Rail - to - rail opamp

Thread Starter

biteriu

Joined May 19, 2020
12
Hello,

I've read something about a rail-to-rail opamp on the internet and I want to design one, just the soft part, and I am not sure if I get some things right. For example:
Let's assume that my Supply Voltages are Vss = -5 V and Vdd=5 V
In a rail-to-rail op amp I need my input to be anywhere in [Vss : Vdd], even... -10mV : 10mV, or -1V and 1V, up to -5V and 5V. [The other input can be the ground]
And the output should always be almost Vss and Vdd.

I am sorry if im not posting this in the right section, I just made the account!
 

dl324

Joined Mar 30, 2015
10,758
Welcome to AAC!
For example:
Let's assume that my Supply Voltages are Vss = -5 V and Vdd=5 V
In a rail-to-rail op amp I need my input to be anywhere in [Vss : Vdd], even... -10mV : 10mV, or -1V and 1V, up to -5V and 5V. [The other input can be the ground]
And the output should always be almost Vss and Vdd.
When describing circuits, you should use schematics to convey thoughts. They're easier to interpret than words.

You mention rail-to-rail, but you neglect to mention whether you're referring to inputs or outputs. You can have one without the other, or have both. There are commonly used techniques to accommodate inputs that don't include the rails, but not for outputs.
 

Thread Starter

biteriu

Joined May 19, 2020
12
You mention rail-to-rail, but you neglect to mention whether you're referring to inputs or outputs.
I'm reffering to a fully rail-to-rail, inputs and outputs. Thats the problem, I dont really understand the concept of the "input rail to rail".. Don't I just have the same signal at the input and output with this one? Sorry if I got something misunderstood there...

That is a comparator, not an opamp.
It's a comparator because I don't have a feedback loop?
 

DickCappels

Joined Aug 21, 2008
6,388
Welcome to allaboutcircuits.com. My understanding is that the term "rail-to-rail" relates to the output swing and though having a common mode input range equal to or exceeding the output range, not being able to work over that input range would not prevent the opamp from being classified as rail-to-rail.
 

dl324

Joined Mar 30, 2015
10,758
I'm reffering to a fully rail-to-rail, inputs and outputs. Thats the problem, I dont really understand the concept of the "input rail to rail".. Don't I just have the same signal at the input and output with this one?
Until recently, comparators didn't have rail-to-rail outputs and the outputs were always open collector. By using voltage dividers on the inputs, you could accommodate the comparator not having a common mode input range that included the rails. The output was always close enough to the negative rail (which was usually ground), and the output could be near the positive rail by choosing an appropriate pull-up resistor.

In cases where you needed the output to source more current and stay close to the positive rail, you'd buffer the output.

The most common technique for achieving rail-to-rail inputs and/or outputs is to employ a charge pump so the opamp, or comparator, increases it operating supplies.
 

Thread Starter

biteriu

Joined May 19, 2020
12
Until recently, comparators didn't have rail-to-rail outputs and the outputs were always open collector. By using voltage dividers on the inputs, you could accommodate the comparator not having a common mode input range that included the rails. The output was always close enough to the negative rail (which was usually ground), and the output could be near the positive rail by choosing an appropriate pull-up resistor.

In cases where you needed the output to source more current and stay close to the positive rail, you'd buffer the output.

The most common technique for achieving rail-to-rail inputs and/or outputs is to employ a charge pump so the opamp, or comparator, increases it operating supplies.
Thank you very much!
 

OBW0549

Joined Mar 2, 2015
3,408
My understanding is that the term "rail-to-rail" relates to the output swing and though having a common mode input range equal to or exceeding the output range, not being able to work over that input range would not prevent the opamp from being classified as rail-to-rail.
I've never seen an op amp described or classified as simply "rail-to-rail." Never.

The term is meaningless unless qualified as referring to the inputs, the output or both. Hence an LMC662 is described as having a rail-to-rail output, while the LMC6482 is described as a "rail-to-rail input and output" op amp, as is the LM6132.
 

Thread Starter

biteriu

Joined May 19, 2020
12
I've never seen an op amp described or classified as simply "rail-to-rail." Never.

The term is meaningless unless qualified as referring to the inputs, the output or both. Hence an LMC662 is described as having a rail-to-rail output, while the LMC6482 is described as a "rail-to-rail input and output" op amp, as is the LM6132.
I am sorry, I thought by " simply rail-to-rail" it means that it is referred to input and output.
 

Thread Starter

biteriu

Joined May 19, 2020
12
I have another question, can you tell me what simulation should I use exactly, DC to see the operating point of my transistors and Transient to see the input and output in time? Is there something else I am missing?
 

dl324

Joined Mar 30, 2015
10,758
I have another question, can you tell me what simulation should I use exactly, DC to see the operating point of my transistors and Transient to see the input and output in time? Is there something else I am missing?
You could start by showing us a schematic for your circuit.

What do you mean by "Transient to see the input and output in time"? Do you want transient response, frequency response, or both?
 

OBW0549

Joined Mar 2, 2015
3,408
The .OP analysis in SPICE will the operating point. The .AC analysis also begins with a .OP analysis; you should be able to find the results for it somewhere in the SPICE output file.
 

Thread Starter

biteriu

Joined May 19, 2020
12
1589908281679.png

I don't think it's working, I still have to choose the W/L for the transistors but:

The input stage is the two differential pair, one NMOS and the other PMOS with two mirror currents (one for each).

The output stage is a class B amplifier.

The intermediate stage I thinks it's the problem here ... the transistors: from M5 to M12.
 

Thread Starter

biteriu

Joined May 19, 2020
12
I hope this is all.

.model AO6408 VDMOS(Rg=3 Rd=4.8m Rs=3.6m Vto=1 Kp=90 Cgdmax=.7n Cgdmin=.25n Cgs=1n Cjo=.36n Is=36p Rb=6m ksubthres=.1 mfg=Alpha_&_Omega Vds=20 Ron=12m Qg=18n)

.model SI3445DV VDMOS(pchan Rg=3 Rd=15m Rs=6m Vto=-.85 Kp=18 Cgdmax=1n Cgdmin=.44n cgs=1.2n Cjo=.1n Is=.045nA Rb=.026 N=1.127 ksubthres=.1 mfg=Siliconix Vds=-8 Ron=34m Qg=15n)
 

Attachments

Top