HUNT ENGINEERING HEL_UNPACK.LIB USER MANUAL
Contents
1. Unpack h HEL PackBin2hl6bitlchan HEL PackBin2hl6bit2chan HEL PackBin2hl6bit3chan HEL PackBin2hl6bit4chan HEL PackBin2hl6bitSchan HEL PackBin2hl6bit6chan HEL PackBin2hl6bit7chan HEL PackBin2hl6bit8chan Description short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel This group of packers take multiple buffers of short data and combine to a single output buffer The output buffer will be compatible with Hunt Engineering the 14 bit GD14 data format right justified packed in 16 bit words top bit used as control Data is limited by ANDing with bitmask before use This can be used to ensure that short data 1s in range for 14 bit GD14 compa2. include HEL Unpack h HEL Unpackl6bitlchan HEL Unpackl6bit2chan HEL Unpackl6bit3chan HEL Unpackl6bit4chan HEL Unpackl6bit8chan HEL Unpackl6bitl2chan HEL Unpackl6bitl6chan Description short input unsigned short bitmask short output int samples_per_channel short errors short input unsigned short bitmask short output int samples_per_channel short errors short input unsigned short bitmask short output int samples_per_channel short errors short input unsigned short bitmask short output int samples_per_channel short errors short input unsigned short bitmask short output int samples_per_channel short errors short input unsigned short bitmask short output int samples_per_channel short errors short input unsigned short bitmask short output int samples_per_channel short errors This group of unpackers take a buffer of data from a 12 bit source typically an ADC and split it into multiple output buffers one per channel The data source must be compatible with Hunt Engineering 12 bit ADCs right justified packed in 16 bit words top 4 bits used as status The top four bits are cleared as the data is unpacked This allows the data to be used directly All status information for an output buffer is logically ORed together and saved to the errors pointer allowing checks to be made for flags being set or ovetflow Parametets Input B
3. Exclusions If HUNT ENGINEERING decides that there is any evidence of electrical or mechanical abuse to the hardware then the customer shall have no recourse to HUNT ENGINEERING its agents In such circumstances HUNT ENGINEERING may at its discretion offer to repair the hardware and charge for that repair Limitations of Liability HUNT ENGINEERING makes no warranty as to the fitness of the product for any particular purpose In no event shall HUNT ENGINEERING S liability related to the product exceed the purchase fee actually paid by you for the product Neither HUNT ENGINEERING its suppliers shall in any event be liable for any indirect consequential or financial damages caused by the delivery use or performance of this product Because some states do not allow the exclusion or limitation of incidental or consequential damages or limitation on how long an implied warranty lasts the above limitations may not apply to you TECHNICAL SUPPORT Technical support for HUNT ENGINEERING products should first be obtained from the comprehensive Support section www hunteng co uk support index htm on the HUNT ENGINEERING web site This includes FAQs latest product software and documentation updates etc Or contact your local supplier if you are unsure of details please refer to www hunteng co uk for the list of current re sellers HUNT ENGINEERING technical support can contacted by emailing support hunteng demon co uk callin
4. int samples per channel short errors HEL UnpackSel bit2chan short input unsigned short bitmask short output int samples per channel short errors Description This group of unpackers take a buffer of data from a xx bit source typically a HERON IO and split it into multiple output buffers one per channel The xx stands for a number of 1 to 15 E g the functions handle both 12 and 14 bit data The data source must be compatible with Hunt Engineering xx bit ADCs right justified packed in 16 bit words top 16 xx bits used as status The top bits are cleared as the data is unpacked The data is then sign extended This means that an array of xx bit raw data samples is ready to be used as an array of signed short integers Example a raw data value of Oxffe will be sign extended to become Oxfffe a short integer value 2 A data value of 0x2 will not change and stay the same All status information for an output buffer is logically ORed together and saved to the errors pointer allowing checks to be made for flags being set or overflow Parameters Input Pointer to 16 bit input buffer containing interleaved samples from all active channels This buffer must be aligned to a 32 bit boundary Bitmask A 16 bit mask used to limit the data For 16 bits use Oxffff 14 bits use Ox3fff 12 bits use OxOfff Output Pointer to first output buffer Subsequent output buffers are located immediately after this one
