0
\$\begingroup\$

The old pair of M-Audio Studiophile AV40 active speakers (one box containing all the amplifier parts and the second is just a passive box) I have have developed a clear low-frequency hum. This hum is present independent if anything is connected to (any of) the input or not, and the hum level is independent from the playback volume. Also, it is present only in the speaker output, and the headphone output is clean. Since the headphone output is taken before the main amplifier chip input, that points towards the chip.

The circuit is based on TDA7265 and is quite close to the one in the AV30 models, from which the schematic can be found online. Internal view of the speaker is available in a repair video. Two-sided power supply without proper regulator, but with a zener, and stereo operation. The cross-over for the one speaker is built on the same PCB as the amplifier.

As the first step, based on online suggestions, I changed the main filter electrolytic caps, 4700 uF and 220 uF ones, even though they seemed visually healthy. This did not remove the hum.

After this, I ended up disassembling the speaker a bit more. Visually I don't see anything suspicious, but I don't have a trained eye for this. During the disassembly, I made the important observation: when the PCB and heatsink are attached to the backplane and they are close to the mains transformer, the hum is present. When increasing the distance between the two, the hum goes away. There is a metal shielding on the transformer and the heatsink is grounded and a piece of sheet metal is between the PCB and the transformer. Given that the speakers did not have the hum earlier, the screening should be enough, but is not anymore. So something has aged badly or some physical damage (though nothing obvious is visible) has broken something.

I would appreciate any suggestion what to check next. I was thinking to swap out the 0.1 uF supply by-pass ceramic caps or the amplifier chip itself.

\$\endgroup\$
7
  • 1
    \$\begingroup\$ Check that the bridge rectifier is working correctly and producing full-wave rectification. If one of the diodes has gone open-circuit in the bridge you will get double the mains hum or worse. \$\endgroup\$
    – Andy aka
    Commented May 30 at 14:07
  • \$\begingroup\$ Measuring in-circuit with a cheap digital multimeter on the diode setting: between the bridge's both AC inputs and the + output the voltage drop is about 480 mV, as is between the - output and the AC inputs. If this is not enough, I can remove the bridge from the circuit and measure alone on the weekend. \$\endgroup\$
    – paulus
    Commented May 30 at 15:03
  • \$\begingroup\$ If it's no bother, take it out or, if you have an oscilloscope you could measure the ripple voltage and post a picture. \$\endgroup\$
    – Andy aka
    Commented May 30 at 15:36
  • \$\begingroup\$ I removed the rectifier bridge from the circuit and measured it. It seems to be functioning correctly: about 520 mV shown for individual diodes and not conducting in the wrong direction. \$\endgroup\$
    – paulus
    Commented Jun 1 at 9:37
  • \$\begingroup\$ One more off the "to-check" list!! If you have some test equipment you might be able to measure the frequency of the hum you hear. \$\endgroup\$
    – Andy aka
    Commented Jun 1 at 9:38

1 Answer 1

Reset to default
0
\$\begingroup\$

The schematic appears to show a shield on the right output, but not the left. I presume that mechanically the right put up board must be closer to the transformer.

The output has an LC filter which could be picking up the AC line frequency (60hz in North America). At that frequency the sheet metal shield would need to be relativ thick (or high permeability) to divert any magnetic field coupling. Both it and the transformer would need to be positioned properly for this to work. Is it possible that this shield has shifted from it’s original position somehow?

The grounding of the shield would be to shield against electric field coupling.

\$\endgroup\$
3
  • \$\begingroup\$ The amplifier and left channel cross-over are built on the same PCB inside the left box. Since there's only a normal 2-conductor speaker wire going to the right-side box, the right side cross-over is inside there and this physical separation is probably what the schematic is trying to convey. Well, the shield is shifted at least now, as it was necessary to detach it to have any real access to the PCB. Still, any manual alignment (there aren't too many alternatives) and making sure the shield is grounded doesn't help. \$\endgroup\$
    – paulus
    Commented May 31 at 13:35
  • \$\begingroup\$ A further observation: the humming is present also in the right channel. I.e., in the channel connected only with the 2-conductor loudspeaker cable and whose cross-over is in the separate box. If the left channel cross-over receives the transformer hum, then it's somehow coupled also in the right channel output. \$\endgroup\$
    – paulus
    Commented Jun 1 at 9:40
  • \$\begingroup\$ I finally had the time and motivation to debug this a bit more. Swapped out the amp chip completely, just in case, but no help. Provided an extra grounding to the metal shielding, but no help. Playing with the alignment between the PCB and the transformer. It does seem that the hum is stronger when the LC filter is closer to the transformer while there is no big effect if the amp chip is close to it. So, Eric's hypothesis of the LC receiving the line hum might be on the right track. All parts installed, there aren't many mm wiggle room in anything, so I'm not sure what could've shifted. \$\endgroup\$
    – paulus
    Commented Jul 26 at 12:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.