I’ve owned my Kurzweil K2500SWx since shortly after it was released in the summer of 1998. It’s always been my primary controller keyboard for the studio, usually sitting between the monitor and the computer keyboard tray. When I lived in Japan for two years, it was the only synth to come with me; the rest went into deep storage or were loaned to friends. I find its VAST architecture very flexible, and it’s still nice to have a hardware sampler, even if it’s limited to 16-bit 48kHz.
So it’s no surprise that it’s had a few mishaps over the years. And while Sweetwater graciously offers “lifetime free technical support,” there’s the hassle of shipping the unit across an international border, and the scarcity of parts to deal with.
This holiday weekend I fixed 4 nagging problems: a digital jitter, a failed front-panel button, failed aftertouch and a fading front-panel display. Here’s how, since it took me a while to research and maybe you’ll need to do it yourself someday.
I had been hearing some low-level LSB jitter noise coming out of the AES/EBU output. Apparently, replugging all the internal connectors to the K-SMP sampling board, tightining the screws holding it in place, and ensuring the grounding bracket made good contact with the aluminum rail was all that was necessary. I love easy fixes.
About 4 years ago a can of ginger ale I had sitting on the desk sprung a pinhole leak on the side. About 1⁄3 of the can ended up all over the K2500, spraying from C4 up to C6 and all over the buttons too. When it dried out, I lost my left arrow key. I’ve been surviving by cycling through options to get back to the left column!
While I had the unit apart, I de-soldered the “left” button using my favourite tool ever, the DEN-ON SC-7000Z. I tried soaking it in rubbing alcohol, then water, but it didn’t help much. Turned out filling a sewing thimble with contact cleaner, putting the button in, then “working” the button with a screwdriver a hundred times or so fixed it right up. Yay, I can navigate menus again!
When you disassemble the front panel to desolder a button, be careful. Popping out the plastic parts can break the little standoffs that hold them into the circuit board. If this happens, don’t worry about re-gluing them or sourcing replacements. Instead just assemble the front panel upside down, placing the button plastic in the hole first, then carefully laying the circuit board down over it.
Right after the dreaded ginger ale incident, I disassembled the keyboard and removed the keys to clean them. (Pro tip: Use a bamboo chopstick to remove and replace each key’s spring by sticking it in the top hole, pressing down to release the bottom spring catch, and pulling the top of the stick towards the front of the keyboard to free it.)
When reassembling, I pinched the aftertouch ribbon cables (here, in a K2661) between the key bed and the pitch/mod end block, breaking the connection inside the ribbon cable and destroying my aftertouch. I was ferociously upset at myself over it; every time I’d see the unit it would sit up and scream “YOU DID THIS TO ME!”
Capacitive aftertouch “Flat Flex Cables” are basically 4 layers: a top plastic layer, a “zipper” layer of extremely thin aluminum foil, a striped layer of low-conductivity dielectric, and a bottom plastic layer. In the K2500 Fatar key bed, this is topped off with a thick layer of felt, against which the key lands when pressed down. I initially tried repairing the break with some conductive paint, but as the cable flexed to be installed, the paint flaked off – not useful. I then tried brute-force soldering some wires to the foil-and-plastic assembly. This worked for about 5 minutes, then it too succumbed to flexing. Worse, each time I worked on the cable, it started delaminating, breaking aftertouch farther and farther up the keyboard.
After years of sulking and unhappiness, a friend suggested I should just attach a new connector. FCI’s 66226-004LF is exactly the right part. Trim the ribbon cable straight, insert into the connector, press hard on the metal bits to get them to puncture the ribbon cable, then fold down and lock the blue tab in place. To connect the ribbon to the wheel assembly, I made a short jumper cable out of an old Molex-to-floppy drive power adapter I had lying around. Because the aftertouch delaminated underneath the keys, I cut a notch in the felt for the new connector and its wires, and was sure when crimping it to position it so that the edges of the keys wouldn’t land on the connector itself. The connector is just small enough! On my keyboard, that placed the connector precisely underneath D1.
I then needed to calibrate the aftertouch sensitivity. Turn on the K2500X, let it fully boot, then press 4-5-6 all at the same time. The machine’s now in diagnostic mode. Play any white key on the keyboard and apply pressure. You will see the display show the pressure reading, which ranges from 0-255. Right underneath the pitch/mod wheel block there are 5 holes for sensitivity; the piano-keyboard side vertically adjacent two holes are white and black aftertouch sensitivity.
The K2000 display has an electroluminescent (EL) backlight, notorious for failing after as little as 5,000 hours of operation. The K2500 used a newer cold cathode fluorescent lamp (CCFL), the sort used in LCD panels up until a few years ago, but it too can have a short lifespan. There is a coarse CONTRAST adjustment knob inside the unit on the Audio Scanner board, but it’s a band-aid; eventually, the EL or CCFL gets too dim to be useful and your display is unreadable without very bright work light.
Fortunately, Newhaven Display makes the NHD-24064WG-ATMI-VZ# (specs here), a fully compatible display assembly with a white LED-powered backlight rated for at least 25,000 hours of continuous use. Here’s the new display with the ribbon cable from the original display soldered on:
As I had to replace the K2000 display unit, I decided to also replace the K2500X’s display and get some future peace of mind along the way. Two modifications to the K2500/X were required to make the replacement display work. First, the new display is a bit shorter and narrower. I cut a small adapter out of a piece of ABS plastic, to which I mounted the display before attaching the display in the machine. Depending on the thickness of your adapter, you may need standoffs to hold the display closer to the acrylic bezel. You could probably 3D print something similar.
The other mod is the backlight power supply, which needs +3.5V instead of the high voltage required for a CCFL. The good news is that a dedicated, regulated +5V was incorporated into the backlight board for the high voltage transformer, and you can remove the old high voltage transformer in the process, eliminating the high-pitched squealing that annoys your children, dogs, and adults still acting like children.
Given the ~80-100 mA draw of the backlight, you will need to use a ~15-33Ω voltage-limiting resistor between the 7805 regulator and the +3.5V input on the display. I measured my unit at 75mA, so used a 38.5Ω 1% resistor I had on hand. I desoldered the transformer, adding in a jumper wire between pin 2 of the 7805 and the old pin 2 of the HV transformer, then the resistor between pins 1 and 3 where the old transformer used to be. I then removed the molex connector and cable from the old display’s power connection, severed the jumper between pins 2 and 4 in the connector (preventing the floating ground output of the 7805 from being shorted to actual chassis ground, passing 10V instead of 5V), and connected pins 1 and 3 to the new display. When wiring the connector, pins 1 and 3 were conveniently pre-marked red and black on the connector itself, which helped me not make a mistake. Here’s a picture of the changes to the backlight board:
Removing the plastic bezel from the old display was a pain, the double-sided sticky tape they used is quite strong. Go slowly with an X-Acto knife and you’ll be fine. I nicked some of the black paint along the underside of the bezel, but it easily touched up with some black nail polish. I bet model airplane paint would work fine, too. For reassembly, I attached the display to the front panel, then pushed the bezel through the front; this seemed easier than trying to get the bezel stuck on the display in exactly the right place.
This was all good practice for my K2000R, suffering from the same issue. I’m planning on selling it in the next few months, and it only felt right to fix the problem first. When I get to that display’s replacement, I’ll post the results here.