5. into multiple output buffers one per channel The data is converted from offset binary to 2 s complement as it is unpacked The data source must be compatible with Hunt Engineering 16 bit ADCs right justified packed in 32 bit words top 16 bits used as channel markers amp status Only the sample data is saved to the output array Thus the output array is half the size in bytes of the input array Parametets Input Bitmask Output Samples per channel Errors pointer to 32 bit input buffer containing interleaved samples from all active channels This buffer must be aligned to a 32 bit boundary A 16 bit mask used to limit the data For 16 bits use Oxffff 14 bits use Ox3fff 12 bits use OxOfff pointer to first output buffer Subsequent output buffers are located immediately after this one This buffer must be aligned to a 32 bit boundary Do not use the input buffer as the output as this will cause data loss Number of samples per channel in the input buffer Pointer to an array of 16 bit error values one array entry per channel This array must be aligned to 32 bit boundary Some I O modules include additional info in flags along with the sample data This is ORed together for each channel and stored here HUNT ENGINEERING UnpackLib USER MANUAL HEL_UnpackSe16bitYchan Syntax include HEL Unpack h HEL UnpackSel bitlchan short input unsigned short bitmask short output
6. 16 HEL 32 17 PACKBin2h 1 reete tie ete ee n e Es 18 PERFORMANCE NOTES de vex sext Eee 19 SOURCE CODE p 21 TECHNICAL SUPPONT UN SEEN 22 3 HUNT ENGINEERING HEL UnpackLib USER MANUAL V1 0 V1 2 V1 3 V1 4 Library History Released 2274 October 2000 1 release of library contained unpackers for 12 amp 16 bit ADCs Released 8 December 2000 2nd release offered a more consistent interface with increased flexibility extended functionality and an improved manual Major additions were e Packers for 12 14 amp 16 bit DACs were introduced These should support all ADCs up to 16 bit for the foreseeable future e Calling conventions for unpackers were changed to allow the wordlength limiting mask to be passed as a parameter Changing the calling convention means that source code modifications are needed to upgrade to this version of the library However they extend the library s abilities to cater for additional wordlength ADCs that we may supply in the future Released 12 Feb 2001 This release fixed a problem when using masks that have the top but set All masks are now defined as unsigned to prevent this problem Released 20 April 2001 Added HEL UnpackSel6BitYchan functions to support HERON IO The functions unpack and sign extend a 12 bit dat
7. ERING HEL UnpackLib USER MANUAL Source Code The source of the library functions is provided on the HUNT ENGINEERING CD in the stc directory Simply follow Getting Started from the CD menu and select libraries This source can be copied onto your hard drive and modified or re built as you wish The library is implemented as a set of linear assembly files To build it run the batch file Build Obj Lib bat This batch file sets the assembler options and builds the library Note that the assembly process will give some warnings relating to large constants these are normal and not a concern The Supplied Library is built for Small memory model This is suitable for most applications Please refer to the HUNT ENGINEERING technical note Choosing a memory model for more details If it is necessary to have a different memory model version of the library it can be built using the sources supplied 21 HUNT ENGINEERING HEL_UnpackLib USER MANUAL Technical Support Technical support for HUNT ENGINEERING products should first be obtained from the comprehensive Support section www hunteng co uk support index htm on the HUNT ENGINEERING web site This includes FAQs latest product software and documentation updates etc Or contact your local supplier if you are unsure of details please refer to www hunteng co uk for the list of current re sellers HUNT ENGINEERING technical support can contacted by emailing s
