White Rabbit issueshttps://ohwr.org/project/white-rabbit/issues2019-02-12T08:15:21Zhttps://ohwr.org/project/white-rabbit/issues/34Fine-tune Synchronous Ethernet compliance document and upload to SVN2019-02-12T08:15:21ZJavier SerranoFine-tune Synchronous Ethernet compliance document and upload to SVNThe document must say in which way a WR node will respect the
Synchronous Ethernet standard. It should be lightweight (not more than 5
pages), like a quick digest of the standard for WR developers.
### Files
* [Synchronous_Ethernet_specification_for_White_Rabbitv2.pdf](/uploads/5f6615b02692c2c8afecddd3c94ac157/Synchronous_Ethernet_specification_for_White_Rabbitv2.pdf)Pedro MoreiraPedro Moreirahttps://ohwr.org/project/white-rabbit/issues/33PTP compliance document, upload to SVN2019-02-12T08:15:20ZJavier SerranoPTP compliance document, upload to SVNThe document must say in which way a WR node will respect the PTP
standard. It should be lightweight (not more than 5 pages), like a quick
digest of the standard for WR developers. Scenarios to cover include
communication between a WR switch and a WR node, but also what a WR
switch will do when hooked to a standard PTP-enabled node.
### Files
* [IEEE_1588_White_Rabbit_.pdf](/uploads/0380813fd34c05d602dec353979fb093/IEEE_1588_White_Rabbit_.pdf)Pedro MoreiraPedro Moreirahttps://ohwr.org/project/white-rabbit/issues/32Finish MCH v2 production and ship to interested developers2019-02-12T08:15:20ZJavier SerranoFinish MCH v2 production and ship to interested developersTomasz WlostowskiTomasz Wlostowskihttps://ohwr.org/project/white-rabbit/issues/31Document procedure to get MCH systems up and running in the lab2019-02-12T08:15:19ZJavier SerranoDocument procedure to get MCH systems up and running in the labTomasz WlostowskiTomasz Wlostowskihttps://ohwr.org/project/white-rabbit/issues/30Document embedded Linux setup for MCH2019-02-12T08:15:19ZJavier SerranoDocument embedded Linux setup for MCHEmilio G. CotaEmilio G. Cotahttps://ohwr.org/project/white-rabbit/issues/29Finish Wishbone Slave generator v12019-02-12T08:15:18ZJavier SerranoFinish Wishbone Slave generator v1Tomasz WlostowskiTomasz Wlostowskihttps://ohwr.org/project/white-rabbit/issues/28Write WR spec v22019-02-12T08:15:18ZJavier SerranoWrite WR spec v2Keep it simple (10 pages max). Use Pedro's docs as a basis. This
document answers the questions "what is WR?" and "what does a node need
to do to be WR-compliant?".Emilio G. CotaEmilio G. Cotahttps://ohwr.org/project/white-rabbit/issues/27Do 2010 planning2019-02-12T08:15:17ZJavier SerranoDo 2010 planningNot very detailed, just the big lines.Javier SerranoJavier Serranohttps://ohwr.org/project/white-rabbit/issues/26Generate high-level block diagram of MCH2019-02-12T08:15:16ZJavier SerranoGenerate high-level block diagram of MCHThis serves two purposes:
\- Make communication easier (we can refer to the drawing in our
discussions).
\- Help in planning, so we make sure there is no missing part in the
plan.Tomasz WlostowskiTomasz Wlostowskihttps://ohwr.org/project/white-rabbit/issues/25Do quick write-up on NIC memory map2019-02-12T08:15:16ZJavier SerranoDo quick write-up on NIC memory mapThis is to help the VHDL developers in the implementation of the NIC
block.Emilio G. CotaEmilio G. Cotahttps://ohwr.org/project/white-rabbit/issues/24WR switch v3 mechanics too wide2019-02-12T08:15:15ZJavier SerranoWR switch v3 mechanics too wideThe current design only leaves 0.5 mm on each side of a 19" rack, so it
might be difficult to install in some non-compliant racks. These width
should be slightly reduced. I just create this issue as a reminder for
the next version.José Gabriel RamírezJosé Gabriel Ramírezhttps://ohwr.org/project/white-rabbit/issues/22WR Switch V3 USB to program MCH not accessible2019-02-12T08:15:15ZErik van der BijWR Switch V3 USB to program MCH not accessibleThe USB connector of the MCH card with the processor on it points to the
side of the switch, while there is no hole in it. So the MCH cannot be
reprogammed without removing it from the switch.José Gabriel RamírezJosé Gabriel Ramírezhttps://ohwr.org/project/white-rabbit/issues/21V3 PCB: R225 value too high2019-02-12T08:15:14ZTomasz WlostowskiV3 PCB: R225 value too highR225 series resistor's value is too high, causing very long rise/fall
times on PLL status signal (PLL\_STATUS output of AD9516 is used by the
FPGA to monitor the external reference clock frequency). Replace with a
value matching the impedance of the trace (33 ohm?)José Gabriel RamírezJosé Gabriel Ramírezhttps://ohwr.org/project/white-rabbit/issues/20V3 SCH/PCB: connect 10 MHz external reference to a GCLK FPGA pin2019-02-12T08:15:14ZTomasz WlostowskiV3 SCH/PCB: connect 10 MHz external reference to a GCLK FPGA pinAdd a 0-ohm resistor (mounted by default) connecting the REF\_10M net to
any 3.3V GCLK pin of the FPGA. This is required for efficient operation
of BB phase detector locking the 125 MHz ethernet clock to external 10
MHz reference. The C185 connecting REF\_10M to AD9516 should not be
mounted by default.José Gabriel RamírezJosé Gabriel Ramírezhttps://ohwr.org/project/white-rabbit/issues/19Live heatsinks in mini backplane power supply2019-02-12T08:15:13ZTomasz WlostowskiLive heatsinks in mini backplane power supplyHeatsinks in the PSU installed in mini-backplanes are live. There's a
dangerous voltage between them (~150 VAC). Accidentally touching the PSU
may end with a very unpleasant shock.
