Design review of CONV-TTL-RS485-RTM-OPT
Held on 15th of December 2017 at CERN, for version 1 of the schematics.
Designer:
- Dimitris Lampridis (BE/CO/HT)
Reviewers:
- Denia Bouhired-Ferrag (BE/CO/HT)
- Tristan Gingold (BE/CO/HT)
- Erik Van Der Bij (BE/CO/HT)
- Tomasz Wlostowski (BE/CO/HT)
Detailed comments on schematics
General comments
- Put "no-ERC" directives to all unused pins, even the NC ones.
- Done.
- Use official project name (CONV-TTL-RS485-RTM-OPT).
- Done.
- Add ESD strips.
- Done.
- Add fiducials.
- Done.
- Some labels appear upside-down.
- Rejected. This is a "problem" in newer versions of Altium. A
symbol rotated by 180 degrees will also have its labels rotated.
To avoid this, just untick
Preferences
Schematic>Graphical Editing->Display Strings as Rotated.
- Rejected. This is a "problem" in newer versions of Altium. A
symbol rotated by 180 degrees will also have its labels rotated.
To avoid this, just untick
Preferences
- Fix typos in copyright notice.
- Done.
- Don't use different symbols and footprints for GND test points.
- Done. All test points (except ESD strips) use 1mm holes. If necessary, single pins can be soldered for GND test points (to make it easier to attach probe ground clips etc).
- Fix page numbering
- Done.
- Add "drawn by" field.
- Done.
- Add "checked by" field.
- Done.
- Check netlists as described in
ed:Schematics-checklist.
- Done. Exported from Altium to OrCad netlist and performed manual check.
Sheet 1: Top-level Schematic
- Add a note to describe the channel configuration (number of
transmitters and receivers per channel).
- Done.
- Add URL of project.
- Done.
- Add note that this project is used in conjuction with
CONV-TTL-RS485.
- Done.
Sheet 2: Optical Channel
- Switch RS485 termination and impedance to 100Ω and explain why (due
to 100Ω being used in the front board - CONV-TTL-RS485).
- Done.
- Make power and ground net connections for quad buffer visible.
- Done, switched to single-part schematic symbol for the quad buffer, which always shows the power pins.
Sheet 3: Optical Receiver
- Add note that design is based on CERN CTDPR receiver.
- Done.
- Add positive feedback to the comparator and attach a series diode,
series resistor and parallel capacitor to form a basic sample and
hold circuit, which could then be digitized to provide a basic
receiver monitoring solution.
- Rejected. The idea of adding power monitoring to the receiver has been already discussed and rejected due to addition of circuit complexity for little gain (very few interventions for our existing optical transceivers were necessary in the previous years, never for a failing transmitter/receiver. All problems were related to faulty 3U power supplies and/or unconnected/wrongly-connected fiber cables).
- Fine-tune the hysteresis of the comparator.
- Rejected. Current setup (no external hysteresis network) has worked for many years on the existing CTDPR receiver and it has been extensively tested in our first prototype as well.
- Improve layout of schematic, do not make connections only with net
labels.
- Done.
Sheet 4: Optical Transmitter
- Add note that design is based on CERN CTDAB transmitter.
- Done.
- Contact Design Office to finalize the schematic symbol (and
footprint) for the laser transmitter, and remove the "preliminary"
note from them.
- Done.
- "Do Not Place" for potentiometer done with Altium design variants
not clear, use the "Mounted: No" parameter.
- Done.
- Remove PECL driver and drive the laser driver directly with the TTL
output of the RS85 receiver, with appropriate biasing.
- Rejected. Existing design is based on CTDAB and it has been extensively tested with our first prototype.
- Connect the signal from the monitoring photodiode to the laser
driver MD pin in order to have the "FAIL" pin of the laser driver
working and driving a panel LED (useful for monitoring).
- Rejected. According to the datasheet of the MAX3766 Rev 1 datasheet, page 15, in order to disable Automatic Power Control (APC), MD and FAIL should be left unconnected. APC is not wanted because it requires a balanced signal and we want to be able to transmit single pulses as well, without encoding. Added relevant design note in schematic.
- Fine-tune laser biasing for optimal timing.
- Rejected. According to the datasheet of the MAX3766 Rev 1 datasheet, page 15, in order to disable Automatic Power Control (APC), BIAS should be connected to 5V and BIASMAX should be left unconnected. Added relevant design note in schematic.
- Add note explaining the function of all MAX3766 pins with fixed
values.
- Done.
- Remove the note regarding the replacing of the 22Ω resistor with 0Ω
in the negative output of the laser driver. It is not necessary and
confusing.
- Done.
Sheet 5: VME64x RTM Connector
-
Harness seems to be inverting RX and TX channels.
- Done. Compared with first prototype, problem verified and fixed.
- Set RTMP ID to not connected (all open).
- Done.
Dimitris Lampridis - 19 December 2017