8. HUNT ENGINEERING bee Chestnut Court Burton Row Brent Knoll Somerset 9 UK Tel 44 0 1278 760188 Fax 44 0 1278 760199 Email sales hunteng co uk http www hunteng co uk http www hunt dsp com HUNT ENGINEERING HEL UNPACK LIB USER MANUAL Unpacking data access library for Hunt Engineering ADC amp DAC modules Many multichannel devices such as ADCs multiplex channels of data onto a single HERON FIFO These channels need to be separated before processing or packed together before output The library described here provides efficient assembly coded routines to pack or unpack the data to from separate arrays and optionally to convert it to a different format Software Rev 1 4 Document Rev 1 4 R Weir 14 11 01 This documentation and the product it is supplied with are Copyright HUNT ENGINEERING 2001 All rights reserved HUNT ENGINEERING maintains a policy of continual product development and hence reserves the right to change product specification without prior warning WARRANTIES LIABILITY and INDEMNITIES HUNT ENGINEERING warrants the hardware to be free from defects in the material and workmanship for 12 months from the date of purchase Product returned under the terms of the warranty must be returned carriage paid to the main offices of HUNT ENGINEERING situated at BRENT KNOLL Somerset UK the product will be repaired ot replaced at the discretion of HUNT ENGINEERING
9. This buffer must be aligned to a 32 bit boundary Do not use the input buffer as the output as this will cause data loss Samples_per_channel Number of samples per channel in the input buffer Errors Pointer to an array of 16 bit error values one array entry per channel This array must be aligned to a 32 bit boundary Some I O modules include additional info in flags along with the sample data This is ORed together for each channel and stored here 12 HUNT ENGINEERING HEL UnpackLib USER MANUAL DAC Packers Data Format Conversion There ate two sets of packers covering 16 bit and 32 bit data sinks The 32 bit packers assume that a single sample is contained in one half of a 32 bit word as in the case with the 16 bit DACs while the 16 bit packers assume that the sample is contained within a 16 bit word In either case the library can pack different wordlengths for example the 16 bit packers can be used with 12 or 14 bit DACs The packers are split into two groups The first group pack the data directly performing no format conversion The second group converts the data to offset binary from 2 s complement Offset Binary is a common format used on many ADCs amp DACs Data Buffer Formats Input data is in linear addressed buffers of 16 bit samples short data It is assumed that the input buffers are consecutive see example The output buffer must be large enough to receive the results note that f
10. a value Added HEL PackBin2h16BitYchan functions to support GD14 data formats It accepts an array of signed shorts and packs them into proper GD14 data format Released 14 November 2001 HEL UnpackSel6BitYchan functions now support any data length not only 12 bit data The functions unpack and sign extend any xx bit data value with xx in 15 1 HUNT ENGINEERING HEL_UnpackLib USER MANUAL Introduction In many cases I O modules used with HERON systems pack their data into a single stream This allows multi channel modules to operate with a single FIFO It is the only way that a GDIO module can implement multiple channels the GDIO specification includes only one channel and it is a very practical way of ensuring data remains synchronised through the system However this data format is not always ideal for processing The incoming data stream may include tag or address information and the multiplex within the stream may make it difficult to access consecutive points with signal processing algorithms It is usually most appropriate to unpack the data into separate arrays For output devices the data may include control information such as the last channel bit ot channel address This must be added before the data is multiplexed for output This library implements a suite of routines which implement unpacking packing for the data types used in Hunt Engineering I O modules In addition the data s format may be conv
11. ce typically an and split it into multiple output buffers per channel The data source must be compatible with Hunt Engineering 16 bit ADCs right justified packed in 32 bit words top 16 bits used as channel markers amp status Only the sample data is saved to the output array Thus the output array is half the size in bytes of the input array Parametets Input Bitmask Output Samples per channel Errors pointer to 32 bit input buffer containing interleaved samples from all active channels This buffer must be aligned to a 32 bit boundary 16 bit mask used to limit the data For 16 bits use 14 bits use Ox3fff 12 bits use OxOfff pointer to first output buffer Subsequent output buffers are located immediately after this one This buffer must be aligned to a 32 bit boundary Do not use the input buffer as the output as this will cause data loss Number of samples per channel in the input buffer Pointer to an array of 16 bit error values one array entry per channel This array must be aligned to a 32 bit boundary Some I O modules include additional info in flags along with the sample data This is ORed together for each channel and stored here HUNT ENGINEERING HEL UnpackLib USER MANUAL HEL_UnpackBin16bitYchan Syntax include HEL Unpack h HEL UnpackBinl6bitl chan short input unsigned short bitmask short output int samples per channel short errors HEL U
