Korg SDD-3000 Program drop out on Audio Peak

[oc disorder]

Hi
Having a few problems with this Korg SDD-3000 delay.
Replaced the backup battery (the old one had leaked) and most of the
electrolytics in the power supply for modern 105 deg types ..On the front panel every thing seems to work fine ie remembers program/patch settings
after power off and on again.. delays work ok can get chorus/flanging type
sounds on very short settings etc BUT when I drive the input into the red on the front panel headroom VU the digit display reverts to the default setting
(program 1).
My initial impression is the audio section is draining power from the digital
section on peaks. If the input is set to a lower level (-5dB ,-10dB) so that it doesn't max out everything seems to work OK.
Anyone any clues ?
In the meantime going to try to trace the supplies to both sections.

[Enzo]

There is no audio power, this isn't a power amp.

How about a shorted peak indicator or swomething? Every time some indicator flashes it winds up shorting out the supply briefly. Just a thought.

[oc disorder]

Evening Enzo , the last led in the vu display lights up ok the program and delay time digits dim then goes back to default program 1..have a copy of schematic with some bits missing now gonna chew it over with a cup of coffee..nearly lunch time here.
I see what you mean "this isn't a power amp" but it sure feels like the audio section draws more power when its got a big signal to handle but I suppose
its just an illusion... the audio chips would constantly be drawing current
so I guess the supply doesn't know if there is audio going through or not
that is the supply would be consistent regardless of the input.
I like your logic... maybe I'll check where the VU gets it's power from and see how it relates to the power to the processor

[Enzo]

The audio circuits are class A. the current is steady, thus the power is steady. Other than LEDs using current, the circuit draw should be fairly stable. I suppose some of the digital circuit couold draw more as it gets more active, but the circuit should easily handle that.

The VU works OK, but the numerical dims - well there is a clue. You could disconnect the display and redo the experiment. You should be able to determine from the sound if the unit defaulted again. If it no longer resets, then the display is troubled. If it still does it, perhaps the display driver?

Try just changing numbers on the display - is ther any number or combination of numbers that dims it? or look in the owners manual. Are they using the comma or decimal point as an indicator of some sort?


Now a completely different thought. Did it do this before new caps were installed? What if some cap is in backwards but doesn;t care until signals get to a certain amplitude that makes it balk somehow?

[jrfrond]

I don't have a schematic in front of me, and will not until tomorrow, but my gut feeling is that this is strictly digital in nature, and that the input VU meter is simply an innocent bystander, indicating that you have crossed some bit threshold in the DSP circuit. The LED display, as well as the reset issue, are all logic-level faults.

[Enzo]

The dimming thing sticks in my mind though.

I don;t know why this didn;t smack me in the face the first time. HOW did you clean up after the battery leaked? The leaked stuff is conductive and there could be a spurious pathway that doesn;t belong there, and it only confuses the system when certain combinations of states are there.

[Steve Conner]

Well, here's what I think's going on.

When you replaced the electrolytics, you used the wrong value, or put one in backwards, or forgot to solder it, or whatever. But the effect is that the +5V rail (or +3.3 or whatever it is) now can't supply enough current to light all of the VU meter LEDs. (Back before today's hyper-efficient LEDs, you used to run them at 20mA each.) When the raspberries go on, the regulator drops out from ripple, causing a voltage brownout that resets the CPU.

You can test this hypothesis by hooking a scope to the rail and looking for brownouts.

[oc disorder]

HOW did you clean up after the battery leaked?

It had dried up and I cleaned it with a cloth soaked in Isopropyl Alcohol.
Hadn't been used for a while.
The battery and power supply are in a seperate compartment to the main boards.
The leakage was only in one section and it was nearest the +'ve 15 supply the 5 volt regulator is further away.
Two multi cables take the supply to the two boards one is mainly analog the other digital.
The analog board also has the led display driver on it (AN6878)and is one of only two things powered by the 5v rail.The other is the mute relay circuit.
I was careful when I fitted the replacement capacitors. The unit is well made and accessible and the boards are clearly silk screened with the component reference number and polarity. The majority of the capacitors were in the analog section the + & - 15volt rails.However I did double check .
When I unplugged the display (Vu) it did not drop out which makes me wonder about loading of the 5volt rail.
The attached AN6878 is the schematic from the data sheet which is virtually identical to the Korg.Thinking of trying different values of the led current limiting resistor although thats not fixing the fault ...more of a temp work round.
When the numerical display goes off it usually displays 0 (only on one of 5 number/letter displays)for a brief second(as opposed to a long second) then reverts to patch 1 with the whole 5 operational with the patch and time setting.This is the same as power up.
It seems to improve when left on for a long time and also I have checked
that it retains it's setting after power off.
I do see a bit of ripple on the scope but minimal when it shuts down.
Have now replaced the regulator 4 x 220uF 16v on the logic board and resoldered the connector pcb mounted sockets.
There is also a row of tantlum capacitors on the logic board but the supply is a minimum of 4.97volts there so I guess they are ok.
My experience with those is they short out.

[Enzo]

Well, if you unplugged just the LED VU bar and the unit no longer resets, then your problem ought to be on that board. With it disconnected, measure the voltage drop across each LED.

[slomax]

Hi there...did you ever get your unit working? I'm experiencing a very similar symptom with my sdd-3000 (once headroom goes into the red, the program resets). Also noticing issues with the numerical display now, it flickers and is dim showing what appears to be all "8's". Any idea?