Need true Rail to Rail part recommendation....

Thread Starter

Gearbreaker

Joined May 23, 2019
24
I am trying to take a DC voltage that varies slowly from 2 millivolts through 50 millivolts and convert it to .2V through 5V linearly. I have a single supply available. Easy task for an op amp right? Each op amp I try gets flaky and non linear on the lower voltage end. I have heard good things about the MAX4239 and the MCP6V01. Does anyone have any idea if either of these would work better than the MC3320 rail to rail unit that I have been trying to use? Any op amp recommendations are welcome!
 

danadak

Joined Mar 10, 2018
4,057
My experience is if you are trying to do high precision work you have
to trial RRIO parts. My rule of thumb is 12 bits over temp, V, all errors,
a challenging design.

For example, still many manufacturers not discussing the crossover distortion
inherent in their input stage design. Here is an example of where due diligence
was done -

upload_2019-6-22_8-16-12.png


Short answer is read datasheet, if you see no discussion of this move on.


Regards, Dana.
 

Thread Starter

Gearbreaker

Joined May 23, 2019
24
Dana,

Thank you. Thats great food for thought. Its reassuring to think that its not me that is doing something wrong. I've been experimenting and have not found even close to a suitable chip. With your help I now know a little more of what to look for and that is one more piece of the puzzle found. Chris.
 

Papabravo

Joined Feb 24, 2006
14,235
Could you add a DC offset to your DC source and use single supply with a virtual ground so that you can cover the range you need within the larger Vcc range?
 

DickCappels

Joined Aug 21, 2008
6,533
If getting all the way to ground is the challenge at hand you can do so with a resistor from the output to ground, even the ancient LM358 can work down to zero volts that way. Not sure about CMOS output structures though.
 

DickCappels

Joined Aug 21, 2008
6,533
If your signal is slow you can try this:
upload_2019-6-24_21-25-22.png

Demonstration circuit allows the output of an LM358 to be pulled through ground by using the collector current of a KSP10 -relies on photoelectric current generated in the collector-base junction driven by photons emitted by the avalanching of the emitter-base junction.

upload_2019-6-24_21-26-3.png
 

Thread Starter

Gearbreaker

Joined May 23, 2019
24
Use a negative rail charge pump to provide a Vee-

http://www.ti.com/lit/ds/symlink/lm7705.pdf
I had never heard of such a chip. I just printed out the datasheet and am sure I can use it in another project somewhere along the line. This project was calmed down enough to work properly just by loading the output of the amp more than I had *was 100K but now 50K. Thank you, the info will be very helpful.
 

Thread Starter

Gearbreaker

Joined May 23, 2019
24
If getting all the way to ground is the challenge at hand you can do so with a resistor from the output to ground, even the ancient LM358 can work down to zero volts that way. Not sure about CMOS output structures though.
You nailed it, loaded more than I had and it all fell into place. Doesn"t go quite to zero but for this project it is close enough. Should I need it end up not being low enough in the end, I will try your other suggestion. Also, Other peoples posts here are possibilities too. I am working with a non isolated ground and going below ground (underground, lol) would prove to be very tricky. Thank you.
 

Thread Starter

Gearbreaker

Joined May 23, 2019
24
I suppose that "tricky ground" is an unambiguous, highly technical term, that would defy your attempt at explanation. Could you possibly be more pretentious?
I do appreciate your input as well as everyone else who took the time to make suggestions. I am learning about circuit design and am off and running without too much experience. The circuit that I have been working on is to monitor part of a modulation circuit in a radio transmitter. Because it is monitoring modulation I was afraid of having any chance of creating a hum or worse having a ground loop affect the audio that the op amp monitors and ultimately controls. My circuit uses a separate power supply via a power transformer for isolation. However, last night I worked with the circuit and read a lot about op amp design and learned quite a few things. Apparently I can simply use a buffer and beyond that I can create a second ground or virtual ground. Although I had already solved my original problem I wanted to learn more about your and everyone else's suggestions. In the next stages of the project I may need to know more about op amps and am fascinated by their applications. With all of that said, I am only a beginner and am not always certain of what I am doing but am asking questions to learn. Isn't that what this forum is all about? Thank you again for your input, it is well appreciated.
I suppose that "tricky ground" is an unambiguous, highly technical term, that would defy your attempt at explanation. Could you possibly be more pretentious?
 

Thread Starter

Gearbreaker

Joined May 23, 2019
24
If your signal is slow you can try this:
View attachment 180294

Demonstration circuit allows the output of an LM358 to be pulled through ground by using the collector current of a KSP10 -relies on photoelectric current generated in the collector-base junction driven by photons emitted by the avalanching of the emitter-base junction.