12. erted to 2 s complement if required This is performed as it is unpacked with almost no time penalty 5 HUNT ENGINEERING HEL_UnpackLib USER MANUAL Where do find the library When you have made a normal software installation from the HUNT ENGINEERING CD the library and header file for HEL_Unpack have been placed in the heron_unpack directory under your API installation directory YHEAPI heron_unpack The Create New HERON Project plug in for Code Composer Studio will automatically include this library into a new project and will add the directory to the include path This means that the functions in this library can be used in any project without concern 6 HUNT ENGINEERING HEL_UnpackLib USER MANUAL Functions ADC Unpackers Data Format Conversion There ate two sets of unpackers covering 16 bit and 32 bit data sources The 32 bit unpackers assume that a single sample is contained in one half of a 32 bit word as in the case with the 16 bit ADCs while the 16 bit unpackers assume that the sample is contained within some a 16 bit word In either case the library can extract different wordlengths for example the 16 bit unpackers can extract 12 or 14 bit samples The unpackers are split into two groups The first group unpack the data directly and perform no format conversion The second group convert the data from offset binary to 25 complement Offset Binary is a common format used on many ADCs a
13. for 12 bit DACs Pointer to 32 bit output buffer Do not use the input buffer as the output as this will cause data loss Number of samples per channel in the input buffer 15 HUNT ENGINEERING HEL_UnpackLib USER MANUAL HEL_PackBin16bitYchan Syntax include HEL _Unpack h HEL PackBinl6bitl chan HEL PackBinl6bit2chan HEL PackBinl6bit3chan HEL PackBinl6bit4chan HEL PackBinl6bit5chan HEL PackBinl6bit6chan HEL PackBinl6bit7chan HEL PackBinl6bit8chan Description short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel short input unsigned short masks unsigned short bitmask short output int samples per channel This group of packers take multiple buffers of short data and combine to a single output buffer converting to offset binary in the pr
14. g the direct support telephone number 44 0 1278 760775 or by calling the general number 44 0 1278 760188 and choosing the technical support option 2 HUNT ENGINEERING HEL_UnpackLib USER MANUAL TABLE OF CONTENTS LIBRARY HISTORY 4 INTRODUCTION c m 5 WHERE DO I FIND THE LIBRARY 6 FUNCTIONS ent C 7 ADC UNPAGCKERS em ded 7 Data Format Conversion 7 Data Bufler Formats 7 Assumptions Noles secte ence ette epe cote tite ice REA 7 A ligument of se b e ettet ee e e et e ERE Ra E RE e COS 7 Unpackl6bitYchndn i aiii seit 8 HEL Unpacki2DItYChOh ried EE ete metn te euet 9 UnpackBin6bitYchan 10 HELD Unpack Bin32 Dit V Chan io RM Me 11 HBE UnpackSe ODitLY Chant e Ae ed A e EE rede 12 DAG PACKERS dete He edet o e 13 Data Format CONVERSION 13 Buffer Formats IERI 13 ZASSUmptiOons d NOL S oit 13 PACKIODILY NGM i 14 Pack32bitYchan ciue eti e a E RR NR A SOE Ae te ad 15 HEE Packbinl6obitYcln x n ot
15. igned int masks unsigned short bitmask short output int samples per channel HEL PackBinl6bit4chan short input unsigned int masks unsigned short Description bitmask short output int samples per channel This group of packers take multiple buffers of short data and combine to a single 32 bit output buffer for 16 bit DACs It also converts the data to offset binary The output buffer will be compatible with Hunt Engineering 16 bit DACs right justified packed in 32 bit words top bits used as control Data is limited by ANDing with bitmask before use This can be used to ensure that short data is in range for a 16 bit DAC Data is ORed with the masks array before output This array contains a mask per input buffer so can be used to set control bits or channel address bits Parameters Input Masks Bitmask Output Samples per channel Pointer to first input buffer Subsequent input buffers are located immediately after this one This buffer must be aligned to a 32 bit boundary Array of 32 bit OR masks one per channel Used to set control bits 16 bit mask ANDed with data before output use Ox3fff for 14 bit DACs OxOfff for 12 bit DACs Pointer to 32 bit output buffer Do not use the input buffer as the output as this will cause data loss Number of samples per channel in the input buffer 17 HUNT ENGINEERING HEL_UnpackLib USER MANUAL HEL_PackBin2h16bitYchan Syntax include
