Question on op-amp current capability for making V-Ground circuit

[BravoV]

I'm going to build a circuit that needs +5 and -5 rails, and looking to make these rails using simple popular circuit like this :



My circuit will only draw max at only 10 mA on +5 and -5 rails, and currently I have abundant of LM358 and LM324 opamps scattered around.

Attached here below are the each datasheet on their current capabilities :






Now my questions :

1. Will this opamps choices are good enough for this job ?

2. If answer of above question is yes, then it looks like LM358 is better than LM324 for this task looking at it's current capability, am I right ?

3. About the R1 resistor, does it really need it ? The reason is I saw lots of examples are using it, what happened if I don't use the R1 ? And what is the suitable value anyway if my current draws on both rails only max out at 10 mA each ?

TIA

[IanB]

I think R1 is to prevent a current surge between the op amp output and C2/C3 on start up. These capacitors will otherwise appear like a transient short circuit to the op amp.

When I built a circuit like that and simulated it with some op amps it went unstable. Some advice I have seen suggests eliminating R1, C2 and C3 entirely. Instead bypass the lower of the 100 k resistors with a small capacitor like 0.1 uF to keep noise out of the input to the op amp and let the output of the op amp be the virtual ground directly.

[grenert]

Sure you don't want to just use a TLE2426?  Only a couple of bucks and easy to implement.

If you decide to stick with the opamp:
This is a pretty undemanding task for an opamp.  I agree the LM358 looks to handle more current.
When I've done this circuit, I didn't have an R1, and no problems arose.

[BravoV]

I think R1 is to prevent a current surge between the op amp output and C2/C3 on start up. These capacitors will otherwise appear like a transient short circuit to the op amp.

When I built a circuit like that and simulated it with some op amps it went unstable. Some advice I have seen suggests eliminating R1, C2 and C3 entirely. Instead bypass the lower of the 100 k resistors with a small capacitor like 0.1 uF to keep noise out of the input to the op amp and let the output of the op amp be the virtual ground directly.

Actually the C2 and C3 are not two caps, they represent the bypass caps that are scattered at the main circuit, like for ICs by pass caps, and probably I will be using plenty of them in the circuit for every ICs.

Good point on the small caps at the lower 100K voltage divider resistor, but only at lower part is enough ?

Sure you don't want to just use a TLE2426?  Only a couple of bucks and easy to implement.

If you decide to stick with the opamp:
This is a pretty undemanding task for an opamp.  I agree the LM358 looks to handle more current.
When I've done this circuit, I didn't have an R1, and no problems arose.

I'm aware of that specialized IC made just for this purpose, its just I don't feel like to buy new one while I got tons of LM358 and LM324 in my drawer. 

Yeah, 10mA current at each rail is not that demanding, thanks, so R1 is not needed.


Another question just pop out in mind :

Say the choice is LM358 which is actually a dual op-amp in single package, will the extra op-amp helps to boost the current capability like wire both of them in parallel ?

How the circuit will look like ? Just parallel them using the same resistor divider, and connect both output on the V-Ground ? Is that it ?

[IanB]

I don't believe you could wire two op amps in parallel. I think they would tend to fight each other for control.

[ciccio]

A question: you must supply only analog devices, such as op-amps?
In this case I use the circuit in the attached picture. I've posted it in a previous topic, but don't know how to post a link to it...
Use an op-amp with some current capability, such as NE5532.
It must be a low noise type (all those used for audio are suitable).
No need for an output protection resistor, the op-amp will take care of overload.

The capacitor will give a "soft start", will filter out any supply noise and will help in keeping a low impedance source. 4.7 to 10 uF will be OK.
The diode will discharge the capacitor versus the supply source (9V) at switching off, preventing a damage to the op-amp input when switching off (without it  the input may momentarily be at an higher voltage than the supply pin). An 1N4148 will be OK.

