Persistent 1.9 V on buffer opamp output when Vin is 0 V

[Tim4444]

Dear everyone,

I am new to opamps and am playing around with some simple designs to try and understand them. Please could someone help me with something I am stuck with.

I have designed a very simple buffer opamp (see picture attached, A). In this design, when I close S1, I want the opamp output to equal the input Vin (two AA batteries (3 V)). The problem I am having is I get a large ouput offset voltage of 1.9 V. When S1 is open, I assumed that there should be 0 V at the non-inverting input, therefore should be 0 V at the opamp's output. When I measure the voltage on the output with S1 open, what I get is a stable 1.9 volts. This 1.9 V persists even when i change the voltage at the source from 3 - 10 V (any lower and the 1.9 V slower decreases). When S1 is closed, this voltage jumps to 3V, as I would expect considering Vin is providing 3 V. I am confused as to where this 1.9V comes from. I know opamps have certain specifications regarding their rail-to-rail abilities, and I know different opamps have specific offset voltages, but i thought this offset voltage was in the mV (or microVolt) range. I am using a LM741CN. Is it simply just that these particular opamps do no reach 0 V that well on their output (I have not seen anything on their datasheet to suggest this)?

I tried removing the battery Vin (3 V) - this circuit is pictured in panel B - but the 1.9 V persisted.

Could someone please help me with where this 1.9 V comes from, and maybe a simple solution to eliminate it (I want a starting voltage, when S1 is open) of 0 V (or as close as i can get, +- 200 mV is okay) . If a new opamp type is required (I would prefer this over adding a lot  more complexity to the circuit), does anyone know of a good one for creating buffer opamps, and is also cheap (under £2)?

Tim

p.s. i recently posted a post about creating a ramp voltage with a buffer opamp. This question is quickly mentioned in that post, so apologies for any repetitions, but I was hoping for a more detailed explanation into this specific issue.

[JPortici]

this type of opamp can't pull the output to the negative supply. one that can is called a "rail to rail output" type

then many of the RRO opamps can only pull the output down to the negative supply, but not to the positive (this is written in the datasheet, for example: "Vo goes from Vee to Vdd-1.5V")

[Ian.M]

Input bias current.   A floating input will NOT be 0V with respect to Gnd, but tends to drift to a specific device and temperature dependent voltage within the supply rails.  Try adding 1Meg from In+ to Gnd.

[Paul Rose]

Both replies are right.

A cheap/common op-amp that includes the negative rail ( handy for single supply operation ):  lm358 or lm324

Also include the "pull down" high value resistor for input current: something like 470k to 1M

lm741 will only go within 1-2 volts of the supply rails, depending on supply voltage. 

With your 741, you could use a another separate 3 volt negative supply ( negative to op-amp negative supply, positive to your "ground" ), your 6 volt supply still connected from ground to the positive supply pin of the 741, and connect the pulldown to the ground point ( not negative supply ). 

[Ian.M]

A LM741 is only specified to operate down to +/-5V supplies, 10V total. (see fig.14 in Fairchild datasheet), and it has piss-poor performance at such a low voltage.

Paul's suggestion to add a -3V negative supply will help, but you'll only have a total of 9V across the OPAMP supply terminals, which is still out of spec, so either increase the positive supply to 9V, or use a -6V negative supply.

Generally when working with 'classic' OPAMPs, a supply of +/-12V or +/-15V is used.  +/-9V from two PP3 batteries with a DPST switch to interrupt both the +9v and -9V rails simultaneously is convenient for breadboarding if you don't have a suitable bench supply.

Caution: Many modern low voltage single supply OPAMPs wont tolerate the typical voltages used by 'classic' OPAMPs - *ALWAYS* check the datasheets and supply voltages before substitution.

[Tim4444]

Thanks for all the great replies, I understand a lot more what is going on now.

Essentially, a lot of op-amps will not be able to get to the negative source rail (V-), or, for example, if this is ground, will only be able to get 1-2 volts within 0 volts. I can easily find the maximum output voltage in data sheets, but I am struggling to find any specifications about how close Vout can get to V-. In the datasheet for the lm324 for example, I could only find a small sentence at the very start (see picture attached, red box), and I am not 100% sure if this is what I am looking for. Is there a general term/symbol for this lower Vout level that is generic across opamp datasheets?

I tried adding a pull-down resistor (tried 10 K and 10 M) at Vin and it made no difference.

Lastly, I will try to add a negative voltage supply to V-. I have not got a power source that goes to negative so I will try and use two 9 V batteries in the configuration in this link: http://www.instructables.com/id/Dual-Voltage-Supply-9V-batteries/     .....     I will let you know how it goes.

Lastly, I think I will also just try another opamp that has good rail-to-rail specifications.

Thanks for the help,

Tim

[Audioguru]

The 741 opamp is 48 years old, long before rail-to-rail.
A lousy old LM324 quad opamp (or its dual opamp sister the LM358) is also not rail-to-rail. Its output goes down almost to ground but goes up to the supply minus about 1.2V when its output current is only 1mA or less and its inputs work at ground but not above the supply minus 2V. They are noisy, have crossover distortion and have trouble at high levels above 2kHz.   