16. itmask Output Samples per channel Errors pointer to 16 bit input buffer containing interleaved samples from all active channels This buffer must be aligned to a 32 bit boundary A 16 bit mask used to limit the data For 16 bits use 14 bits use Ox3fff 12 bits use OxOfff pointer to first output buffer Subsequent output buffers are located immediately after this one This buffer must be aligned to a 32 bit boundary Do not use the input buffer as the output as this will cause data loss Number of samples per channel in the input buffer Pointer to an array of 16 bit error values one array entry per channel This array must be aligned to a 32 bit boundary Some I O modules include additional info in flags along with the sample data This is ORed together for each channel and stored here HUNT ENGINEERING HEL UnpackLib USER MANUAL HEL_Unpack32bitYchan Syntax include HEL Unpack h HEL Unpack32bitl chan int input unsigned short bitmask short output int samples per channel short errors HEL Unpack32bit2chan nt input unsigned short bitmask short output int samples per channel short errors HEL Unpack32bit4chan int input unsigned short bitmask short output int samples per channel short errors HEL Unpack32bit8chan nt input unsigned short bitmask short output int samples per channel short errors Description This group of unpackets take a buffer of data from a 16 bit sour
17. mp DACs Data Buffer Formats Output data is in consecutive linear addressed buffers It is assumed that the output buffers are consecutive see example The output data is 16 bit Buffers allocated must be large enough to receive the results Assumptions amp Notes 1 All code is interruptible to ensure maximum performance in HERON systems 2 It is assumed that the first sample in the input buffer is from Channel 0 and that all subsequent samples are from consecutive channels This is the normal startup sequence for HERON IO systems 3 The unpacker functions will not detect a FIFO overflow If FIFO overflow occurs synchronisation will be lost and must be regained externally 4 In all cases it is assumed that each input buffer will contain at least 8 samples and that each output buffer will contain at least 2 samples 5 It is also a condition that all buffers should contain an even number of samples and be aligned to 32 bit boundaries This allows the library to treat data as 32 bit values allowing faster data transfers Alignment of arrays Many of the functions use arrays of shorts that must be aligned to 32 bit boundaries This is mentioned in the tables below along with This means that the data must be forced to be aligned as follows allocate the errors array pragma DATA ALIGN errors 4 short errors NO_OF CHANNELS 7 HUNT ENGINEERING HEL_UnpackLib USER MANUAL HEL_Unpack16bitYchan Syntax
18. npackBinl6bit2chan short input unsigned short bitmask short output int samples per channel short errors HEL UnpackBinl6bit3chan short input unsigned short bitmask short output int samples per channel short errors HEL UnpackBinl6bit4chan short input unsigned short bitmask short output int samples per channel short errors HEL UnpackBinl6bit8chan short input unsigned short bitmask short output int samples per channel short errors HEL UnpackBinl6bitl12chan short input unsigned short bitmask short output int samples per channel short errors HEL 16611 short input unsigned short bitmask short output int samples_per_channel short errors Description This group of unpackers take a buffer of data from a 12 bit source typically an ADC and split it into multiple output buffers one per channel The data is converted from offset binary to 2 s complement as it is unpacked The data source must be compatible with Hunt Engineering 12 bit ADCs right justified packed in 16 bit words top 4 bits used as status The top four bits are cleared as the data is unpacked This allows the data to be used directly All status information for an output buffer is logically ORed together and saved to the errors pointer allowing checks to be made for flags being set or ovetflow Parametets Input Bitmask Output Samples per channel Errors pointer to 16 bit inp