If you need more current, use two op-amps, paralleling the + inputs, leave feedback as in the picture, do not parallel the outputs but put a small value (10 ohm) resistor in series with each of them and connect the two resistor at the other side.
Sometimes a diode from this junction to V+ will help at start-up
Best regards

[BravoV]

A question: you must supply only analog devices, such as op-amps?
In this case I use the circuit in the attached picture. I've posted it in a previous topic, but don't know how to post a link to it...
Use an op-amp with some current capability, such as NE5532.
It must be a low noise type (all those used for audio are suitable).
No need for an output protection resistor, the op-amp will take care of overload.

It will be used in a circuit that has analog and logic ICs.

I checked the NE5532 datasheet, it doesn't mention about the current sink/source capability, is it better than the LM358 for this purpose ? Maybe I missed it somewhere, cause looks like the datasheet has less detail info than LM358.

The capacitor will give a "soft start", will filter out any supply noise and will help in keeping a low impedance source. 4.7 to 10 uF will be OK.
The diode will discharge the capacitor versus the supply source (9V) at switching off, preventing a damage to the op-amp input when switching off (without it  the input may momentarily be at an higher voltage than the supply pin). An 1N4148 will be OK.

If you need more current, use two op-amps, paralleling the + inputs, leave feedback as in the picture, do not parallel the outputs but put a small value (10 ohm) resistor in series with each of them and connect the two resistor at the other side.
Sometimes a diode from this junction to V+ will help at start-up

Thanks for the circuit, so as your note on parallel op-amp, I made the diagram below, please verify.

Also you mentioned this -> "Sometimes a diode from this junction to V+ will help at start-up", I'm really curious what is this all about ? Put this diode in the circuit as well too.

[slateraptor]

I don't believe you could wire two op amps in parallel. I think they would tend to fight each other for control.

http://www.intersil.com/data/an/an1111.pdf

[BravoV]

I don't believe you could wire two op amps in parallel. I think they would tend to fight each other for control.

http://www.intersil.com/data/an/an1111.pdf

Thanks slateraptor, a valueable reference.

As stated since my circuit draws only 10 mA at each pos and neg rails, with the op-amp's offset voltage of 5 mV (max as per datasheet), and also assuming the worst case scenario my circuit will draw twice the number (eg 20 mA), looking at the below pic on the LM358 current sink/source capabilities, especially by using both op-amps in parallel, this should be more than enough right ?

Really appreciate any confirmation on this matter, btw I'm learning this as well, especially on reading opamp's datasheet. 

Also what is the optimal value for those two ballast resistors in order to have better power efficiency ? Since this circuit will be powered using a battery.



Click the this below thumbnail to see the LM358 current capabilities.

[olsenn]

10mA is nothing for any op-amp that I can think of, but fortunately for you, Texas Instruments offers free samples of the TLE2426 and next day delivery.

[BravoV]

10mA is nothing for any op-amp that I can think of, but fortunately for you, Texas Instruments offers free samples of the TLE2426 and next day delivery.

Nope, not an option cause TI wont ship any sample to my country, the only option is to purchase it online there, and the problem is they charge $70 just for s/h & handling. 

[slateraptor]

Also what is the optimal value for those two ballast resistors in order to have better power efficiency ?

Yes...

10mA is nothing for any op-amp that I can think of, but fortunately for you, Texas Instruments offers free samples of the TLE2426 and next day delivery.

Digi-Key can think of >400 general purpose op amps in stock that can't. Being aware that certain components are at an engineer's disposal is nice, but it appears that BravoV has the flexibility to explore a subtle application that many EE students don't, so I can't knock him for daring to stray from the path of least resistance.

[Bored@Work]

One question hasn't been asked. Are you sure you need a split power supply? Shifting voltage levels in the circuit, the part we didn't see, can mitigate the need to have a split power supply.

[Zero999]

I don't believe you could wire two op amps in parallel. I think they would tend to fight each other for control.

http://www.intersil.com/data/an/an1111.pdf

There's also less chance of a mismatch if both of the op-amps are on the same IC.

