Advertisement

Blog

Op Amps: Transconductance versus Voltage Feedback

Although I used op amps in many feedback loops over the years, the subject of transconductance op amps versus voltage feedback op amps has always intrigued me. Therefore, I thought it would be a good issue for the Planet Analog audience to debate. Incidentally, these devices are often referred to as OTA’s for Operational Transconductance Amplifiers.

OTA’s perform a transconductance operation where a voltage input becomes an output current. Voltage Operational Amplifiers output a voltage based on the differential voltage of an input.

First and foremost, I don’t consider myself to be an authority on the subject of op amps. I simply closed feedback loops as per the type of op amp that I chose in the PWM IC that I employed to control a power supply. Most of these supplies switched at a rate of hundreds of kHz with a range up to the MHz on occasion. Some were in the tens of kHz range. With Nyquist criteria for a roll off of less than half the switching frequency, it’s easy to see that I wasn’t even close to the bandwidth of a typical op amp. Furthermore, as long as I employed proper signal acquisition and precautions such as grounding and sensing, I really didn’t have to worry too much about noise considerations.

However, realize that a power supply and a switching power supply in particular, generates a fair amount of noise. There were some simply “rules of thumb” that often made the creation of a stable feedback loop much easier. One of which was when using a capacitor and resistor in series while feeding back a voltage op amp, the capacitor should be tied to the low impedance output in order to avoid coupling noise into the high impedance input. To this day, I see these components reversed. And yes, that was once an interview question that I aced.

One of the benefits of my career was the ability to work with the IC designers in addition to working in the creation of the module level power supply that used the controllers. I once asked a designer why they used transconductance amplifiers versus voltage feedback types.

The response was a shrug of the shoulders and a comment as to “it was easier”.

Easier how? I never did inquire further. Recently however, I did find a reference to transconductance amplifiers being easier however there really was no explanation as to why (“OTAs – in comparison to voltage opamps – can be easily realized as integrated circuits (simpler design, less stages).”). However I noticed that OTAs are often followed by a buffer stage that converts the output from a voltage to a current. So how can they be a simpler design unless both stages together are “simpler” than a voltage operational amplifier? I would welcome a designer’s comments here.

I also recalled reading that voltage amplifiers are often internally compensated. Who gave you permission to take away the fun of closing the feedback loop? Give it back, give it back.

My research did lead to some interesting comments on control signals, noise immunity, and bandwidth of operation. One blog response talked about the ability to use a “third” control of an OTA bias current thus enabling the ability to control the volume of multiple audio channels with a single signal. Xavier Ramus of Texas Instruments had both an IEEE presentation and an Application Report on noise immunity as well as bandwidth comparisons.

Another point that was made was OTA capacitors reference ground and should never have a capacitor across the feedback resistor. I read it, but I just can’t find the reference again. There were also references to OTA’s having faster slew rates. There were distortion comparisons of each of the devices as well as some hints on applications.

As with any research effort, there were some niceties that surfaced. I find the most helpful little details in application notes and this search was no exception. As an Applications Engineer, I coined the phrase, “An app note a day. You’re smarter that way.” It rang true here as there was a very useful reference to tracking down oscillations in the TI presentation by Ramus.

In the same manner that the research provided additional insight, there was some historical information as well. A tutorial that was originally posted in 1985 and upgraded in 2000 was a stroll back in time to the evolution of today’s integrated circuit. Of particular interest was the onset of CMOS and the references to biCMOS as technology transitioned. OK that period of time was only two thirds of my career however I’d never admit to being that old.

Of course being a Bob Pease fan, I also searched “Pease transconductance” and found articles on transistors as well as op amps written six years apart. As for Bob’s personality, it shines through in the transistor article. BTW Bill Gates, transconductance is a common industry term so why is your software continuously underscoring it with an angry, red line?

One of these days I’m going to sit down with a couple of evaluation boards, a scope, and perhaps some SPICE models and see for myself how this works. I’ll do it right after I scan all of the photos of my kids growing up and paint all of the gorgeous scenes I see daily while travelling the west. In other words, fat chance of this ever happening. Finding two, well matched op amps of the voltage feedback and OTA style would be the first challenge and in itself could be an impossible task.

There is no closure in this particular blog as I would rather open a subject for the audience to debate and provide further information on. And for you the application engineer that is looking for a subject to write about, this debate seems to offer an opportunity to further the quest for analog information. The latest references I found were from 2013. An editor once told me that magazine subjects in technical journals resurfaced on the average of eight months. Maybe it’s time for an update.