19. ocess The output buffer will be compatible with Hunt Engineering 14 bit or 12 bit DACs right justified packed in 16 bit words top bits used as control Data is limited by ANDing with bitmask before use This can be used to ensure that short data is in range for a 12 bit or 14 bit DAC Data is ORed with the masks array before output This array contains a mask per input buffer so can be used to set control bits or channel address bits Parameters Input Masks Bitmask Output Samples per channel Pointer to first input buffer Subsequent input buffers are located immediately after this one This buffer must be aligned to a 32 bit boundary Array of 16 bit OR masks one per channel Used to set control bits 16 bit mask ANDed with data before output use Ox3fff for 14 bit DACs OxOfff for 12 bit DACs Pointer to 16 bit output buffer This buffer must be aligned to a 32 bit boundary Do not use the input buffer as the output as this will cause data loss Number of samples per channel in the input buffer 16 HUNT ENGINEERING HEL UnpackLib USER MANUAL HEL PackBin32bitYchan Syntax include Unpack h HEL PackBinl6bitl chan short input unsigned int masks unsigned short bitmask short output int samples per channel HEL PackBinl6bit2chan short input unsigned int masks unsigned short bitmask short output int samples per channel HEL PackBinl6bit3chan short input uns
20. or some packers the output data is 32 bit hence the output buffer will be twice the size of the input buffer Assumptions amp Notes 1 All code is interruptible to ensure maximum performance in HERON systems 2 In all cases it is assumed that each input buffer will contain at least 2 samples and that each output buffer will contain at least 8 samples 3 It is also a condition that all buffers should contain an even number samples and be aligned to 32 bit boundaries This allows the library to treat data as 32 bit values allowing faster data transfers 13 HUNT ENGINEERING HEL UnpackLib USER MANUAL HEL Pack16bitYchan Syntax include Unpack h HEL Packl6bitlchan HEL Packl6bit2chan HEL Packl6bit3chan HEL Packl6bit4chan HEL Packl6bit5chan HEL Pack1 6bit6chan HEL Packl6bit7chan HEL Packl6bit8chan Description short input unsigned short masks unsigned short bitmask short output int samples_per_channel short input unsigned short masks unsigned short bitmask short output int samples_per_channel short input unsigned short masks unsigned short bitmask short output int samples_per_channel short input unsigned short masks unsigned short bitmask short output int samples_per_channel short input unsigned short masks unsigned short bitmask short output int samples_per_channel short input unsigned short masks unsigned short bitmask short out
21. put int samples_per_channel short input unsigned short masks unsigned short bitmask short output int samples_per_channel short input unsigned short masks unsigned short bitmask short output int samples_per_channel This group of packers take multiple buffers of short data and combine to a single output buffer The output buffer will be compatible with Hunt Engineering 14 bit or 12 bit DACs right justified packed in 16 bit words top bits used as control Data is limited by ANDing with bitmask before use This can be used to ensure that short data is in range for a 12 bit or 14 bit DAC Data is ORed with the masks array before output This array contains a mask per input buffer so can be used to set control bits or channel address bits Parameters Input Masks Bitmask Output Samples per channel Pointer to first input buffer Subsequent input buffers are located immediately after this one This buffer must be aligned to a 32 bit boundary Array of 16 bit OR masks one per channel Used to set control bits 16 bit mask ANDed with data before output use Ox3fff for 14 bit DACs OxOfff for 12 bit DACs Pointer to 16 bit output buffer This buffer must be aligned to a 32 bit boundary Do not use the input buffer as the output as this will cause data loss Number of samples per channel in the input buffer 14 HUNT ENGINEERING UnpackLib USER MANUAL HEL_Pack32bitYchan S
22. the number of samples per channel HEL_Unpack16bitYchan Test 1 Channel 36 N 1 0 2 Channel 52 0 8 3 Channel 40 N 1 0 4 Channel 52 N 0 9 8 Channel 76 N 0 8 12 Channel 72 N 1 3 16 Channel 76 N 1 3 HEL_UnPack32bitY chan Test 1 Channel 28 N 1 0 2 Channel 24 N 1 0 4 Channel 52 1 0 8 Channel 52 N 1 0 HEL UnpackBin16bitYchan Test 1 Channel 44 N 0 8 2 Channel 44 N 0 9 3 Channel 44 N 1 0 4 Channel 52 N 0 8 8 Channel 72 N 1 0 12 Channel 68 N 1 2 16 Channel 76 N 1 2 HEL_UnPackBin32bitYchan Test 4 Channel 20 N 1 0 8 Channel 56 N 1 0 19 HUNT ENGINEERING HEL_UnpackLib USER MANUAL HEL Pack16bitYchan Test 1 Channel 44 N 0 6 2 Channel 52 N 0 9 3 Channel 56 1 0 4 Channel 68 N 0 8 5 Channel 72 N 0 8 6 Channel 88 N 0 7 7 Channel 104 N 0 8 8 Channel 112 N 0 8 HEL Pack232bitYchan Test 1 Channel 44 TN 1 1 2 Channel 40 N 1 0 3 Channel 48 N 1 3 4 Channel 60 N 1 4 HEL_PackBin1 6bitYchan Test 1 Channel 60 0 5 2 Channel 72 0 8 3 Channel 68 N 1 1 4 Channel 24 1 3 5 Channel 84 N 0 8 6 Channel 88 N 0 8 7 Channel 108 N 0 8 8 Channel 120 N 0 8 HEL PackBin22bitYchan Test 1 Channel 60 N 1 3 2 Channel 60 N 1 0 3 Channel 52 1 3 4 Channel 140 N 1 3 HEL_UnpackSe16bitYchan Test 1 Channel 40 N 1 1 2 Channel 56 N 1 1 20 HUNT ENGINE
