I have a ~20 years old professional film photography scanner manufactured by Fujifilm. Fujifilm produced a few units of this scanner worldwide, and today they are completely discontinued.
Since few days ago I’m facing a critical issue with its high-speed USB 2.0 connection. Basically its USB device is not detected by the host.
KL5BUDV003 - The Kawasaki KL5KBUDV003 is a high performance device that transfers data between the USB2.0 high-speed BUS and the PCI 33MHz, 32 bit BUS. This 1 .
datasheetspdf.com
I have checked all DC voltages and clocks applied to this device and everything seems to be ok in the PCB.
Then, using an oscilloscope, I have sampled the D+ and D- lines, RPU_ENA and VBDET pins.
RPU_ENA: This pin is used to indicate a high-speed USB connection to the host. The host recognizes a new USB device only if this pin is 3V (pull-up to the D+ line)
VBDET: Pin indicating that a host is connected (based on the 5V USB power supply)
D+ and D- lines: USB data lines
And then I reach the following conclusions:
Turning scanner on with host/USB cable connected: VBDET is maintained at 3V, and when scanner is turned on RPU_ENA goes high (3V), but only during ~3 seconds. Then, it stays 0V forever. During the ~3 seconds high, an interrupt is detected in the host (Windows device manager refresh), but then no USB device is showing up. The following figures show the D+ D- and RPU_ENA signals during scanner boot up (they are not synchronized because I have only an oscilloscope port to sample the signals):
D+ line:
D- line:
RPU_ENA:
Plug out/in the USB cable with the scanner on: VBDET goes high (3V), but RPU_ENA is maintained always at 0V. No interrupt is detected in the host (so, no refresh is observed in the Windows device manager).
Are you using exactly the same USB cable, computer and port on that computer, as has worked OK in the past?
I have quite a few devices that will not work with any new USB cables, but work fine with the decades old cables I have.
Many new cables have conductors that are far too thin, to save on copper, and are too high resistance to work with devices that need low resistance connections.
Yes, the cable is the original.
I tried a different one also, but still without success. But I might try again with others just to check (maybe a shorter cable?).
Anyway, since when I plug out/in the USB cable with the scanner on, the RPU_ENA is maintained always at 0V (i.e. no interrupt is sent to the host), I believe this is indeed a hardware problem.
The only other thing I can think of at present that may be worth trying is to add an extra decoupling capacitor across the power to the IC, in case the existing ones now have high internal resistance.
A power glitch as it starts to connect could easily mess things up.
It appears it may have its own regulator, just above it in the photo? The number is too blurred to read, but it look like a reg?