Noise of filter

Discussion in 'General Electronics Chat' started by fragrance2008, Dec 15, 2009.

  1. fragrance2008

    Thread Starter New Member

    Nov 18, 2009
    22
    0
    Hello, I am in trouble again. I designed an active filter. When I built it on PCB, it has a nosiy output. Could someone help me to check my circuit and tell me the possible reasons?

    Attached is the schematic of my circuit. Thank you for your time.
     
  2. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    In general, a 741 op amp is only going to work predictably with +/-15 volts as the supplies. I would highly recommend a nice inexpensive TLO71 as a better exemplar of an op amp.

    If you need the signal to be able to assume some maximum level, check the specifications on the op amp. Some will go rail-to-rail, meaning they can have an undistorted output level equal to the supply rail. Most are limited to a level that is a volt or two less than the supply.

    The 741 will have its output limited to somewhere between 2 and 3 volts of the supply voltages. Some of that noise you are experiencing could be resulting from the output clipping - being unable to keep following the signal. The signal goes from a slope resembling the signal to a flat line.

    There are no bypass capacitors shown on the supply lines. It is best practice to place a .1 uF cap close to the power supply pins of an op amp.
     
  3. fragrance2008

    Thread Starter New Member

    Nov 18, 2009
    22
    0
    Could you explain this part in more details? how can I figure out this problem?
     
  4. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    The 741 opamp is 42 years old. Back then nobody cared that it is very noisy (hissss). It has many other problems such as a max output level frequency of only 9kHz. Above 9kHz its output is ramps (slopes).

    The TL071 opamp works perfectly with a total supply voltage as low as 7V and is low noise. It is used in many hi-fi sound systems. It has a max output level frequency of 100kHz.
     
  5. fragrance2008

    Thread Starter New Member

    Nov 18, 2009
    22
    0
    sorry, i forgot to mention that I only use 741 for simulation, in the real circuits, i used AD797.
     
  6. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    The difference between the extremely low noise and very wide bandwidth AD797 and the lousy old 741 opamp are like the difference between an F-16 jet fighter and a Model T Ford car. Thanks for wasting our time.

    Because the AD797 operates up to very high frequencies then maybe it is oscillating in your circuit due to input-output or ground coupling or because its supplies are not properly bypassed with capacitors.

    Please describe the noise.
     
  7. fragrance2008

    Thread Starter New Member

    Nov 18, 2009
    22
    0
    Sorry for wasting your time.:(

    the output appears in the similar sine wave as input signal. the only difference is the output signal looks more thick, just like drawing something with a big brush. When i zoom in the signal, it seems that there is some very high frequency components. is it because the high caps, 1uF?
     
  8. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    Your op amp is probably oscillating. As beenthere said,
    and Audioguru said,
     
  9. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    1) I do not see 0.1uF and 4.7uF bypass caps from the opamp supply pins to ground. See figure 35 in the datasheet, page 13.
    2) What kind of capacitors are you using? Electrolytic would not be good for this purpose.
     
  10. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    SgtWookie, I know that you know this, but our OP probably doesn't.
    That's one of the problems with simulations. Unless you explicitly add power supply distribution parasitics (R, L, and C), the supply will have zero impedance and will not need bypassing. In the real world, they are almost always required.
     
  11. fragrance2008

    Thread Starter New Member

    Nov 18, 2009
    22
    0
    thank you so much.
    since somebody has mentioned caps before. i have already tried to add 0.1uF capacitors both to the positive and to the negtive supply, but the output signal remains the same. And i also changed the electronlytic caps. When i tested the whole circuit, the first two stages are ok. The last stage is the filter, could it be something wrong about the design? I simulated it, it was all right. but my tutor said you might design a new circuit and he also said some circuits you got from internet might not work practically. (he is very busy, and doesn't have much time to guide me.)
     
  12. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    OK, are you feeding in 0.1VPP in, and what are you getting out? Are you doing this at resonance to the filter?
     
  13. fragrance2008

    Thread Starter New Member

    Nov 18, 2009
    22
    0
    when the input is connected to the ground, it has an output of around 3.3 Vpp, and the frequency is around 2-4 MHz.
    BTW: what do you mean by resonance? i calculated the transfer function, but ideally, this circuit can't resonate.
     
    Last edited: Dec 16, 2009
  14. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    Are you placing the 0.1uF caps as close to the op amp power pins as possible, with very short leads? Do you have a good ground plane?
    Why are you using such a wideband amplifier for a 10Hz lowpass filter? Power supply bypassing can be tricky for wideband amplifiers, especially on a breadboard. Have you considered a more "docile" amplifier, such as TL071?
     
  15. KL7AJ

    AAC Fanatic!

    Nov 4, 2008
    2,040
    287

    Hey! I'm older than that and I'm STILL perfectly serviceable. :)
     
  16. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    You don't use electrolytic capacitors in an accurate lowpass filter. They have inductance that cause them to be a useless capacitor at high frequencies. Then your high frequency opamp oscillates. Also their tolerance is horrible. Also the capacitors in your circuit have AC across them and most electrolytic capacitors are polarized.

    I use 5% metallized plastic film capacitors up to 1uf and they look like this one:
     
  17. fragrance2008

    Thread Starter New Member

    Nov 18, 2009
    22
    0
    when the input is grounded, the tested value of this input is about 30 mVpp, and the output is about 170 mVpp. is the circuit still oscillating?
    Sorry about this silly question, i really don't know.
     
  18. fragrance2008

    Thread Starter New Member

    Nov 18, 2009
    22
    0
    yes, i have already added caps to the op amp, but not very close, i will try to rebuild the circuit in a new pcb.

    The ground in is a long metal line at the bottom, i got 30 mVpp noise when i tested the ground. how can i improve this ground?

    My tutor said the op amp AD797 is easy to design circuits due to high gain and wide bandwidth. so far i have only ordered AD797. i suggested changing op amps, my tutor said the problem was not the main reason for oscillation. was he right or not?
     
  19. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    You don't need pads for the caps, solder them directly to the power supply leads on the op amp. Industry does this all the time. If you are worried about shorts you can add insulation on the exposed lead wires, usually scavenged off regular wire.
     
  20. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    Does your tutor actually have experience with using AD797?
    Personally, I would use the TL071 series for your application.
     
Loading...