Hello Everyone,

I have been trying to get a voltage follower to work for my application with no success.
I am using voltage followers in my design in several places and have no problems, but for this particular configuration my input voltage is very small ( 0.17 mV) and I am measuring 60 to 80mV output voltage at the output.
I read several papers and tried different approaches ( on this website and other ) but I can't seem to find a way around this issue.
My input voltage comes from a thermistor linearizing resistor network and will range from 0.17 ( T=25C) to 0.28mV ( T=50C)
Here is my initial setup :

I started with the TLV2370 .which has an input offset voltage in the order to ~2mV. I though that was my problem so I switched to an LTC2050 ( have it handy) with input offset voltage is in the uV range.
I still measure about the same output voltage of around 60 to 80mV.
I tried this setup ( Although some member of this forum disapproved in some posts ) and it didn't help. I was able to increase the offset but not decrease it.

I also tried matching the feedback resistor with an input resistor (here) ( 1K trying to limit the input bias current ) but still measure the same output voltage of 60mV when the the input is 0.16 mV.


I do not want to use an opamp with a nulling input and endup with a pot on the board.

I think I am missing something and can;t figure it out by myself. I hope someone smarter can help me out here.

Thanks
AJ

 

What are you using for +V and -V?

You could try adding a 1K resistor in series with the non-inverting input. Then change R7 to 1K also.

The source impedance from your resistive network (top of R8) is very low (<10 ohms), do you really need a buffer on that line?

 

If TLV23728 was two words it would be a googlewhack!

 

What are you using for +V and -V?

You could try adding a 1K resistor in series with the non-inverting input. Then change R7 to 1K also.

The source impedance from your resistive network (top of R8) is very low (<10 ohms), do you really need a buffer on that line?

Thanks for your reply
I am using 3.3V for V+ and an AGND for V-
I tried what you suggested in an effort to limit the biased in the nin inverted input and changed the feedback resistor to 1k as well but I still got the same output voltage level.

Regarding your question on the need for a buffer: my intension in to feed this signal as a bias voltage to a non inverting opamp that amplifies an IR detector signal ( thermopile) in order to perform some sort of ambient temperature compensation. This IR sensor is mounted on a surface that heats up and its temperature varie from 25 to 50C and comes with an embedded thermistor.
I was trying to avoid connecting the resistor network in series with R1 (replaces VREF in below schematic) because I think that it will also make my gain change with temperature. ( becauss of the thermistor)


I want the offset to change with Temp. and gain to remain constant.

 

If TLV23728 was two words it would be a googlewhack!

Hello Albert,

The opamp in the schematic is TLV2372 . Eagle called it TLV23728MSOP . 8MSOP is the name of the Variant as this opamp comes in multiple packages.

 

I might be misunderstanding these specs (happens more often than I'd like to admit,) but it looks like the lowest output it's capable of can easily be 100-150mV above ground in the 2.7-5V Vdd range, and is "typically" 50-100mV in that range.

 

Probably going to need at least a couple of -volts on that -V pin. You are trying to operate with input and output right at the rail.

 

Probably going to need at least a couple of -volts on that -V pin. You are trying to operate with input and output right at the rail.

You are right.
I used a lab power supply and fed +/- 3.3v to that opamp and the follower is working now.
I don;t have anegative rail on my board at the moment. Although it is not impossible to add one, I still have some space limitation on it.
Can you please suggest a way of providing a -v rail to the opamp.
It looks to me like Inverting DC/DC Converters are switching regulators and will require an inductance , a relatively large cap and a diode at the output. I looked at the LT1617 datasheet and few similar products.
Your suggestions are greatly appreciated

Thanks

 

This chip is designed specifically for the purpose.

 

This chip is designed specifically for the purpose.

This is extremely helpful, thank you very much four your help. This solves my problem

 

This is extremely helpful, thank you very much four your help. This solves my problem

The 7705 only outputs -0.23 volts. May be enough, may not. Test before committing to it.