Clock Card firmware

Related topics:

► Clock Card

► Firmware

▼ Clock Card Firmware‎ (17 P)

  Clock Card firmware

  Arbitrary Waveform Generator

  High rate acquisition

  How to Configure Which Readout Cards Report Data

  How to Read a Column at the Frame Rate

  Internal Commands

  MCE backplane protocol

  MCE commands

  MCE fibre protocol

  MCE flat-file format

  MCE temperature

  MCE Timing Diagram

  MCE Timing Diagram with older firmware

  Pre-v5 firmware

  Ramp Generator

  Recommended firmware versions

  Rectangle Mode Data

• Pre-v5 firmware

Firmware links

• Firmware Programming Files
• Clock Card firmware revisions may implement different data packet header formats. All of the different formats are documented at MCE flat-file format

Firmware Revision Listing

Revision 6.0.2

• Filename: cc_v06000002_20160530.sof

• Features: support of upper 32 word needs an additional wait cycle for ack_read
• Features:fpga_clr and critical_error_rst commands added
• Note: skip 6.0.1 due to losing track of all_cards
• bug: When stop command is issued, two frames are sent with both last-frame-bit and stop-bit set. This is currently benign as mas ignores the extra frame, but needs to be fixed regardless.
• bug: When data-timing-err bit is set, the last-frame-bit gets set but then the ret_dat counter doesn't not get reset, so when sync-dv starts coming in, clock card continues to send out frames again.

Revision 5.0.e (stable)

• Filename: cc_v0500000e_15may2012.sof

• Features: added pcb_rev interface to be able to read pcb revision starting RevC Clock Card.
• bugfix: awg is now realigned with data-acquisition and the phase can be adjusted using ramp_step_phase.
• bugfix: A time-delay is added between assertion of epc_sel and config_n to remedy the occasional failure of config_app and config_fac commands.
• bug: When stop command is issued, two frames are sent with both last-frame-bit and stop-bit set. This is currently benign as mas ignores the extra frame, but needs to be fixed regardless.
• bug: When data-timing-err bit is set, the last-frame-bit gets set but then the ret_dat counter doesn't not get reset, so when sync-dv starts coming in, clock card continues to send out frames again.

Revision 5.0.d (test)

• Filename: cc_v0500000d_10may2012.sof

• Features: Internal commands are issued at address-return-to-zero (ARZ) to provide more deterministic timing. This means the data_rate has to be at least 2, or internal commands will stop with a data_rate of 1. In previous versions, internal commands were being issued asynchronous to ARZ. The phase of the ramp or awg can be adjusted using ramp_step_phase.

• bugfix: stop cmd works now and sets the stop-bit and last-frame-bit properly in the frame status word of the last frame.
• bugfix: the ramp/awg value reported in the header is the one applied last. Previous versions reported what was going to be applied in the next iteration. If the command was not issued due to collision, then the value in the header was still being updated which was a bug.

• bug: awg mode doesn't work properly.

Revision 5.0.c (test)

• Filename: cc_v0500000c_26mar2012.sof

• Features: frame header revision 7 with dv_pulse_fibre_i encoded as bit 9 of the frame-status word in the header.
  • Based on 5.0.a

• Bugs: stop cmd may be broken?!It works but stop bit and last-frame bits are not set and a fakestop needs to be issued in mas.

Revision 5.0.b (test)

• Filename: cc_v0500000b_06feb2012.sof

• Features: based on 5.0.a, frame header revision 7 with dv_pulse_fibre_i encoded as bit 9 of the frame-status word in the header.

• Bugs: stop cmd may be broken?!It works but stop bit is not set and a fakestop needs to be issued in mas.

Revision 5.0.a (test)

• Filename: cc_v0500000a_06feb2012.sof

• Features: Based on 5.0.9

• Bugfix: last-frame-bit was not being set in 5.0.9 and it returns right number of frames!

• Bugs: stop cmd may be broken. It works but stop bit is not set.

Revision 5.0.9 (test)

• Filename: cc_v05000009_26jan2012.sof

• Features:
  • added ramp_step_phase command (par_id=xBB) to adjust the phase of Ramp Generator or Arbitrary Waveform Generator relative to data acquisition.
  • Based on 5.0.7

• Bugfix:
  • Internal housekeeping and internal ramp/awg commands can be run as tight as possible (data_rate=2 and ramp_step_period=2) alongside data acquisition without affecting data timing. Previous versions would hold the ramp-value refresh till data was collected first. In this version when there is a collision between ret_dat and ramp-command, the ramp value is still increased, so the next time it is due, it would apply the right value.

• Bugs:
  • config_app command is broken, it works in 5.0.3, but not in this version. to be investigated.
  • command issue/reply translator blocks are overhauled, so watch out for bugs related to stop command. None discovered yet though.

