I looked further into the power supply in my Nikon LS-50 (Cool Scan V ED) and made some measurements that should be helpful.
After turning the scanner off, all of the voltages drop to near zero in about a second, including the high-voltage side, which was 175 volts AC peak-to peak, or 124 VAC RMS (a little high for 110-120 volt residential supply).
The ~ 175 p-p voltage is present on the heat sink of the switching transistor, so there is a shock hazard there when the scanner is on. The temperature of this heat sink stabilized at 105 degrees F at room temperature with the case off and the scanner on for a while. High voltage is in other places too, but the heat sink is more exposed to clumsy or naive fingers.
The resistances (in ohms) at the power supply header for the supply voltages were:
3.55 k for the 15.3 volt supply, 2.6 k for the -12 volt, and 53 for each of the 5 volt lines.
The 53 ohm load could be why the supply voltage drops so fast after turn-off, as well as a low current capacity.
If there is something wrong with the circuit, the voltage may not drop this fast and the shock hazard would remain longer -- maybe much longer.
The ripple on all the lines was less than 2 mV, and there was about 5 mV of high frequency noise (in the many mega hertz range) on all of them. SMPS's tend to be noisy, anyway, but some of the high frequency noise could be FM radio signals picked up by the oscilloscope probe cable -- I did not try to separate this out, but have seen similar levels on non-SMPS devices and picking up FM signals is not unusual on longish cables.
The above voltages and resistance measures can help diagnose a problem with a scanner. The 8000 model is probably different, but it has similar voltage levels. Values widely different from these would indicate a problem. If the loads (resistance) on the header are reasonably close (and not very low, like < 5 ohms, or infinite), that would point to an okay circuit beyond the power supply. The power supply can be disconnected and loaded with appropriate resistors to see if it produces voltages close to these.
Of course there could be circuit problems that these measures would not detect, but getting them is easy and can point to obvious problems. The power supply would be a lot easier to fix than the rest of the circuitry. G. Shtengel's website mentioned in post #2 is helpful.
After turning the scanner off, all of the voltages drop to near zero in about a second, including the high-voltage side, which was 175 volts AC peak-to peak, or 124 VAC RMS (a little high for 110-120 volt residential supply).
The ~ 175 p-p voltage is present on the heat sink of the switching transistor, so there is a shock hazard there when the scanner is on. The temperature of this heat sink stabilized at 105 degrees F at room temperature with the case off and the scanner on for a while. High voltage is in other places too, but the heat sink is more exposed to clumsy or naive fingers.
The resistances (in ohms) at the power supply header for the supply voltages were:
3.55 k for the 15.3 volt supply, 2.6 k for the -12 volt, and 53 for each of the 5 volt lines.
The 53 ohm load could be why the supply voltage drops so fast after turn-off, as well as a low current capacity.
If there is something wrong with the circuit, the voltage may not drop this fast and the shock hazard would remain longer -- maybe much longer.
The ripple on all the lines was less than 2 mV, and there was about 5 mV of high frequency noise (in the many mega hertz range) on all of them. SMPS's tend to be noisy, anyway, but some of the high frequency noise could be FM radio signals picked up by the oscilloscope probe cable -- I did not try to separate this out, but have seen similar levels on non-SMPS devices and picking up FM signals is not unusual on longish cables.
The above voltages and resistance measures can help diagnose a problem with a scanner. The 8000 model is probably different, but it has similar voltage levels. Values widely different from these would indicate a problem. If the loads (resistance) on the header are reasonably close (and not very low, like < 5 ohms, or infinite), that would point to an okay circuit beyond the power supply. The power supply can be disconnected and loaded with appropriate resistors to see if it produces voltages close to these.
Of course there could be circuit problems that these measures would not detect, but getting them is easy and can point to obvious problems. The power supply would be a lot easier to fix than the rest of the circuitry. G. Shtengel's website mentioned in post #2 is helpful.
