect (not to mention useful for tweaking simulations to pinpoint possible problem areas).
 

MrChips

Joined Oct 2, 2009
30,711
Potentiometers do not randomly drop out to zero ohms. They do sometimes lose contact and give infinite ohms instead.

This is why when you are using a pot as a simple variable resistor, you connect the center wiper to the unused terminal. Instead of giving infinite resistance the pot will be limited to the nominal value of the pot.

When using a pot as a voltage divider, you can mitigate the random loss of contact by installing fixed resistors from the wiper contact to the outer arms, depending on what you desire as your fail-safe result.
 

Thread Starter

xox

Joined Sep 8, 2017
838
Potentiometers do not randomly drop out to zero ohms. They do sometimes lose contact and give infinite ohms instead.

This is why when you are using a pot as a simple variable resistor, you connect the center wiper to the unused terminal. Instead of giving infinite resistance the pot will be limited to the nominal value of the pot.

When using a pot as a voltage divider, you can mitigate the random loss of contact by installing fixed resistors from the wiper contact to the outer arms, depending on what you desire as your fail-safe result.
Thanks, that's pretty useful stuff to know, actually!

But as far as those went, I'm fairly certain it was a zero reading. The meter I used flashes a "-" for infinite and "0" for, well, zero ohms. I'll recheck though if I can find them just to be sure.

@MrChips: Found one of these pots and was able to confirm that it does indeed drop to zero moving along the dial. I wanted to post a video, but can't seem to upload it here, sorry...
 
Last edited:

OBW0549

Joined Mar 2, 2015
3,566
I've also noticed that it isn't entirely uncommon for components to be "out of spec", so it pays to do some tests and measurements before plugging stuff into a circuit too. I've run into things like potentiometers that literally dropped out to zero ohms randomly along the dial, and mislabeled capacitors that ended up causing me several hours of frustration before I thought to verify their values.
As I said above:
It's pretty rare anymore to get a defective electronic component (other than from a shady supplier)-- and even more rare to get a whole batch that are bad.
When I need to make sure I'm getting reliable, in-spec parts from a reputable manufacturer (which is most of the time), I always buy from a trustworthy distributor (Digi-Key, Mouser, etc.) who will certify what they sell. The only time I'll buy from any of the low-cost overseas suppliers, such as uxcell via Amazon, is if I'm just playing around and literally don't care whether the parts meet specifications-- because all too often, they don't.

You get what you pay for.
 

MrChips

Joined Oct 2, 2009
30,711
Thanks, that's pretty useful stuff to know, actually!

But as far as those went, I'm fairly certain it was a zero reading. The meter I used flashes a "-" for infinite and "0" for, well, zero ohms. I'll recheck though if I can find them just to be sure.

@MrChips: Found one of these pots and was able to confirm that it does indeed drop to zero moving along the dial. I wanted to post a video, but can't seem to upload it here, sorry...
Dropping to zero ohms and dropping to zero output are two different phenomena.
Check to make sure it is one and not the other
 

Thread Starter

xox

Joined Sep 8, 2017
838
Dropping to zero ohms and dropping to zero output are two different phenomena.
Check to make sure it is one and not the other
Okay, so just to double-check I've placed the pot in series with an LED and its current-limiting resistor. As I turn the dial very slowly the LED gradually fades but then at several points along the way it suddenly brightens (if it had been a pocket of infinite resistance, the LED would have gone out).
 

Audioguru

Joined Dec 20, 2007
11,248
A linear opamp circuit usually needs to have its (+) input biased at half the supply voltage so the input and output can swing equally up and down. If the supply has + and - then half the supply is 0V. if the supply is only + then do it like this:
 

Attachments

Thread Starter

xox

Joined Sep 8, 2017
838
A linear opamp circuit usually needs to have its (+) input biased at half the supply voltage so the input and output can swing equally up and down. If the supply has + and - then half the supply is 0V. if the supply is only + then do it like this:
Wow, thank you, I would have never figured that out myself!

But now just to be clear, this applies only when used as an amplifier? I mean for something like a simple comparator circuit a straight single supply should work fine, correct?
 

OBW0549

Joined Mar 2, 2015
3,566
But now just to be clear, this applies only when used as an amplifier? I mean for something like a simple comparator circuit a straight single supply should work fine, correct?
No. What applies to op amps applies equally to comparators: whether single supply or dual supply, both inputs must remain within the specified Input Common Mode Voltage range at all times for the chip to operate properly.

There's some good reading on the subject here, here and here.
 

Thread Starter

xox

Joined Sep 8, 2017
838
No. What applies to op amps applies equally to comparators: whether single supply or dual supply, both inputs must remain within the specified Input Common Mode Voltage range at all times for the chip to operate properly.

There's some good reading on the subject here, here and here.
Thanks for the clarification, I'll go ahead and read up on that now.
 
Top