Revision 5.0.8 (not issued)

• Filename: cc_v05000008_12jul2011.sof

• Features:
  • This tag was a place holder for the final version of Clock Cards with Ethernet support (CCwE). This tag currently references unfinished code in CVS. The Quartus project file in this tag is designed for the Altera Stratix I Development Board.
  • Header Version 6
  • Based on 5.0.7

Revision 5.0.7 (stable)

• Filename: cc_v05000007_14may2010.sof

• Features:
  • Header Version 6
  • Based on 5.0.6
  • Compatible with MCE Jam Player -- SVN revision 16 (~/jp_25/mce_jam/trunk)
  • Simulation compatible with firmware v5.0.0 for the other cards.
  • Fixed a bug that caused a Clock Card with version 5.x.x firmware installed to return stale data for cards that had version 4.x.x. firmware installed on them.
  • Fixed a bug that caused a Clock Card to return stale data if a card was not present, or not configured. Now the Clock Card returns 0x00000000.

• Bugs:
  • internal_ramp and internal_awg only refresh on readout-frame rate (data_rate). So effective ramp_step_period is data_rate.
  • When both internal and data commands are scheduled to be issued on multiplexing frame N, then at ARZ, data command is issued first, followed by the internal command. This may cause problems when the internal command is not completed (too many words and row_len x num_rows not long enough) before next ARZ.
  • config_app command is broken (sometimes fails), it works in 5.0.3, but not in this version.

• FPGA Resource Usage (clk_card.fit.rpt):

; Total logic elements     ; 18,745 / 32,470 ( 58 % )                      ;
; Total pins               ; 243 / 598 ( 41 % )                            ;
; Total virtual pins       ; 0                                             ;
; Total memory bits        ; 957,952 / 3,317,184 ( 29 % )                  ;
; DSP block 9-bit elements ; 10 / 96 ( 10 % )                              ;
; Total PLLs               ; 2 / 6 ( 33 % )                                ;
; Total DLLs               ; 0 / 2 ( 0 % )                                 ;

• Timing Analyzer Summary (clk_card.tan.rpt):

; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk1' ; 0.279 ns  ;
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk0' ; 0.437 ns  ;
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk2' ; 3.224 ns  ;

Revision 5.0.6 (buggy)

• Filename: cc_v05000006_21apr2010.sof

• Features:
  • Header Version 6
  • Based on 5.0.5
  • Fixed a bug that prevented the Clock Card from loading firmware from its Factory Configuration Device when sw1:p1 is set to open (to enable remote configuration).
  • This version of firmware works in conjunction with modified Jam Player code that was recently committed to SVN tag: ~/jp_25/mce_jam/trunk : revision 16.

• Details:
  • To learn how to use remote configuration: Remote Firmware Update.
  • For .jam file conversions, see: MCE Programming File Conversions.
  • Removed crc_error functionality for now. It will get added back in when it is working. It was found to conflict with the Remote Configuration functionality by preventing the Clock Card from configuring from its Factory Configuration Device.

• Bugs:
  • Has a bug that causes the Clock Card with version 5.x.x firmware installed to return stale data for cards that have version 4.x.x. firmware installed on them.
  • Has a bug that causes the Clock Card to return stale data if a card is not present, or not configured.

• FPGA Resource Usage (clk_card.fit.rpt):

; Total logic elements     ; 18,652 / 32,470 ( 57 % )                      ;
; Total pins               ; 243 / 598 ( 41 % )                            ;
; Total virtual pins       ; 0                                             ;
; Total memory bits        ; 957,952 / 3,317,184 ( 29 % )                  ;
; DSP block 9-bit elements ; 10 / 96 ( 10 % )                              ;
; Total PLLs               ; 2 / 6 ( 33 % )                                ;
; Total DLLs               ; 0 / 2 ( 0 % )                                 ;

• Timing Analyzer Summary (clk_card.tan.rpt):

; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk0' ; 0.590 ns 
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk1' ; 1.683 ns  
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk2' ; 3.277 ns

Revision 5.0.5 (buggy)

• Filename: cc_v05000005_05mar2010.sof

• Features:
  • Header Version 6
  • Based on 5.0.4
  • Implemented unpacking logic for TMS and TDI signals, and inferring logic for TCK. This is the solution to the JTAG packing problem.
  • This version of firmware works in conjunction with modified Jam Player code that was recently committed to SVN tag: ~/jp_25/mce_jam/trunk : revision 8.

• Details:
  • Added the following commands: upload_fw, config_jtag, tdo_sample_dly, tck_half_period
  • This firmware solves the TMS and TDI packing problem such that a "wb cc upload_fw" command contains the following:
    • Word 0: total number of valid bits contained in words 1-n
    • Word 1-n: (tms,tdi) pairs starting from word 1 (bits 1,0), word 1 (bits 3,2), etc.
  • The TDO packing is done differently: the tdo bits are captured by a shift register, and shifted from LSB to MSB, up to a maximum of 16 TDO bits per 32-bit fibre word.

