SDK Errata

In the SDK package there is a Changelog.txt file that details all known additions, changes and bugs at the time of release. This page contains a list of SDK and TES Errata found after release, which will be updated as each bug is reported or resolved. 

Last Edited: 6th July 2021

FPGA source code not included

Bug:

FPGA source code was omitted from the SDK 17 release. 

Customer Impact:

Users can download a new release of SDK from the product page (version 17.1) that has the FPGA source code included.

This is fixed in SDK 17.1

MCS Operation not working 

Bug:

MCS operation in SDK version 17 did not function as expected and as a result was not useable. 

Customer Impact:

Users can download a new release of SDK from the product page (version 17.1) that has resolved this issue. 

This is fixed in SDK 17.1

Segmentation fault due to large memory allocations on the stack

Bug:

By default, ADI_FPGA9001_MAX_CAPTURE_BYTES is set to the absolute maximum supported by the FPGA. This may cause a segmentation fault due to large memory allocations on the stack.

Customer Impact:

This can be resolved by either:
Modify the makefile CFLAGS to include -DADI_FPGA9001_USE_DYNAMIC_MEMORY. This makes the ADI_FPGA9001_MAX_CAPTURE_BYTES value irrelevant, instead allocating memory on the heap.
Modify the value of ADI_FPGA9001_MAX_CAPTURE_BYTES to be smaller - ideally only as large as the largest required data capture/transmit

MCS clock is zero in generated TES sample Code:

Bug: 

Generated sample code doesn't work if the customer's configuration includes MCS.

Customer Impact:

The user will need to implement this fix in the generated code to get the MCS signal to work. The generated code from MCS enabled profile, for example LTE 61.44 MHz, will have the following code:

AnalogDevices.EvalClient.Fpga9001.fpga9001_McsCfg_t(6, 0, 8, fpga9001_McsEdge_e.FPGA9001_MCS_EDGE_RISING);

# Description of the parameters

• adi_fpga9001_McsCfg_t->mcsPeriod = 0, we set that to 1000000
• adi_fpga9001_McsCfg_t->mcsPulseWidth = 8, we set that to 2

Correct code

AnalogDevices.EvalClient.Fpga9001.fpga9001_McsCfg_t(6, 1000000, 2, fpga9001_McsEdge_e.FPGA9001_MCS_EDGE_RISING);

This is fixed in SDK 16

TES relative paths not saved:

Bug: 

When the user tries to upload a file, for example, a Hop Table or a Table Select Sequence, the file browser dialog won't know where to initialize if the desired path is relative. The file browser dialog will open to the /last/ location the customer selected to upload a file, not the intended location.

The GUI has default suggestions for where to upload files based on relative paths. To be clear, /all/ of our example files are located in relative paths. So that means all hop tables, table index by pin sequences, table select by pin, gain and attenuation tables and sequences, all of these have relative paths that are meant to guide the file browser to the right location.

Customer Impact:

The user will have to manually find the correct file locations for the different tables and example files needed for upload to the TES. When the SDK and TES are installed, these files can be found in the Examples directory of the TES installation. 

This is fixed in SDK 16