Electronic Repairs

Korg B1 Music Keyboard C, E, G# Keys Not Working

With this faulty keyboard most of the C, E, and G# keys weren't working, only one octave from the lowest end had all keys working, but with this type of fault with repeated same notes not being detected it suggests a problem with the keyboard matrix as multiple notes are connected together. I did look online for a similar fault and actually found a site where someone had the same problem with the same keys not responding:

https://www.ifixit.com/Answers/View/655548/Korg+B1+-+specific+notes+don't+work,+problem+with+circuitry

The site linked above was a great help to get an idea as to what could be wrong and after taking the keyboard apart and cleaning I eventually found what looked to be liquid damage to one of the tracks which I've highlighted with red arrows as follows:

There are two switches for each key so that pressure can be sensed when a key is pressed and each of the 2 switches per key has a thick track which is quite resistive and perhaps made of carbon. Despite apparent damage to two pads for the second C key, as seen in the photo above, the bottom one is fine and there is still a working connection to it, however, the track going off to the right of the top pad has been severed.

The fix was to clean the pad and scratch with a knife to reveal some copper which I also did for the track to the right of the pad and then I was able to solder across as to join the track back to the pad as can be seen below:

It's not attractive but it gets the job done and I'll need to clean the PCB further but repairing that one track restored all non-working notes.

SK-20061B Music Keyboard Certain Keys Not Working

I was asked to help with the repair of a music keyboard having model number SK-20061B, which had the fault that one of the keys was not working and it was also noticed that upon power on only the first key would briefly light instead of each key in turn. Checking all of the keys in turn it was revealed that actually 14 keys were unresponsive, and grouped into pairs, and the pairs were 8 keys apart (including black notes). Opening up the keyboard with the power off I checked the suspect key switches worked by using my multimeter and indeed when pressed my multimeter responded, confirming the switches was working. But even after cleaning one of the faulty switches both on the PCB and the carbon disc that makes contact with it I still could not get the note to play by pressing the key.

The key switch PCB has two ribbon cables that connect to the function button (instrument select/volume/etc.) PCB which in turn connects to the main board (note: the main PCB is poorly soldered with multiple bodges, which was a surprise). By directly connecting a few of the connections from the key switch PCB to the main board the dead keys worked and upon power on each of the keys lit in turn. Thus the function button board was at fault so it made sense to remove it from the keyboard to see if there was anything obviously wrong. Then I saw what was wrong: one of the function buttons had been forced into the on position and by re-fitting the rubber button the keyboard was now fully functioning.

Going back to the original fault, because a function key was always pressed the keyboard quickly cancelled the light show when powered on and a number of keys became dead since they shared either the row or column with the function key and thus effectively the keyboard only saw one button pressed (the function button). Yes it took me longer to find the fault than was perhaps necessary, but I at least did confirm the basics were working (the actual switches worked) which could have been at fault.

SMPS blows fuse immediately

Switch Mode Power Supplies (SMPS) are much more efficient than traditional linear power supplies, however, they are a lot more complicated and thus, there is greater chance for them to fail. I had obtained a small SMPS that generated +5V, +12V and -12V DC from 240VAC but it blew the 2A mains fuse as soon as I connected the power. I checked that the fuse had indeed blown, which it had, and then I looked to see if any components looked damaged before checking them with my multimeter. One electrolytic capacitor, which was on the +5V line looked ever so slightly out of shape so I desoldered it as I didn't have an ESR meter to check it.

When testing components that are part of a circuit it's very important to take into account any other components that are in parallel with the component under test; if need be you can desolder one of the component's connections. To help with getting a good view of how the circuit works and what components may affect the reading of others you can draw out the circuit, if it isn't too difficult. One way to help with working out the circuit is to take a photo of the front and back of the circuit board, transfer the photos to a computer and then draw over them to build up the connections and then to add the component symbols. When you have the circuit diagram you could rearrange it to make it easier to read and you can confirm connections where possible using a multimeter and reasoning.

Having not found any other problems with the power supply I replaced the fuse and turned the power on; the fuse didn't blow. It's very important to find the fault before connecting power again otherwise the fuse could blow once more and the circuit could be harmed. In my case it seemed the capacitor had been the fault but the circuit wasn't quite fixed yet. I measured the outputs and found the 5V line to be about 0.5V and the +12V and -12V outputs at about +18V/-18V. As the capacitors on the output lines were rated at 25V you can see why it's important to use capacitors that are rated at least twice as high as the maximum voltage, should there be a fault.

I soldered a new capacitor for the 5V output and after removing the capacitors on the +12V and -12V lines and putting in new ones the circuit was now working with 5.2V for the 5V output and about +11V and -11V for the +12V/-12V lines. Note that the capacitors appeared to be fine when I tested them out of the circuit using my capacitor meter and this is why an ESR (Equivalent Series Resistance) meter is much better at telling you whether a capacitor is faulty or not.

As a multimeter does not have much of an effect on a circuit it sometimes doesn't show the whole picture. To further test the power supply I connected a motor to the +5V output and saw that the voltage stayed at 5V while the motor was connected. I then measured the current, which was about 50mA for the motor's initial start-up current before settling down to about 19mA. This test showed that the 5V output was working correctly, at least under a small load.

As for the +12V/-12V I was disappointed to find that they struggled to power an LED via a limiting resistor; the voltage dropped to about +8V/-8V under the load. This result, coupled with being a volt below the nominal output voltage, suggested that there was still a fault on the +12V/-12V lines. However, this is where we must understand how a circuit works before jumping to the conclusion that there is a problem with it. Power supplies are sometimes designed to take advantage of a particular scenario such as always being under load, as this can simplify the circuit but for the tester can complicate things. By looking at the power supply circuit diagram I could see that it relied on the 5V line for feedback which in turn adjusted the +12V/-12V outputs. When I connected the motor to the 5V output and an LED via limiting resistor to the +12V output I found that the +12V output was about 9V. By connecting two motors to the 5V line, drawing around 50mA, and with the LED running off the +12V output, the +12V line was now showing 11V. Thus, the +12V output will only work correctly if under load and if the +5V output is under an appreciable load. The same works for the -12V as that too is dependent on the +5V output, however, both the +12V and -12V outputs will rise above 12V is they are not under load while the +5V line is.

All content of this and related pages is copyright (c) James S. 2015-2021