• Bugs:
  • Does not configure from its Factory Configuration Device upon power-up
  • (non-critical) when a card does not return a valid reply, the clock card does not send an error reply back to the PC. Instead, it fills the data with whatever it had in its buffer from previous command and sends it to mas.

• FPGA Resource Usage (clk_card.fit.rpt):

; Total logic elements     ; 18,699 / 32,470 ( 58 % )                      ;
; Total pins               ; 261 / 598 ( 44 % )                            ;
; Total virtual pins       ; 0                                             ;
; Total memory bits        ; 957,952 / 3,317,184 ( 29 % )                  ;
; DSP block 9-bit elements ; 10 / 96 ( 10 % )                              ;
; Total PLLs               ; 2 / 6 ( 33 % )                                ;
; Total DLLs               ; 0 / 2 ( 0 % )                                 ;

• Timing Analyzer Summary (clk_card.tan.rpt):
  • Note that the clock slack on clk0 has diminished significantly over the past few revisions. However, on this version, it increased again to a reasonable level.

; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk1' ; 1.164 ns  ; 
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk0' ; 1.276 ns  ; 
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk2' ; 3.299 ns  ;

Revision 5.0.4 (buggy)

• Filename: cc_v05000004_26feb2010.sof

• Features:
  • Header Version 6
  • Based on 5.0.3
  • Added JTAG control registers that emulated a parallel port to allow Jam Player software to write to the MCE from a MAS PC and configure devices via JTAG.
    • JTAG0 -- Output data
    • JTAG1 -- Input data
    • JTAG2 -- JTAG Chain control

• Details:
  • This version of firmware is compatible with ported JAM Player software that has been temporarily committed to CVS under \\mce\cards\clk_card\config_fpga\source\unix_code.

• Bugs:
  • Does not configure from its Factory Configuration Device upon power-up
  • (non-critical) when a card does not return a valid reply, the clock card does not send an error reply back to the PC. Instead, it fills the data with whatever it had in its buffer from previous command and sends it to mas.

• FPGA Resource Usage (clk_card.fit.rpt):

; Total logic elements     ; 17,827 / 32,470 ( 55 % )                      ;
; Total pins               ; 261 / 598 ( 44 % )                            ;
; Total virtual pins       ; 0                                             ;
; Total memory bits        ; 949,760 / 3,317,184 ( 29 % )                  ;
; DSP block 9-bit elements ; 10 / 96 ( 10 % )                              ;
; Total PLLs               ; 2 / 6 ( 33 % )                                ;
; Total DLLs               ; 0 / 2 ( 0 % )                                 ;
; M512s                    ; 66 / 295 ( 22 % )          
; M4Ks                     ; 171 / 171 ( 100 % )        
; M-RAMs                   ; 3 / 4 ( 75 % )              

• Timing Analyzer Summary (clk_card.tan.rpt):

; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk0' ; 0.771 ns  ;
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk1' ; 1.576 ns  ;
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk2' ; 2.859 ns  ;

Revision 5.0.3 (tested)

• Filename: cc_v05000003_13jan2010.sof

• Features:
  • Header Version 6
  • Based on 5.0.2
  • Added a Stratix I generic parameter for synthesis-time library selection.
  • Made a top-level modification that makes the interface compatible with the LM95235, while maintaining backwards compatibility.
  • Added the following commands for applying maximum-length sequences to MCE outputs: AWG_SEQUENCE_LEN, AWG_DATA, AWG_ADDR. See Arbitrary Waveform Generator .

• Details:
  • Arbitrary Waveform Generator (i.e. Maximum Length Sequences for Complex Impedance Measurements)

• Bugs:
  • (non-critical) when a card does not return a valid reply, the clock card does not send an error reply back to the PC. Instead, it fills the data with whatever it had in its buffer from previous command and sends it to mas.

• FPGA Resource Usage (clk_card.fit.rpt):

; Total logic elements     ; 18,095 / 32,470 ( 56 % )                      ;
; Total pins               ; 255 / 598 ( 43 % )                            ;
; Total virtual pins       ; 0                                             ;
; Total memory bits        ; 949,760 / 3,317,184 ( 29 % )                  ;
; DSP block 9-bit elements ; 10 / 96 ( 10 % )                              ;
; Total PLLs               ; 2 / 6 ( 33 % )                                ;
; Total DLLs               ; 0 / 2 ( 0 % )                                 ;

• Timing Analyzer Summary (clk_card.tan.rpt):

; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk0' ; 1.547 ns  ;
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk2' ; 1.985 ns  ;
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk1' ; 2.067 ns  ;

Revision 5.0.2

• Filename: cc_v05000002_test00_tagged.sof

• Features:
  • Header Version 6
  • Based off of 5.0.1 and in parallel with 4.0.c (equivalent version)
  • The Sync Box PLL was re-instated to the top level and routed to dv_rx. It was mistakenly removed after v4.0.9, which means that the sync box sequence number decoding has not worked since then! When collecting data while triggering off the Sync Box DV, the Clock Card would also sporadically trigger in between data frames.
  • Fixed a bug in ret_dat_wbs that did not handle wb num_rows_reported and wb num_cols_reported commands correctly.

