Difference between revisions of "Bias Card firmware"
From MCEWiki
(→Revision 5.1.0 (recommended)) |
|||
Line 1: | Line 1: | ||
* Recommended Firmware Revisions [http://e-mode.phas.ubc.ca/mcewiki/index.php/Compatible_Versions_(CC,RC,BC,AC) here] | * Recommended Firmware Revisions [http://e-mode.phas.ubc.ca/mcewiki/index.php/Compatible_Versions_(CC,RC,BC,AC) here] | ||
− | == Revision 5.1.0 ( | + | == Revision 5.1.0 (Stable) == |
; Filename | ; Filename | ||
: bc_v05010000_25oct2011.sof | : bc_v05010000_25oct2011.sof | ||
Line 36: | Line 36: | ||
; Details | ; Details | ||
: When installed on Rev. D file, tes_mapping has to have offset =11 in mce.cfg. | : When installed on Rev. D file, tes_mapping has to have offset =11 in mce.cfg. | ||
+ | |||
== Revision 5.0.7 (recommended) == | == Revision 5.0.7 (recommended) == | ||
; Filename | ; Filename |
Revision as of 11:27, 15 November 2011
- Recommended Firmware Revisions here
Contents
- 1 Revision 5.1.0 (Stable)
- 2 Revision 5.0.7 (recommended)
- 3 Revision 5.0.6 (lab use only)
- 4 Revision 5.0.5 (recommended)
- 5 Revision 5.0.4 (lab use)
- 6 Revision 5.0.3 (Rev. E cards)
- 7 Revision 5.0.2 (lab use)
- 8 Revision 5.0.1 (lab use)
- 9 Revision 5.0.0
- 10 Revision 1.4.2
- 11 Revision 1.4.1
- 12 Firmware Links
- 13 Wiki Links
Revision 5.1.0 (Stable)
- Filename
- bc_v05010000_25oct2011.sof
- Features
- In multiplex mode: DACs are clocked early in the row-visit, chip-select is always low and only strobed high at clock-cycle 1 during row-visit.
- In non-multiplex mode: flux_fb values are all loaded at the start of the frame. (Previous versions updated DACs row-aligned as oppose to frame-aligned and this had the undesired effect of refreshing some DACs on one row and rest of the DACs on the following row depending on which clock cycle the command was received. This only matters when running internal ramp on bias-card parameters.)
- ln_bias lines are only refreshed once and frame-aligned! a bugfix!
- Details
- When installed on Rev. D file, tes_mapping has to have offset =11 in mce.cfg.
Revision 5.0.a (lab use)
- Filename
- bc_v0500000a_24oct2011.sof
- Features
- based on 5.0.9, with fixing the incomplete ifelse in spi_dac module. Was this causing the improper chip-select?
- Details
- When installed on Rev. D file, tes_mapping has to have offset =11 in mce.cfg.
Revision 5.0.9 (lab use)
- Filename
- bc_v05000009_05oct2011.sof
- Features
- based on 5.0.5, but instead of loading a new DAC value at the start of the row, it is now 32 clock cycles after row switch.(or clocking starts at row-switch)
- Details
- When installed on Rev. D file, tes_mapping has to have offset =11 in mce.cfg.
Revision 5.0.8 (lab use)
- Filename
- bc_v05000008_04oct2011.sof
- Features
- based on 5.0.5, but instead of loading a new DAC value at the start of the row, it is now 20 clock cycles after row switch.
- Details
- When installed on Rev. D file, tes_mapping has to have offset =11 in mce.cfg.
Revision 5.0.7 (recommended)
- Filename
- bc_v05000007_22jun2011.sof
- Features
- based on 5.0.5, but instead of loading a new DAC value at the start of the row, it is now 10 clock cycles after row switch.
- Details
- When installed on Rev. D file, tes_mapping has to have offset =11 in mce.cfg.