References

  1. OTAs – in comparison to voltage opamps
  2. CMOS transconductance amplifiers, architectures and active filters: a tutorial” 1985, updated in 2000, Sanchez-Sinencia and Silva-Martinez.
  3. Voltage feedback vs. current feedback amplifiers: Advantages and limitations”, IEEE presentation, Xavier Ramus, Texas Instruments, 2009.
  4. Demystifying Operational Transconductance Amplifier” SBOA117: A Texas Instruments Application Report, Xavier Ramus, May 2009 revised April 2013.
  5. What’s All This Transimpedance Amplifier Stuff, Anyhow? (Part 1)” Jan 8, 2001Bob Pease | Electronic Design.
  6. What’s All This Transconductance Stuff, Anyhow?” Nov 21, 1994Bob Pease | Electronic Design.

10 comments on “Op Amps: Transconductance versus Voltage Feedback

  1. Michael Dunn
    June 24, 2016
  2. vikas.rahul77
    June 27, 2016

     

    Hey!
    One question always bothers me, The concept of virtual ground !
    Can anyone provide a detailed explanation of same? I read many literature but the more I understand the more I get confused!!
    Please provide some good explanation 🙂

  3. studleylee
    June 29, 2016

    Virtual ground is usually a concept used when you have a single supply, say +12vdc and you want to create a fake bipolar supply to use( as an example) an opamp in a more usual way of having both a postive and negative supply( relatively). By creating a voltage that's 1/2 of the +12vdc, ie +6vdc, and calling that “virtual ground”.

    Then using this virtual ground as your common referance ground, the circuit now has a -6vdc( was gnd), 0vdc(=VirtualGround(was 6vdc)), and a +6vdc supply(was 12vdc).

    So it's a biasing thing and a relative thing. One important note is when creating the 1/2 voltage source it has to both source and sink in good regulation. If you are using a quad opamp you can dedicate one section as a unity follower hooked to a voltage divider across the 0vdc and 12vdc to create the 6vdc tap( or your new virtual ground point ) Also remember that your output is still 6vdc up from the “real world” ground so if its an audio signal, you will need a dc blocking coupling cap to bring the signal out as regualr gnd referenced AC. 

    Hope this helps, -Lee Studley

     

     

  4. vikas.rahul77
    June 29, 2016

    Hey Lee!!
    Thanks for your explanation my queation was regarding OP-AMO virtual ground concept of non-inverting and inverting terminals. Why does it happen in negative feedback configuration?

  5. vikas.rahul77
    June 29, 2016

    Hey Lee!!
    Thanks for your explanation my queation was regarding OP-AMO virtual ground concept of non-inverting and inverting terminals. Why does it happen in negative feedback configuration?

  6. rambirsinghjal
    June 30, 2016

    good post like dis

  7. studleylee
    June 30, 2016

    I got you now. That refers to the op amp being a differential engine.  Negative feedback operation is such that the output will do what it needs to do voltage-wise to cause/drive the voltage divider comprised of Rfeedback, intersectiing at Vin-, then to Rinput to be equal to the Vin+ input. It does this by being a high gain element drivien by the difference of of Vin+ – Vin-. 

    So it is a voltage servo.

    The opamp  in negative feedback connection drives the output to cancel the difference voltage 2 inputs. Does that help? Take a meter and measure across the Vin+ abd Vin- and you will see they should be equal. If you are using a bipolar supply it will usually be at zero volts then.

  8. D Feucht
    July 4, 2016

    Scott,

    Some confusion over OTA terminology can occur because as you are using OTA , it refers to what otherwise are called current-feedback amplifiers (CFAs) or what others prefer to call transconductance-feedback amplifiers, which they are. These op-amps with low-resistance inverting inputs do not have a closed-loop transconductance transfer function; they are voltage amplifiers. However, their feedback path is often best analyzed as a transconductance because the output of the feedback path is a current. That means the summing junction of the classic feedback block diagram sums current, not voltage, and the closed-loop amplifier must then have a pre-loop transfer function that is also a transconductance: a voltage to current conversion.

    OTAs, as I understand the expression, refer to amplifiers such as the LM13700, which have current out, voltage in amplification which is transconductance. But they are not CFAs. I explain CFAs in my new book, Transistor Amplifiers, linked at innovatia dot com and analyze an amplifier as first a VFA, then as a CFA. The closed-loop gain in both analyses is the same. This is possible because any amplifier with resistance > 0 ohms at the inverting input can be analyzed either way.

  9. Effective-Technical-Writing dot com
    July 5, 2016

    Thanks for the clarification.  Glad you wrote a book and shared your knowledge.

  10. Effective-Technical-Writing dot com
    July 5, 2016

    Thanks for the clarification.  Glad you wrote a book and shared your knowledge.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.