• Bugs:
  • (non-critical) when a card does not return a valid reply, the clock card does not send an error reply back to the PC. Instead, it fills the data with whatever it had in its buffer from previous command and sends it to mas.

• FPGA Resource Usage (clk_card.fit.rpt):

; Total logic elements     ; 17,598 / 32,470 ( 54 % )                      ;
; Total pins               ; 254 / 598 ( 42 % )                            ;
; Total virtual pins       ; 0                                             ;
; Total memory bits        ; 818,688 / 3,317,184 ( 25 % )                  ;
; DSP block 9-bit elements ; 10 / 96 ( 10 % )                              ;
; Total PLLs               ; 2 / 6 ( 33 % )                                ;
; Total DLLs               ; 0 / 2 ( 0 % )                                 ;

• Timing Analyzer Summary (clk_card.tan.rpt):

; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk0' ; 1.965 ns  ;
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk1' ; 2.041 ns  ;
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk2' ; 3.548 ns  ;

Revision 5.0.1 (buggy)

• Filename: cc_v05000001_12may2009.sof

• Features:
  • Based on 5.0.0 and in parallel with 4.0.b (equivalent version)
  • Header Version 6
  • STOP commands are meant to work in this revision. The successful execution of STOP commands through the system depends also on the PCI card firmware, and PCI driver. Modifications have been made to these, and their version numbers have been bumped to...

• Details:
  • The STOP Command

• Bugs:
  • There may be a problem with decoding sync numbers from the sync box.

• FPGA Resource Usage (clk_card.fit.rpt):

; Total logic elements     ; 18,286 / 32,470 ( 56 % )                 ;
; Total pins               ; 259 / 598 ( 43 % )                       ;
; Total virtual pins       ; 0                                        ;
; Total memory bits        ; 1,537,536 / 3,317,184 ( 46 % )           ;
; DSP block 9-bit elements ; 10 / 96 ( 10 % )                         ;
; Total PLLs               ; 1 / 6 ( 17 % )                           ;
; Total DLLs               ; 0 / 2 ( 0 % )                            ;

• Timing Analyzer Summary (clk_card.tan.rpt):

; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk0' ; 1.995 ns  ;
; Clock Setup: 'clk_switchover:clk_switchover_slave|cc_pll:pll0|altpll:altpll_component|_clk2' ; 3.585 ns  ;
; Clock Setup: 'altera_internal_jtag~TCKUTAP'                                                  ; 5.644 ns  ;

Revision 5.0.0 (buggy)

• Filename: cc_v05000000_22dec2008.sof (tagged as sys_v05000000_22dec2008)

• Features:
  • Header Version 6
  • IMPORTANT: Must be used in conjunction with firmware v05000000 of all other cards!!!
  • This version is based on 4.0.a. That is, it includes all of the features that were under development in 4.0.a, even though 4.0.a was not released for telescope use.
  • Added the ability to read out a single column of data continuously from one Readout Card
  • New commands include: readout_col_index, readout_priority, num_cols_reported
  • To allow enough data bandwidth, the spare LVDS line from each card to the Clock Card is now used

• Details:

• Bugs:
  • There may be a problem with decoding sync numbers from the sync box.

• FPGA Resource Usage (clk_card.fit.rpt):

; Total logic elements     ; 26,607 / 41,250 ( 65 % )                 ;
; Total pins               ; 358 / 616 ( 58 % )                       ;
; Total virtual pins       ; 0                                        ;
; Total memory bits        ; 406,016 / 3,423,744 ( 12 % )             ;
; DSP block 9-bit elements ; 76 / 112 ( 68 % )                        ;
; Total PLLs               ; 1 / 6 ( 17 % )                           ;
; Total DLLs               ; 0 / 2 ( 0 % )                            ;

• Timing Analyzer Summary (clk_card.tan.rpt):

; Clock Setup: 'rc_pll:i_rc_pll|altpll:altpll_component|_clk0' ; 2.558 ns  ;
; Clock Setup: 'rc_pll:i_rc_pll|altpll:altpll_component|_clk2' ; 3.892 ns  ;
; Clock Setup: 'rc_pll:i_rc_pll|altpll:altpll_component|_clk3' ; 16.987 ns ;