When I run board level BSCAN testing, the Boundary Scan Length Test fails due to issues with the power supplies and the signal integrity of the JTAG lines.
What other problems can result with this test in the Spartan-3E 1200 and 1600 devices?
The Spartan-3E 1200 and 1600 devices have known issues with the Stepping 0 silicon in regards to the Boundary Scan Register. These know issues relate to internal timing aspects of the BSCAN circuitry. These issues have been fixed in Stepping 1 devices.
Symptoms include a failing Boundary Scan Register Length Test.
This has been resolved in the Stepping 1 devices and is further outlined in the Spartan-3E Data Sheet.
If you have devices not listed and are still having problems with the Boundary Scan Register Length Test please check the following:.
1 - Power Supplies: All power supplies must rise monotonically and within the specified minimum and maximum values in the Data Sheet. There should not be any sections of flat or negate slope in the rise of each of the supplies.
2 - JTAG Signal Integrity: Double clocking or glitching on the JTAG lines can cause problems with BSCAN testing. In particular, TCK needs to be scoped to ensure no double clocking is occurring. TMS and TDI should also be scoped for glitching which can cause the device to exit out of a shift state driving TDO High or place invalid data on the input to the tap controller.