EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: kvresto on October 17, 2014, 07:09:30 am
-
Hi. I'm re-posting...I think I need an expert.
I want to convert a bipolar +12V to -12V ac signal into a single ended 0V to 3V signal. I have been simulating in SPICE a general purpose amp for this example and I have settled on the design shown in the pic. What I don’t know is whether this configuration with the feedback set up as an attenuator is stable. Will this design work or is it doomed when I actually build the cct? Is there an amp which is best suited for this, or even another method, I’m still searching for the best method.
As the cct is inverting, would a non inverting configuration do as good a jog or other?
Thanks
Cheers
k.v
-
What I don’t know is whether this configuration with the feedback set up as an attenuator is stable. Will this design work or is it doomed when I actually build the cct? Is there an amp which is best suited for this, or even another method, I’m still searching for the best method.
Hi kvresto,
1.) At first, the circuit should work as desired (use 1.333V instead of 1.3V).
However, you should increase the supply voltages at least up to 4 volts.
2.) Regarding stability - is the LMV358 unity gain stable? On the other hand- because the gain is below unity there might be a phase margin that is too small.
3.) In this case, it could be wise simply to reduce the input voltage range (simple voltage divider) and to increase gain (and change the dc offset correspondingly).
4.) In the latter case, a non-inverting configuration is preferrable (why not unity gain?) because of the large input resistance (no loading of the voltage divider)
-
Hi LvW.
Thanks for your input, on a phase margin error, although I don’t know at this stage how to check for it, any ideas? Some guide where to look and research.
Based on your thoughts I reconfigured the cct and came up with this one using a non-inverting unity gain amp. I won’t be using the LMV358, I want something that has a tight Vos Ios etc, any suggestions?
I think this will avoid any unseen issues and work just fine when I put it together on a breadboard and start testing. I would welcome any more thoughts and ideas.
Thanks again.
k.v
-
I would welcome any more thoughts and ideas.
OK - if you can afford to use three active components.
My former recommendation was based on a much simpler circuit, for example:
* Input voltage division (1k/11k) with a factor of 1/12.
* A non-inverting gain stage with G=1+0.5=1.5 gives an amplitude of 1.5 volts.
* Biasing the negative input with -3 volts shifts everything by +1.5 volts.
* This gives a peak-to-peak signal between 0 and +3 volts.
-
An inverting amplifier with resistive attenuation will be stable if the opamp is unity gain stable.
It is "noise gain" that matters and it is 1.125 times in the circuit of the first post so should not be an issue if the opamp used is unity gain stable.
-
An inverting amplifier with resistive attenuation will be stable if the opamp is unity gain stable.
It is "noise gain" that matters and it is 1.125 times in the circuit of the first post so should not be an issue if the opamp used is unity gain stable.
Yes - it will be stable. However, the phase margin is at its minimum value which might not be sufficient for some applications.
-
(https://www.eevblog.com/forum/projects/amp-attenuator/?action=dlattach;attach=113620;image)
:)
-
Well - of course, a pure passive realization is possible, however an output buffer will be necessary for further processing.
And, therefore, one has to decide which alternative uses the least amout of parts and/or has a suitable input resistance.
-
Well - of course, a pure passive realization is possible, however an output buffer will be necessary for further processing.
And, therefore, one has to decide which alternative uses the least amout of parts and/or has a suitable input resistance.
Yes of course, I was just trying to show that you don't need it to be active. Since I have no idea what impedances we are looking at and what the rest of the circuit will be I can't say that this is the best solution, just making a point here...
-
...an output buffer will be necessary for further processing.
Since kvresto never revealed such important things as source and load impedances, that is pure speculation.
-
...an output buffer will be necessary for further processing.
Since kvresto never revealed such important things as source and load impedances, that is pure speculation.
Yes - therefore I will add the word "perhaps". OK now?
-
.
Hi all, I want to build a front end that will interface to sensors plus be able to make differential measurements. Some examples are:
Resistive Photocell Flex, Force Sensors.
Potentiometers.
Analog Hall Effect Sensors.
Analog Temperature & Humidity Sensors.
Analog Range & Proximity & Depth Sensors.
Analog Gyros & Accelerometers.
To me this means I need a high impedance input, hence the two op-amps feeding into an ins_amp. There would be some instances where the “sensor” in question may need to source current, but I can work on that later I guess. For now reading the sensors listed above would be the start of a research project on accurate data acquisition, and my current research has led me to what I have posted .
Cheers
KV
BTW, PA0PBZ great cct idea thanks for the input.