This results from bad PSU design - the active components are insulated
from their heatsinks, but somebody connected the heatsinks to different
nets (shouldn't they both be connected to the PE wire?). There's no
warning label at all on the PSU.
This is a serious safety threat to the developers working on the switch.José Gabriel RamírezJosé Gabriel Ramírezhttps://ohwr.org/project/white-rabbit/issues/18miniBackplane18 - Fan noise too high2019-02-12T08:15:12ZErik van der BijminiBackplane18 - Fan noise too highThe fan noise is really high. This fan is in the back of the box.
- See if lower noise can be used.
- See if design can be made to lower the fan speed.
- Check possibilities of measuring correctly the temperature so a
feedback loop can be made.https://ohwr.org/project/white-rabbit/issues/17SCB I2C Temperature Sensor2019-02-12T08:15:12ZJosé Gabriel RamírezSCB I2C Temperature SensorThe I2C temperature sensor chips (TMP101NA) mounted on the SCBs V3.0 and
V3.1 are wrong. They must be changed by the TMP100NA P/N in order to
allow access to the 4 sensors. This bug is already corrected on V3.2
version of the switch.José Gabriel RamírezJosé Gabriel Ramírezhttps://ohwr.org/project/white-rabbit/issues/16V3.0-V3.2: switching DDMTD clock frequency2019-02-12T08:15:11ZTomasz WlostowskiV3.0-V3.2: switching DDMTD clock frequencyThe DDMTD clock frequency is currently fixed at 62.5 MHz, causing
routing problems for firmwares with \> 8 ports (ISE is unable to use a
global clock network, because there are too many non-clock loads on the
signal). In order to fix this, one should distribute a 125 MHz DDMTD
clock divided by 2 (single FF) just before sampling (in "reversed" mode
- i.e. DDMTD clock sampled by the measured clock). I recommend enabling
the FPGA to select between 62.5 and 125 MHz DDMTD clock.
In the already-produced PCBs, the fix can be done by:
\- removing R62,
\- cutting the +3V0\_VM53\_EN trace (the signal is not used in the WRS
design) on the bottom layer right next to a via between 'MidLayer2' and
'Bottom layer' (near RP5),
\- wiring pin AN15 of the FPGA to pin 14 of IC13 (from the nearby via,
~4mm of wire).Tomasz WlostowskiTomasz Wlostowskihttps://ohwr.org/project/white-rabbit/issues/15V3 - Some mini-USB connectors don't fit to front of switch2019-02-12T08:15:11ZErik van der BijV3 - Some mini-USB connectors don't fit to front of switchIt seems there is not enough space around the mini USB connector.
See https://www.ohwr.org/project/wr-switch-sw/wikis/Troubleshooting
7Sols is studying the problem.https://ohwr.org/project/white-rabbit/issues/14V3 - Protective Earch Connection2019-02-12T08:15:10ZDietrich BeckV3 - Protective Earch Connectionwe received three prototype switches from 7Sols for "experimental use".
We have asked two people at GSI to have a look at the switches from the
point of view of electrical safety. Here is their report.
- **Connection Missing**. The metallic housing must be directly
connected with PE, usually using a earth bolt at the housing. The
connection must be marked with the "earth symbol" or a label “PE”.
- **Connection of Cover**. The metallic cover must have a low
resistive connection to PE (e.g. to the earth bolt of the housing)
- **Coding**. PE connections (cables) must be achieved in
“green-yellow” color.
- **SUB-D Connector**. The housing of the SUB-D connector is not
connected to PE. It should be connected.