[ciccio]

I checked the NE5532 datasheet, it doesn't mention about the current sink/source capability, is it better than the LM358 for this purpose ? Maybe I missed it somewhere, cause looks like the datasheet has less detail info than LM358.

Yes, never found a spec for output current.
We can try to calculate it ( using a rough approximation, purists please forgive me)
This is a low noise audio op-amp, that can drive a 600 ohm load at +20 dBm with  30 V total supplies (dont' have a data sheet on hand, but I remember these values).
Using one on-line calculator http://www.daycounter.com/Calculators/Decibels-Calculator.phtml
we see that this is equivalent to 7.75 V dc or 11.0 peak ac volts.
7.75/600 = 13 mA.  (I've sourced or sinked higher currents without any problem, but the chip was hot).
I've seen somewhere an audio POWER amplifier built with LOTS on paralleled 5532s.

EDIT : an NJM4580 data sheet http://akizukidenshi.com/download/NJM4580.pdf states +/- 50 mA output current. This is a cheap op-amp, low noise, reliable.
Actually I use it in place of 5532,

Your schematic is correct.

Please consider that I used this circuitry to simulate the ground reference to the non inverting inputs of multiple op-amps,  with single power suppply.
The few logic chips are usually CMOS, powered by the positive supply with a series zener diode (that will take care of the difference : the ground pin is connected to the negative supply, the +V pin goes to the positive supply via a series zener. The logic circuits I use are usually very slow, so this trick works.

[BravoV]

One question hasn't been asked. Are you sure you need a split power supply? Shifting voltage levels in the circuit, the part we didn't see, can mitigate the need to have a split power supply.

Nothing fancy actually, this circuit will be used for powering a unique diy esr meter that is using synchronous detector method (what a mouthful, at least this is what I read  ) using 4 wires measurement method, and it is also an adapter to a DVM for it's reading, to me its sort of like Dave's ucurrent but for cap esr thingy, link here -> http://kripton2035.free.fr/analog%20esr/esr-lowohm-silic.html

I'm no EE, its just I can't see if its easy or worth the trouble to modify the circuit in order to use single supply, but I'm all ear if you have anything in your sleeve to make it happened easily.

Your schematic is correct.

Please consider that I used this circuitry to simulate the ground reference to the non inverting inputs of multiple op-amps,  with single power suppply.
The few logic chips are usually CMOS, powered by the positive supply with a series zener diode (that will take care of the difference : the ground pin is connected to the negative supply, the +V pin goes to the positive supply via a series zener. The logic circuits I use are usually very slow, so this trick works.

Thanks ciccio, you've been helping me alot, really appreciate it.

Btw, about the additional diode for preventing hiccup on the start-up as you mentioned, mind elaborate further on that part ?

[Mechatrommer]

this circuit will be used for powering a unique diy esr meter

how many amps the esr meter requires? remember that amps will go in and out of your virtual ground. i only worry impedances will screw that up, never tried though, so cant tell. the only way for sure is to make it in real life.

[BravoV]

the only way for sure is to make it in real life.

Thats the plan. 

[Mechatrommer]

the only way for sure is to make it in real life.

Thats the plan. 

so my 2cnt advice is dont waste your time too much in sim. you know, time is gold (in fact "more precious than")

[Zero999]

Go ahead and try and see if it works , if it doesn't or it blows up ,
you have many of those 

There's no risk of it blowing up, the output is normally current limited so that can't happen. If it doesn't work, there's more chance of it behaving unpredictability such as oscillating.

[codeboy2k]

Here's how I would go about it..

I would build that circuit first with a real +/- 5V supply if you can, and debug it and get it working, BEFORE trying to make it run from a single supply
with a virtual ground.

As someone else has said, the vground can bounce around all over the place, especially if something is not right with your circuit,
so you really don't want that while  you're trying to debug the rest of the circuit.

I would recommend using two power supplies that you can hook up back to back and build it up first. If you don't have two power supplies, then simply
two 9-volt batteries (or 12V batteries) with two LM7805 type linear regulators. That gives you two floating power supplies that you can wire up to get +/- 5V