View attachment 180295
After many attempts on variations of this concept I have learned that it does work but linearity suffers. Again, I am just learning and the more I learn, the more I learn that I have a lot to learn.
 

Thread Starter

Gearbreaker

Joined May 23, 2019
24
The plot thickens..... I finally have come to the conclusion that the only way to keep linearity is to create a split supply. I have ordered a center tapped transformer to do this. However, in my bench testing I have found that many op amps have linearity problems and accuracy suffers below 20mv in. My bench supply is +5v and -5v (but I should easily be able to squeak out +9v and -9v after the new transformer arrives). My new criteria is that I only need to go, on the high end, to 350mv in to get 3.5v out. All of this has eliminated the need for a rail to rail op amp. During my most recent testing I have made attempts with many parts. Perhaps it's a fluke but I found the NTE941M to be fairly accurate throughout the range (accuracy of approximately +/- 2mv out for any given input throughout the range). NTE's data sheet is pretty vague and I don't see anything which talks about the linearity or at least I don't see it if it is there. I am not partial to NTE components and am looking for suggestions for another part that will keep linearity in check as this one does. The only catch is that I need a quad version as the initial amp I have been speaking of is only the front end of a bigger project. Thank you to everyone so far for your help.
 

Thread Starter

Gearbreaker

Joined May 23, 2019
24
My experience is if you are trying to do high precision work you have
to trial RRIO parts. My rule of thumb is 12 bits over temp, V, all errors,
a challenging design.

For example, still many manufacturers not discussing the crossover distortion
inherent in their input stage design. Here is an example of where due diligence
was done -

View attachment 180197

That was a great piece of information. I think it's the key to me finding what I need now. Thank you.

Short answer is read datasheet, if you see no discussion of this move on.


Regards, Dana.
 

TeeKay6

Joined Apr 20, 2019
572
If your signal is slow you can try this:
View attachment 180294

Demonstration circuit allows the output of an LM358 to be pulled through ground by using the collector current of a KSP10 -relies on photoelectric current generated in the collector-base junction driven by photons emitted by the avalanching of the emitter-base junction.

View attachment 180295
@DickCappels
An effect I had never heard of. Is the KSP10 unique or can the same effect be seen in any BJT? Thanks!
 

DickCappels

Joined Aug 21, 2008
6,533
It works on many other transistors, such as: 2N222, 2N440, 2SD882L, BC337, BC546B, BF423, 2SB1068, 2SA775, LTA1266, KSP94, 2N3906, BC557, KSP2907, 2SA1020Y, MPS6519, MPSH55, MPS6562, and 2N2907. Come to think of it I cannot recall this not working with any transistor tested, though some are much worse than the others. With a 0.1% current transfer ratio, the KSP10 was the highest I found.
 

DickCappels

Joined Aug 21, 2008
6,533
The plot thickens..... I finally have come to the conclusion that the only way to keep linearity is to create a split supply. I have ordered a center tapped transformer to do this. However, in my bench testing I have found that many op amps have linearity problems and accuracy suffers below 20mv in. My bench supply is +5v and -5v (but I should easily be able to squeak out +9v and -9v after the new transformer arrives). My new criteria is that I only need to go, on the high end, to 350mv in to get 3.5v out. All of this has eliminated the need for a rail to rail op amp. During my most recent testing I have made attempts with many parts. Perhaps it's a fluke but I found the NTE941M to be fairly accurate throughout the range (accuracy of approximately +/- 2mv out for any given input throughout the range). NTE's data sheet is pretty vague and I don't see anything which talks about the linearity or at least I don't see it if it is there. I am not partial to NTE components and am looking for suggestions for another part that will keep linearity in check as this one does. The only catch is that I need a quad version as the initial amp I have been speaking of is only the front end of a bigger project. Thank you to everyone so far for your help.
Good move. No doubt that going to a split supply will make your job a lot essier
 

Thread Starter

Gearbreaker

Joined May 23, 2019
24
Good move. No doubt that going to a split supply will make your job a lot essier
Yes, I can't believe how much easier it is with a split supply. For my application, although not a quad, I believe the INA106 will be a perfect fit. I have the INA105 and it tracks perfectly. I have ordered the INA106 to test. The best part of my attempts to use a single supply is that I learned a lot of the limitations of going that route and read up on a lot of tips and tricks... I found a great learning video series.... https://training.ti.com/getting-started-current-sense-amplifiers Great for beginners like me. Thank you again for your help and forum too!
 
Top