I'm running the test procedures from the service manual on a 1980s Korg DS-8 synthesizer and I've hit a snag.
Annotated schematic (see the yellow):
The test procedure is to press a key on the keyboard with a certain basic patch in effect and adjust VR2 so as to remove the DC offset from the signal at test point TP1. The problem is that there's a whopping 7.44V positive offset at TP1 and adjusting VR2 has no effect on it at all (the audio AC component is present but it's closer to half a volt peak-to-peak). I've verified that all power supply rails have good voltages and don't have any significant ripple, and I've checked for proper voltages at VR2's legs (block-B is +15.11 measured, 15 nominal). VR2's wiper's voltage is showing up at pin 7 of NE572 compander IC10 as expected.
With a key pressed, there's an audio AC component into pin 7 (delta G) of IC10. Apparently IC10's output at pin 5 must be communicating to IC9 in current because there's no AC voltage component there; its DC bias is +2.497V which is the same as what I read at pin 6 which, per datasheet, is a nominal +2.5V reference voltage pin.
So I was thinking at first that some +15V supply is somehow punching through internally to IC9's output but then I got curious about the supply voltages for IC9 (on schematic, NJM4558; on silkscreen "4558DV JRC"). I thought one possibility was that maybe its negative supply were disconnected. It turns out that pin 4, which on the datasheet is supposed to be -Vcc, is grounded in this Korg board (see sub-schematic at lower left). That's what the DMM is also telling me. I notice that there is no actual ground pin on the 4558 and since the 4558 is often used in guitar pedals that only have a +9V supply, it seems to be the case that this chip is often implemented with a unipolar supply. But the test procedure has to be faulty; how can a unipolar-powered op-amp circuit like this ever produce a signal that swings negative - and therefore how could one ever hope to trim out the DC offset? Apparently Korg shipped a lot of units like this and apparently it kinda-sorta works, but my intuition tells me that some distortion might be caused by the limited AC headroom.
Does my troubleshooting make sense, and should I hack the circuit to move IC9 pin 4 from ground to the -15V supply?
Annotated schematic (see the yellow):
The test procedure is to press a key on the keyboard with a certain basic patch in effect and adjust VR2 so as to remove the DC offset from the signal at test point TP1. The problem is that there's a whopping 7.44V positive offset at TP1 and adjusting VR2 has no effect on it at all (the audio AC component is present but it's closer to half a volt peak-to-peak). I've verified that all power supply rails have good voltages and don't have any significant ripple, and I've checked for proper voltages at VR2's legs (block-B is +15.11 measured, 15 nominal). VR2's wiper's voltage is showing up at pin 7 of NE572 compander IC10 as expected.
With a key pressed, there's an audio AC component into pin 7 (delta G) of IC10. Apparently IC10's output at pin 5 must be communicating to IC9 in current because there's no AC voltage component there; its DC bias is +2.497V which is the same as what I read at pin 6 which, per datasheet, is a nominal +2.5V reference voltage pin.
So I was thinking at first that some +15V supply is somehow punching through internally to IC9's output but then I got curious about the supply voltages for IC9 (on schematic, NJM4558; on silkscreen "4558DV JRC"). I thought one possibility was that maybe its negative supply were disconnected. It turns out that pin 4, which on the datasheet is supposed to be -Vcc, is grounded in this Korg board (see sub-schematic at lower left). That's what the DMM is also telling me. I notice that there is no actual ground pin on the 4558 and since the 4558 is often used in guitar pedals that only have a +9V supply, it seems to be the case that this chip is often implemented with a unipolar supply. But the test procedure has to be faulty; how can a unipolar-powered op-amp circuit like this ever produce a signal that swings negative - and therefore how could one ever hope to trim out the DC offset? Apparently Korg shipped a lot of units like this and apparently it kinda-sorta works, but my intuition tells me that some distortion might be caused by the limited AC headroom.
Does my troubleshooting make sense, and should I hack the circuit to move IC9 pin 4 from ground to the -15V supply?