Revision 5.0.6 (lab use only)
- Filename: bc_v05000006_11may2011.sof
- To Do:
- Features:
- Changes the timing of loading the FB_COLxx values from 32 clock cycles before the start of a new frame, to the start of a new frame. This firmware was generated to find the reason for the downturns in raw data at the end of rows, as seen by Jeffrey Flilippini:
- See details here: http://spiderwiki.princeton.edu/spider/AnalysisLogbook/TestCryostat/20110427_muxing/index.html
- FPGA Resource Usage (bias_card.fit.rpt):
; Total logic elements ; 6,278 / 10,570 ( 59 % ) ; ; Total pins ; 209 / 427 ( 49 % ) ; ; Total virtual pins ; 0 ; ; Total memory bits ; 133,120 / 920,448 ( 14 % ) ; ; DSP block 9-bit elements ; 8 / 48 ( 17 % ) ;
- Timing Analyzer Summary (bias_card.tan.rpt):
; Fast Model Clock Setup: 'bc_pll:pll0|altpll:altpll_component|_clk0' ; 6.930 ns ; Fast Model Clock Setup: 'bc_pll:pll0|altpll:altpll_component|_clk1' ; 6.980 ns ; Fast Model Clock Setup: 'bc_pll:pll0|altpll:altpll_component|_clk3' ; 16.383 ns
Revision 5.0.5 (recommended)
- Filename
- bc_v05000005_20jul2010.sof
- Features
- based on 5.0.4, but updated for Rev. F/D hardware pinout.
- card_type parameter now includes the pcb-revision information as well as the card_type. card_type is specified as 0x01 as the lower byte (same as before). Reading back card_type parameter returns: 0x0F01 in Rev. D cards, and 0x0601 in Rev. F cards. (Previously, reading back card_type parameter only returned 0x01).
- Details
- When using tes_bias lines, make sure the count is 12 for bias parameter in mce.cfg file.
- When installed on Rev. D file, tes_mapping has to have offset =11 in mce.cfg.
- To do
- none
- Bugs
- none
- FPGA Resource Usage
: Total logic elements ; 60% : M512s ; 49% : M4Ks ; 100% : M-RAMs ; 100%
Revision 5.0.4 (lab use)
- Filename
- bc_v05000004_20may2010.sof
- Features
- build based on 5.0.3 for Rev. E cards
- Biases are refreshed 1 clock cycles after the start of a new row, regardless of whether they are running in multiplex mode or not and whether it is a regular bias line or low-noise bias line.
- Details
- NOTE If this firmware is loaded on a Rev. D card, it's harmless. The only functionality loss is that the bias command will not work as unfortunately DAC CS and CLK pins are swapped for the bias DAC between the two revisions of the card.
- To do
- none
- Bugs
- none
- FPGA Resource Usage
- Total logic elements ; 60%
- M512s ; 49%
- M4Ks ; 100%
- M-RAMs ; 100%
Revision 5.0.3 (Rev. E cards)
- Filename
- bc_v05000003_12may2010.sof
- Features
- build based on 5.0.2 for Rev. E cards
- added fb_col0 to fb_col31 and enbl_mux commands to accommodate row-based sq2fb switching. When enbl_mux is asserted for a column, then the DAC is refreshed on every row visit with the value specified by fb_col command.
- Details
- The original Bias Card design has a buffer after the DAC that is too slow followed by kHz range RC filters, so in order to enable the multiplexing feature, one has to make sure that the BC hardware is modified as per ECO-xxx (to be added here).
- card_type returns 5 (indicating a Rev. E)
- If this firmware is loaded on a Rev. D card, it's harmless. The only functionality loss is that the bias command will not work as unfortunately DAC CS and CLK pins are swapped for the bias DAC between the two revisions of the card.
- combinational loops present in previous versions are removed now
- There is 9+16*2=41 clock cycle (820ns) delay for the bias to be applied after the start of a new row. This combined with another 400ns delay due to inherent DAC delay is about 60 clock-cycle delay.
- To do
- The excessive 41 clock cycle delay to apply a new bias has to be reduced. It is conceivable to preload the DACs and reduce this delay to 1 to 2 clock cycles.
