DC Amplifier Oscillation Issue

Discussion in 'General Electronics Chat' started by Federov, Nov 6, 2010.

  1. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    The idea of this project was to build a current source that drives 3 high powered LEDs represented in this circuit by the resistances R9-R11. I realize LEDs are not linear but this is there maximum resistance resulting in the hardest condition for the current source. My problem is that X1 seems to be oscillating with added oscillation when the LEDs are turned off. I have come to this conclusion that X1 oscillates by itself because grounding the bottom of the 5Ω resistor still results in circuit oscillation. I was using a 741CN originally, but than changed to a 301CN and this actually seemed to reduce the problem. What really helped reduced the problem farther was adding capacitors on the positive and negative rails of my power supply (more/less the negative rail had the much larger effect). Ideally I need an op-amp that can handle 20V rails so I bought a couple of UA748INs, but they seem to oscillate REALLY badly(if anyone knows a fix for it you would be saving me some money! =P). After reducing the problem a lot I ran into the problem that when my switches (M2,M4,M5) were on (LEDs off) the oscillation became worse. My first guess has been that they are adding a capacitive load to the op amp, but you would think that Q1 would help stop this effect. I am stuck on this problem with no real solution or ideas on how to fix it or if my thoughts are even correct. The oscillation is roughly a volt(peak-peak).

    So what am I am looking for is a fix for the op-amps i bought and/or just a new part that performs like the 301CN and can handle +20/-20V, and help on how to fix the oscillation when the LEDs are off. All help is appreciated! A picture of the circuit is shown below.

    NOTE: I know the switch design is not ideal but with the project specifications the way they are I had no better choices.


    [​IMG]
     
  2. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    X2 provides positive feedback to X1 which causes oscillation.

    What do you need to build exactly?
     
  3. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    An accurate voltage controlled current source that provides at least 300mA. How come X2 only causes a lot of oscillation problems when the LEDs are off? =/ I guess i could convert the circuit to the more common voltage controlled current source that uses a sense resistor. Any better ideas?
     
  4. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
  5. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    Couple problems with that setup. First off its no longer a current source its a current sink =/ and my professors are very picky about that one. Second off my switching circuit doesn't work under that condition.
     
  6. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
  7. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    Thanks for the fast responses, its appreciated. The second circuit could work, just will have invert my input signal but thats not to hard. Any suggestions on a favored op amp that supports +/- 20V?
     
  8. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    I can't suggest you an op amp since I am not using them very often. Other people in this forum will do you if you are patient. :p

    You will need to consider the accuracy of the current source and the bandwidth in your choice of the op amp.
     
  9. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    Alright im back after building and testing the current source you so kindly pointed me towards (Thanks again mik3). I understand how the circuit works very well besides resistor R2. The only thing i can see it doing is pulling some of the current that normally would go into the base of Q2 away. Is this just so you dont begin to drive the current through the BJT larger than desired? Any help on what this resistor is doing would be appreciated, thanks for all the help so far.
     
  10. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    Possible uses are to reduce the gain of circuit as to improve stability and it helps the BJT to turn off faster.
     
  11. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    Thanks for all the help mik3. The current source is up and working well (simplified circuit of just the current source below). I have had to make some design modifications to the circuit you pointed me towards in order to properly use the LED load switches (swapped the rails around). I also removed the FET because it was causing oscillation issues (180 degree phase shift around the loop). The op-amp I have can source enough current that it isn't really a problem at this point. Also driving the op-amp with 17.5-20V (needed to get an Iout of 0-.5A) was a problem because it is driven by an op-amp with 20V rails as well so I couldn't get an output that close to the rails. This is why there is an added 750k and 250k which increases the input by 4/3's, hence the input only needing to be from 13.125V to 15V. Okay so the questions that I'm really just confused about.. sorry if some are obvious.. been stressed lately so probably just not thinking properly. All help from anyone on these is greatly appreciated. I need to understand this stuff by next week or i'm going to fail this project =(.

    1. Why was a FET and BJT used in the original design? what are the benefits? (just the op-amp doesnt need to source current? or?)

    2. Why is the feedback to the positive terminal? This is normally never done unless it is a comparator I thought.. How/why does it work in this situation?

    3. Can I replace the BJT Q2 with a MOSFET in order to eliminate the error due to the base current in the BJT? or does this lead to a problem I'm not seeing? I know a MOSFET adds more capacitance for the circuit than a BJT but are there other bad effects?

    4. Isn't a characteristic of a good current source is to have a really high output impedance? How does this circuit have a high output impedance while looking at the emitter of the BJT?

    [​IMG]
     
  12. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    Common (American) single opamps rated for a total supply up to 44V are the MC34071 and the TLE2241. Their inputs work as low as the negative supply and the outputs go down almost to the negative supply so they work well from a single positive supply.
     
  13. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    Thanks for the reply! I realize there are rail-to-rail opamps that can get me very close to what I needed but using the parts I had at hand this setup seemed to work just fine. The biggest problem is this design is sorta already laid out on perf board =x. Can you shed any light on the some of the questions I had? still greatly confused =P
     
  14. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,795
    951
  15. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    Thank you for the knowledge, just briefly read through the oscillation sections in the link and enjoyed they way they approached it all. The thing is I have fixed oscillation in the circuit by compensating the op-amp. I only care about DC signals so over compensating really isn't an issue. I'm still lost on the four questions I asked, if you have a link to answer any of those I would love it =).
     
  16. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,795
    951
  17. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    Haha i just read the entire article and thoroughly enjoyed it. Thank you for the link =). Unfortunately it still doesn't help answer any of the questions that I have about the current source mik3 pointed me too =P:

    1. Why was a FET and BJT used in the original design? what are the benefits? (just the op-amp doesnt need to source current? or?)

    2. Why is the feedback to the positive terminal? This is normally never done unless it is a comparator I thought.. How/why does it work in this situation?

    3. Can I replace the BJT Q2 with a MOSFET in order to eliminate the error due to the base current in the BJT? or does this lead to a problem I'm not seeing? I know a MOSFET adds more capacitance for the circuit than a BJT but are there other bad effects?

    4. Isn't a characteristic of a good current source is to have a really high output impedance? How does this circuit have a high output impedance while looking at the emitter of the BJT?
     
  18. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    How do you know the op amp was oscillating?

    Did you see it on a scope, or is this a simulation only?

    What was the frequency of oscillation.

    It is really annoying when a circuit starts oscillating, but this can only happen if the circuit has gain at the oscillation frequency.

    So a simple response would be to limit the bandwidth of the op amp. I know the 741 is internally compensated, but you can still add external compensation to curtail its bandwidth further.
     
  19. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    Yes it was oscillating on the scope and as mentioned previously i have used the compensation pins on the op-amp and the oscillation issue has been resolved. So the issue is not the oscillation its that i'm unsure about some of the design decisions made in the circuit mik3 pointed me too. The questions(4) I have are listed above any help with these questions will be appreciated, thank you =).
     
  20. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    The FET needs very little current into its gate, thus the op-amp is not loaded very much. This allows power BJTs to be used as to make a current source into the Amps range.

    Because the output is 180 degrees phase shifted. This is like negative feedback.

    You can but it may become unstable due to the higher input capacitance and the higher transconductance (Io/Vin gain) of the MOSFET.

    Small signal analysis will give you the answer. :p
    Think of it logically. If the output voltage changes, the current stays constant. This is like having a high impedance (current source) in series with a low impedance load. If you change the load impedance (as long as it stays well below the current source high impedance) the output current does not change much (ideally constant). This is because the high impedance dominates in the determination of the current.
     
Loading...