[Ian.M]

Alkaline PP3 batteries aren't cheap,  unless you are lucky enough to find them at the dollar store, so a switch to disconnect the batteries easily is strongly recommended so you don't waste current while you are scratching your head thinking.   Its a really bad idea to disconnect one supply rail without disconnecting the other as it can cause current to flow through IC input terminals and damage the chips, so pulling or inserting one wire at a time or unclipping the batteries is not satisfactory.   The minimum requirement is a DPST switch - two independent sets of contacts in the same switch body, and of course you can also use a DPDT switch wired as a DPST one, however if you are going to be experimenting with OPAMPs + power or logic circuits that you are going to be using an extra power rail for then ideally you should get a 3PST (or 3PDT wired as 3PST) switch and bring out the spare pole to a pair of screw terminals you can wire the extra power rail through so it switches at the same time as your OPAMP supply.

You can mess about with salvaged terminal plates from PP3 batteries as per that instructable, or simply get standard PP3 battery clips, compete with red and black wires.  The connections inside the clips are often rather fragile and its well worth squirting a drop of hotglue, impact adhesive or electrical grade silicone caulk into the vinyl cover where the wires enter to provide strain relief.   

Join the red and black wires from separate clips to get your 0V connection.  Splice it to a breadboard jumper.   Solder the remaining red and black wires to separate poles of the switch and  solder two more jumpers to the other contact of each pole to provide your switched +9V and -9v rails.   You may decide to use a 3A 'scotchblock' screw terminal block with short wires to it from the switch instead of breadboard jumper wires so you can easily change the jumpers to the circuit.  If you have a third pole on the switch wire it to two spare ways on the terminal block for future use.

ALWAYS use a consistent colour code for the circuit wiring as mixing up your +9V, and -9V rails will usually blow any chips on your board!   I favour either: red for positive, black for negative and green for 0V/Gnd (although US NEC for three wire DC supplies says white for 0V) or: red for positive, black for 0V and blue for negative.  DON'T mix and match or use those colours for any other breadboard wiring except for the supply and ground rails.  While I'm on the subject of wiring colours: breadboard jumpers colour coded for length are an abomination that can only lead to mistakes and circuit damage.  Either cut your own from suitable solid core wire, or buy 'rainbow' jumper assortments of various lengths so you can colour code your breadboard wiring by function or type of signal. 

explanation of switch poles and ways

[Tim4444]

Hello,

I have manged to fix the problem using two batteries in series to obtain the V+ and the V- for the opamp.

Firstly, I tried using two 9 V batteries in series (picture, ' 2x9V') - I took the middle point between the batteries as ground, then the positive of the first battery (left, picture) for V+ (9 V) and the negative of the second battery for V- (-9 V). This now though seemed to give me a persistant -7 V on the output of my opamp-buffer.

The next thing I tried was to swap the 2nd 9 V battery with a 1.5 V battery (picture, 'V- at -1.5V'), again taking the point the two batteries as ground. Now I get a persistent 400 mV on the output, not the original 1.9 V I had when I had V- tied to ground. This kind of makes sense to me because, when V- equals ground, Vout = 1.9 V, and when V- = -1.5 V, Vout = 400 mV - 1.9 V - 1.5 V = 400 mV.

Essentially, what I have learnt from this is that, if I have a unwanted voltage offset on my buffer opamp, what I can do to eliminate this (in theory) is to add the negative of this offset (in volts) to the opamp's source voltage, e.g. if i have Vout offset of 1.5 V, I can just add -1.5 V to V- to get a Vout of 0 V, or, if I have an Vout offset of -1.5 V, just add 1.5 V to V+.

Is this correct?

Thanks,

Tim

[Zero999]

Hello,

I have manged to fix the problem using two batteries in series to obtain the V+ and the V- for the opamp.

Firstly, I tried using two 9 V batteries in series (picture, ' 2x9V') - I took the middle point between the batteries as ground, then the positive of the first battery (left, picture) for V+ (9 V) and the negative of the second battery for V- (-9 V). This now though seemed to give me a persistant -7 V on the output of my opamp-buffer.

The next thing I tried was to swap the 2nd 9 V battery with a 1.5 V battery (picture, 'V- at -1.5V'), again taking the point the two batteries as ground. Now I get a persistent 400 mV on the output, not the original 1.9 V I had when I had V- tied to ground. This kind of makes sense to me because, when V- equals ground, Vout = 1.9 V, and when V- = -1.5 V, Vout = 400 mV - 1.9 V - 1.5 V = 400 mV.

Essentially, what I have learnt from this is that, if I have a unwanted voltage offset on my buffer opamp, what I can do to eliminate this (in theory) is to add the negative of this offset (in volts) to the opamp's source voltage, e.g. if i have Vout offset of 1.5 V, I can just add -1.5 V to V- to get a Vout of 0 V, or, if I have an Vout offset of -1.5 V, just add 1.5 V to V+.

Is this correct?

Thanks,

Tim

The first schematic is correct. I don't understand the second one though: what are you trying to do, get +9V & -1.5V?

Ideally you should design the circuit so it works from a single power supply. How many modern electronic devices need two 9V batteries?

If the circuit amplifies AC signals, such as an audio amplifier, then this can be done by biasing the op-amp at half the power supply voltage and coupling the signal with capacitors, which pass AC, not DC.

If the circuit is DC, then you can use an op-amp which will work down to the 0V rail such as the old LM358/324, TS272, MCP602 etc. or a virtual earth circuit.