What is the lowest reference voltage of LM311 N?
I cannot get it working for reference voltages less than about 100 mV (pins 5-6 are shorted, and I use a decoupling capacitor several tens pF). At about 80 mV ref. voltage at the inverting input, the output flips to Vcc...any advice? I need to let pulses with amplitudes greater than 50 mV.

 

The allowed CM input range is -




So it does not work close to the rails. There are alternative comparators that
do. just search at vendor sites for RRIO or rail to rail.

Regards, Dana.

 

do you mean something like
LMV7219

 

hi dud,
Your diagram shows a circuit input to the LM311 which is not close to the power rails.??
E

 

hi dud,
Your diagram shows a circuit input to the LM311 which is not close to the power rails.??
E

Depends on the power supply, right? R2 can be adjusted down until the inverting input reaches ground. If the op amp is powered from single supply, then this will require a rail to rail amp. If it's got a bipolar supply, then any amp should have no problem handling small voltages.

 

Hi,

I am using R2 to change the reference voltage and therewith the discrimination level (I want to cut everything with amplitudes lower than 50 mV).

Unfortunately below about 100 mV at the inverting input, the output shifts to Vcc.

Maybe a workaround is to amplify the input signal by3 then I can use 150 mV as a threshold? but then I get all the issues with temperature stability and etc...

 

hi,
OK,
A point to be aware of is that most low cost pots have an 'end' resistance, they are not 0 Ohms when full turned to one end.
Use a voltmeter to check the actual minimum voltage at the Inv input of the LM311

What is the source voltage input at the 1meg Ohm.?
E

 

What is the source voltage input at the 1meg Ohm.?
E

the multi meter reads something like 2.4 mV.

Depends on the power supply, right? R2 can be adjusted down until the inverting input reaches ground. If the op amp is powered from single supply, then this will require a rail to rail amp. If it's got a bipolar supply, then any amp should have no problem handling small voltages.

I am suing a single power supply to simplify the things. I guess dual power supply might be better than rail-to-rail or pre-amplification?

 

hi,
If you can use a dual power supply, that should solve the problem.
Say +/-5V.
E

 

If all you need the negative rail for is the op amp, you won't need much current and can use something like the TC7660 (and the caps it requires) to generate your negative supply. It might be cheaper and/or more compact than a full, dedicated bipolar supply. Naturally, it just depends on what else you've got going on in the circuit.

There are also chips dedicated specifically to making negative rails just for op amps, but in very specific voltage range limitations. I don't think this is what you need at the moment, but they're interesting to know about. The LM7705 is one such example. It generates -0.232V from a positive supply up to 5.25V. The idea is that some op amps can handle a total power supply of only 5.5V, so your negative rail can only go a little negative. It's just enough to get rail to rail amps, which usually can't actually reach all the way to zero, the rest of the way down.

Anyway, if you don't otherwise need a bipolar supply, I'd check out the TC7660.

 

Probably I'll do it with a bipolar Power Supply: I'm still prototyping ... it worked with the bipolar supply following the circuit here:

I had to remove the coupling capacitor at the entrance. I will experiment a bit

 

The LM311 has a nominal input bias current of 100nA which will generate an offset voltage of 100mV through the 1 megohm resistor at the plus (+) input.

 

The LM311 has a nominal input bias current of 100nA which will generate an offset voltage of 100mV through the 1 megohm resistor at the plus (+) input.

yes, I am measuring about -3 mV, so there is a current sink?

 

The LM311 has a nominal input bias current of 100nA which will generate an offset voltage of 100mV through the 1 megohm resistor at the plus (+) input.

Maybe I misinterpreted the schematic, but l thought there was an external signal being applied to that input at the far left of the schematic. In that case, wouldn't the signal source impedance be the deciding factor in how much the bias current impacts offset voltages? The 1Meg only creates 100mV offset if there's no other input, right? (I promise, I'm not really arguing here, but trying to understand. I still have to think pretty hard to make sure I understand things when considering factors like input bias current. I get turned around quite easily with the nuances of op amps.)

 

yes, the signal comes from a pre-amplifier with an output impedance 93 Ohms.

 

hi,
I read your opening post the same way as described in post #14
What is the purpose of the project.?
E

 

to discriminate low level pulses (less than a threshold - or reference voltage) - low level discriminator (LLD).

Typical input pulses are 2 - 170 mV.

 

hi,
If you want such low level thresholds you must use a stable 5V ref at the top end of the divider chain, also some power rail decoupling for starters.
E

 

yes, the signal comes from a pre-amplifier with an output impedance 93 Ohms.

But did you have that connected when you were adjusting the pot and getting the offset in post #1?

 

Most comparators will not operate correctly with either input close to the rails. In my case it was the LM393/LM339. You need to read the datasheet VERY VERY carefully. The LM311 has a tendency to have stability problems so watch your decoupling and output impedance carefully. The best way to build an oscillator is to build a high gain amplifier. A comparator certainly qualifies as a high gain amplifier.