- Bugs
- none
- FPGA Resource Usage
- Total logic elements ; 6,140 / 10,570 ( 58 % ) ;
- Total memory bits ; 133,120 / 920,448 ( 14 % ) ;
- M512s ; 48 / 94 ( 51 % ) ;
- M4Ks ; 60 / 60 ( 100 % ) ;
- M-RAMs ; 0 / 1 ( 0 % ) ;
Revision 5.0.2 (lab use)
- Filename
- bc_v05000002_xxjan2010.sof
- Features
- build based on 5.0.1 for Rev. E cards
- Independent control for ln_bias lines 0 to 11
- Details
- If this firmware is loaded on a Rev. D card, it's harmless. The only functionality loss is that the bias command will not work as unfortunately DAC CS and CLK pins are swapped for the bias DAC between the two revisions of the card.
- To do
- fix combinational loops on read-ram register
- Bugs
Revision 5.0.1 (lab use)
- Filename
- bc_v05000001_19jan2010.sof
- Features
- supports the new low-noise bias lines (total of 12) introduced in Bias Card Rev. E
- card_type parameter is set to 5 (bias-card Rev. E)
- All DACs are loaded at once as oppose to previous revisions that loaded them one after next.
- DAC clock is now 25MHz and generated by PLL, previous firmware had 12.5MHz clock generated by dividing down.
- Details
- If this firmware is loaded on a Rev. D card, it's harmless. The only functionality loss is that the bias command will not work as unfortunately DAC CS and CLK pins are swapped for the bias DAC between the two revisions of the card.
- To do
- All ln_bias lines are controlled at once and this should be modified to independent control.
- Bugs
ln_bias_0 doesn't work, because the data line for ln_bias_0 is not connected in firmware.
Revision 5.0.0
- Filename: bc_v05000000_22dec2008.sof (tagged as sys_v05000000_22dec2008)
- To Do:
- Features:
- IMPORTANT: Must be used in conjunction with firmware v05000000 of all other cards!!!
- To allow enough data bandwidth, the spare LVDS line from each card to the Clock Card is now used
- Details:
- Bugs:
- None yet reported
- FPGA Resource Usage (bias_card.fit.rpt):
; Fitter Status ; Successful - Wed Jan 14 11:19:37 2009 ; ; Quartus II Version ; 8.1 Build 163 10/28/2008 SJ Full Version ; ; Device ; EP1S10F780C5 ; ; Total logic elements ; 3,356 / 10,570 ( 32 % ) ; ; Total pins ; 187 / 427 ( 44 % ) ; ; Total memory bits ; 70,144 / 920,448 ( 8 % ) ;
- Timing Analyzer Summary (bias_card.tan.rpt):
; Fast Model Clock Setup: 'bc_pll:pll0|altpll:altpll_component|_clk1' ; 7.060 ns ; ; Fast Model Clock Setup: 'bc_pll:pll0|altpll:altpll_component|_clk0' ; 7.125 ns ; ; Fast Model Clock Hold: 'bc_pll:pll0|altpll:altpll_component|_clk0' ; 0.383 ns ; ; Fast Model Clock Hold: 'bc_pll:pll0|altpll:altpll_component|_clk1' ; 0.384 ns ; ; Fast Model Recovery: 'bc_pll:pll0|altpll:altpll_component|_clk0' ; 16.037 ns ; ; Fast Model Removal: 'bc_pll:pll0|altpll:altpll_component|_clk0' ; 0.575 ns ; ; Total number of failed paths ; ;
Revision 1.4.2
- bc_v01040002_15jul2008.sof
Features:
- Added the tga_i signal to the sensitivity list for the bc_dac_ctrl FSM
- potential bug may have been introduced with resetting the DAC counter due to inadvertently using the wrong bc_dac_ctrl_core.vhd version.
Revision 1.4.1
- bc_v01040001_25jan2008.sof
Features:
- Added card_type and scratch commands
- Integrated fw_rev and slot_id as part of all_cards
- Added provisions for safe state machines to fix the reset problem.
- Added lvds_tx = 0 so clock-card can check whether the card is plugged in using this spare lvds line.
Bugs: None so far