Roland E-35 keyboard with temporary noise/hum

[marcinado]

I have a Roland E-35 keyboard that is working great, however after playing music, when sounds are stopping, then there's noise/hum for couple of seconds. Problem is very obvious on internal speakers. On "headphone output" is less significant and I can't hear it on "line output".
I couldn't find "E-35 service notes" in the Internet but E-35 have similar "CPU Board" and "GSE Board" like in E-70 model and "Power Supply & Amplifier Board" is like in E-15 model.
I think capacitor/capacitors are guilty but optically all capacitors on boards looks ok. Can you help me localize source of problem? Then I can unsolder some elements and check them in LCR tester.
I've attached example audio file which is amplified and you can hear that after 3 second hum is gone and only noise floor is present.

[Audiorepair]

I assume you have ruled out the external power adaptor as the culprit?

[marcinado]

Yes. Checked on two different power adapters.

[marcinado]

Here are additional infos and some pictures with schematic diagrams and photos of the instrument boards.

The instrument has weird volume control circuit - the slider adjusts the sound level right after the DAC system, before the input to the power amplifiers. This would suggest that the problem is somewhere on the Power Supply & Amplifier Board.

Interestingly, even when the volume slider is set to minimum where the sounds generated by the instrument cannot be heard, the hum I described can be heard. Also, when the volume slider is at minimum and you start playing on the instrument then various noises can be heard on the speakers.

I cannot localize the source of the problem the only solution in my mind is to replace all electrolytic capacitors. But is it possible that other components apart from capacitors have died and can cause the problem? Can you advise me what kind/producer of caps should I buy with long life?

[Vertamps]

Its over 30 years old, caps are generally great for 15 years and really go downhill after 20 depending on the quality.
Go with Nichicon and Chemi-Con and its good for another 30 years, and there are so few caps the order is going be
$30 including shipping from Digikey or Mouser.
I always replace capacitors on most music gear older than 2003.
*Check you have the correct DC voltages from the power supply.
-William

[Audiorepair]

According the to data I have, the E15 requires 9v 1A,  the E35 requires 12v 1.6A.

This still may be just a power supply issue.

[marcinado]

According the to data I have, the E15 requires 9v 1A,  the E35 requires 12v 1.6A.

Yes I realize that.

[marcinado]

I've replaced all electrolytic caps with Chemi-Con and Panasonic. I found only two 4.7uF capacitors with 7.8uF, but the rest was OK. Unfortunately replacing all caps didn't help
It was hard work because PCBs are double-sided, with large ground fields and plated through holes. One leg of capacitors was soldered to such a large ground field and trying to unsolder it was a real nightmare. The electric desoldering iron I have only removes the outer layer of tin on such big ground fields. My soldering station with a thick tip and a maximum temperature of 480 degrees wasn't able to help much either. Fortunately, apart from some scratches on the PCBs, there were no other losses and the instrument works after assembly
I think it's some sort of ground/shielding problem. It looks like that one element/circuit is "sending" interference and another is receiving it. Or maybe somewhere is problem with weak ground connection? But it's only my small speculation.

[Vertamps]

Due you have an oscilloscope? Check backwards (or forwards if you prefer) from sound output all the way back to the DAC for the hum you are hearing, low frequency oscillations should show up on the scope.
Also check your 1/4" cable is good, verify the output is the same against the head phone jack output hum.

Also check befor and after IC 22 for hum just after the DAC. If you do not have an oscilloscope, construct a probe to check the signal lines and use an external amplifier/mixer and listen to the signal path to find the hum.m Just don't short any pins, which could end up being costly via your time and some parts. I know, iv shorted an $150 unobtainable VCO chip before.

[marcinado]

Good idea. I have simple PC USB oscilloscope. I will compare signals in different places.

[marcinado]

I couldn't capture the problem by oscilloscope probe because of high noise floor on idle signal. I've made another test. I've recorded signals from three different audio outputs. Speaker out, phones out and line out. As you can hear, noises on speaker output are most hearable. Phones out is much better, almost free of problem. On line out problem doesn't exist.
I've connected external speakers to speaker cables instead of internal 4 ohm speakers. On speaker with 4 ohm impedance noise is the same as on internal speakers. With 8 ohm speaker is better but not very much.
Internal speakers and phones out are connected in the same place. The only differences are resistors and ferrite beads in serial for phones output.
I have another keyboard Roland E-15 which has almost identical Power & Amplifier Board. There are only two differences between them. E-35 has additional resistor 2.2 ohm instead of jumper in power section (maybe because of different voltage value on input? E-15 9V and E-35 12V). And another difference is 18k resistor and 27nF capacitor connected to TDA IC. What's strange it's only on the right channel. On left channel there are empty pads like on both channels on E-15.  And problem with temporary noise doesn't exist on E-15 keyboard.
Do you think it's worth to replace TDA1905 chips? I can't spot another source of the problem different than TDA1905 chips.