Low Frequency Distortion?

Discussion in 'The Projects Forum' started by Ben_C, Apr 11, 2014.

  1. Ben_C

    Thread Starter Active Member

    Oct 19, 2010
    65
    4
    Hi guys, it's been while!
    I have a problem with a Mackie SA1521 woofer, theres very noticeable distortion at low frequencies between 40-70Hz, I've tested the majority of the transistors and referenced with a working unit I have and still cant find any fault with the components.
    I'm thinking it's around the low pass op-amp but not quite sure, do any of you have "I'm 99% sure it'll be this" answers?
    It works fine on any other frequencies it just seems it may be lagging or something..

    Thanks,
    Ben.
     
  2. t06afre

    AAC Fanatic!

    May 11, 2009
    5,939
    1,222
    In audio the lowest frequencies have the highest amplitude. This problem could be related to problems in your power supply unit. Like dried out filter caps. Or a bad bad diode(s) in the rectifier. A transformer with inadequate power/voltage rating may also lead to this
     
  3. Ben_C

    Thread Starter Active Member

    Oct 19, 2010
    65
    4
    Thanks for your reply :) Ive tested all caps and all seem fine, tested rectifier and seems to be all good, the transformer is separate to this board and I placed a working output board to this amp and it works fine so the problem definitely lies within this output board.
    Logic does tell me its clipping the edges which causes the distorted sound, It's just really frustrating how it measures good compared to the working one but still is distorted at low frequencies
     
  4. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    At the lower frequencies the woofer has more time for cone travel, so it could be a mechanical issue.

    If that is the case, you are simply driving the woofer with more power than it can handle so (at that frequency) it hits the end stops.
     
  5. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,542
    1,251
    I had a similar problem in my car after only 1 year, and it was loose voice coil wires. How old are the speakers?

    ak
     
  6. Ben_C

    Thread Starter Active Member

    Oct 19, 2010
    65
    4
    I put the amp from the other "working fine" speaker into this unit and drove the woofer, it sounded fine at all frequencies.

    The amp I removed (The one that sounds "farty" at 50Hz) I linked up to an Oscilloscope and increased the frequency slowly and noticed a chopped off more square wave between 47Hz to 80Hz then went back to sine thereafter..
     
  7. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,801
    1,105
    What cap tests did you do?
     
  8. Ben_C

    Thread Starter Active Member

    Oct 19, 2010
    65
    4
    Removed them from the circuit and tested with an Atlas meter ( ESR + Capacitance) and they measured around what they where valued at.. Do you think it strongly is related to capacitance?

    I've added a screenshot of the output if it explains it better.

    sine2.jpg
    clipped.jpg
    sine1.jpg
     
    Last edited: Apr 11, 2014
  9. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,542
    1,251
    If the input was constant at the three freqs, then it looks like the amp is peaking and running out of headroon. 25V Vcc, 21 V clip, about right. So, what component failure or cold solder joint could cause a narrow band freq response rise at 50 Hz?

    ak
     
  10. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    Photo 3 shows a sine reproduced ok at 80Hz and 20.179v.

    Photo 2 shows bad clipping at 50Hz and 21.000v, but the shape of the wave shows your sine signal was far in excess of 21v, looks more like 30v.

    All you are doing is overdriving the amp, so it clips the signal.

    Why is there any mystery here?
     
  11. t06afre

    AAC Fanatic!

    May 11, 2009
    5,939
    1,222
    You should show the input signal in another channel at the same time. That may help narraw down your problem. But as said before this look very like plain old clipping
     
  12. Ben_C

    Thread Starter Active Member

    Oct 19, 2010
    65
    4
    The signal input to the amp was constant and fairly low, the level starts to distort around half volume when adjusted the pot on the amp.

    The mystery lies where the other amp module, with the exact same level input, can turn up to max volume and sound great.

    So you think it could be just a simple dry solder joint? The solder does look a bit dull so this is a solution I can try on Monday. It seems far fetched but worth a shot :)

    I'll keep you updated no doubt ;) and thanks for your help guys!

    @to6afre I will upload images on Monday :)
     
  13. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,542
    1,251
    The amp is fine at 45 Hz and 80 Hz, but seriously clipping at 52 Hz. To me that sounds like a high-Q tank circuit rather than a linear amplifier, so I wonder what kind of component change could turn this thing into a 50 Hz tank.

    Or <strange thought> is the 50 Hz a clue about something synpathetic picked up from the power supply?

    ak
     
  14. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    Is the "other amp" identical in every way?

    If not, there is no point to comparing it.

    It's like saying "my car won't pull the caravan up the hill, but my other car does it ok". It's only valid if the two cars are identical types.


    Subwoofer amps often have "bass boost" of some type incorporated into the amp. Some low freq band will be accentuated. That is normal and to be expected.

    I would test the clipping as an output signal;
    100Hz; scope the largest possible sine out the speaker terminals, before clipping
    50Hz; do the same

    If both frequencies can output a 21v p/p sine without clipping, and both clip when the output tries to be 22v p/p, then it is simply clipping of the output amp.

    That eliminates issues like bass boost from the test procedure.
     
    Ben_C likes this.
  15. Ben_C

    Thread Starter Active Member

    Oct 19, 2010
    65
    4
    Yes, they're identical, same value components etc.. Both SA1521 Amps. The one in question, looks a bit older than the one Im comparing it to though as in solder joints looks "cloudy" rather than have a shine to them..
    I will keep you updated over the next few days. Thanks very much for your input. :)
     
  16. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    Make sure to compare the PSU voltage rails between the two amps.

    If they use different PSU voltages they will clip at different output amplitudes. Even with the same components the power transformers are bought from external suppliers and may have slightly different voltages.

    And a newer model of the amp might have a "tweaked" bass boost. That might be as simple as they changed one component value to increase or decrease the bass boost.
     
  17. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,542
    1,251
    I don't think this is plain old clipping, or component tolerance error. The 50 Hz photo isn't just clipped. The slope of the sinewave around zero is significantly larger, indicating more gain, and the peaks are seriously crunched. I ran a few numbers on the main amp schematic and didn't see a shelf equalizer or anything else that could cause this. In fact, the main amp is DC coupled with no servo and no freq components in the feedback loop. The power section seems overly complex, so there might be something in there.

    Also not sure how you get 500 W into 16 ohms with 25 V rails.

    ak
     
  18. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    Yeah I saw that and mentioned it in post #10. Looks like a standard bass boost to me bringing up the amplitude of the lower frequencies and I've scoped many brands of subwoofer amps over the years. :)

    If the power amp itself has no bass boost it might be in the preamp. Or it may even be an issue with the signal generator adjustment.

    Standard practice for gain and clipping testing is to use one scope channel on the amp input and one channel on the amp output, then normalise the input sine to 2v p/p (etc) for each test.

    The OP has only given us a couple of examples of the output waveform and we're all guessing at the rest of his test procedure. ;)
     
  19. t06afre

    AAC Fanatic!

    May 11, 2009
    5,939
    1,222
    Here is a tip Ben. I see from your screen shoots. That you use a LabView based scope application. So I guess your scope setup is PC based. Instead of taking a picture of the screen. You can just use the print-screen button on the keyboard to take a copy of the full screen. Then just paste this image into say Windows Paint. Then just save the picture. For forum work I recommend NOT using the BMP format but the PNG file format. As it give much smaller files.
    http://www.makeuseof.com/tag/awesome-screenshots-windows-7/
    http://www.homeandlearn.co.uk/BC//win7/taking_screenshots.html
     
Loading...