23. tible data Data is ORed with the masks array before output This array contains a mask per input buffer so can be used to set the GD14 data s control bit Parameters Input Masks Bitmask Output Samples per channel Pointer to first input buffer Subsequent input buffers are located immediately after this one This buffer must be aligned to a 32 bit boundary Array of 16 bit OR masks one per channel Used to set control bits 16 bit mask ANDed with data before output use Ox3fff for the 14 bit GD14 Pointer to 16 bit output buffer This buffer must be aligned to a 32 bit boundary Do not use the input buffer as the output as this will cause data loss Number of samples per channel in the input buffer 18 HUNT ENGINEERING HEL UnpackLib USER MANUAL Performance Notes Each of the packers unpackers is implemented in optimized linear assembly code Best results will be achieved using the unpackers on data in on chip RAM ideally in different memory blocks It is impossible to code a packer unpacker so that it does not cause memory bank contention We have tried to reduce this as far as possible but it is the most significant factor influencing performance Wherever possible arrange for the input and output buffers to be in separate memory banks The following figurs were generated using a TMS320C6201 with data in THE SAME BANK of on chip memory These should therefore be pessimistic numbers Note N is
24. upport hunteng demon co uk calling the direct support telephone number 44 0 1278 760775 or by calling the general number 44 0 1278 760188 and choosing the technical support option If you are in North America South America or Canada contact our strategic partner Traquair Data Systems at www traquair com company support html for support information and contact details 22 HUNT ENGINEERING HEL_UnpackLib USER MANUAL
25. ut buffer containing interleaved samples from all active channels This buffer must be aligned to a 32 bit boundary A 16 bit mask used to limit the data For 16 bits use Oxffff 14 bits use Ox3fff 12 bits use OxOfff pointer to first output buffer Subsequent output buffers are located immediately after this one This buffer must be aligned to a 32 bit boundary Do not use the input buffer as the output as this will cause data loss Number of samples per channel in the input buffer Pointer to an array of 16 bit error values one array entry per channel This array must be aligned to a 32 bit boundary Some I O modules include additional info in flags along with the sample data This is ORed together for each channel and stored here 10 HUNT ENGINEERING HEL UnpackLib USER MANUAL HEL_UnpackBin32bitYchan Syntax include HEL UnpackBin32bitl chan HEL UnpackBin32bit2chan HEL UnpackBin32bit4chan HEL UnpackBin32bit8chan Description HEL Unpack h int input unsigned short bitmask short output int samples per channel short errors int input unsigned short bitmask short output int samples per channel short errors int input unsigned short bitmask short output int samples per channel short errors int input unsigned short bitmask short output int samples per channel short errors This group of unpackers take a buffer of data from a 16 bit source typically an ADC and split it
26. yntax include HEL_Unpack h HEL Packl6bitlchan HEL Packl6bit2chan HEL Packl6bit3chan HEL Packl6bit4chan Description short input unsigned int masks unsigned short bitmask short output int samples_per_channel short input unsigned int masks unsigned short bitmask short output int samples per channel short input unsigned int masks unsigned short bitmask short output int samples per channel short input unsigned int masks unsigned short bitmask short output int samples per channel This group of packers take multiple buffers of short data and combine to a single 32 bit output buffer for 16 bit DACs The output buffer will be compatible with Hunt Engineering 16 bit DACs right justified packed in 32 bit words top bits used as control Data is limited by ANDing with bitmask before use This can be used to ensure that short data is in range for a 16 bit DAC Data is ORed with the masks array before output This array contains a mask per input buffer so can be used to set control bits or channel address bits Parameters Input Masks Bitmask Output Samples per channel Pointer to first input buffer Subsequent input buffers are located immediately after this one This buffer must be aligned to a 32 bit boundary Array of 32 bit OR masks one per channel Used to set control bits 16 bit mask ANDed with data before output use Ox3fff for 14 bit DACs OxOfff
Download Pdf Manuals
Related Search