Created by [ Rowan Dempster], last modified by [ James Jo] on Apr 24, 2020
[\@Power_Systems] I've prepared a Q&A for the Rdar PCB based on the questions I've been asked.
Q: Should the power for the radars and media convertor be supplied
from the PCB
A: *Yes*, the main intention of the PCB is to centralize the
connections of all radars and convertors into one board.
Q: Why cant we just connect all the radars to a single media
convertor?
A: Because the media convertor manipulates the data of *1 radar only*
and supplies it out in ethernet. If we do that, we'll have to think
about 1) Separating data packets from different radars, and 2) Can the
media convertor handle data from 6 radars?
Q: Should we have 6 fuses or 1 big fuse?
A: We will have a diode for current direction verification, so there is
no need for 1 big fuse. However, we will need the 6 fuses (1 for each
radar) because if one goes bad during the comp, we still want the other
to work!
Q: Adding LEDs at each radar path? Should I add a transistor/other
circuitry to control it?
A: That's a nice idea. However, adding an LED as is good enough. You
don't have to make it complex i.e. LED with no transistors/other
circuitry. Also, if this is going to take time (it shouldn't) then
focus on getting the basic PCB done first. PLEASE DON'T ADD NEW
FEATURES IF THE BASIC ONES ARE NOT DONE YET.
Q: How big should the PCB be?
A: We don't have an official restriction on the size, but we won't
need to make it big anyway. Aim for as small as possible (which should
be less than the sample PCBs we've seen at the bay)
Q: How will the external connections connect to the PCB?
A: We will be using terminal blocks similar to the ones we've seen in
the LIDAR connection box (the ones with screws). All you need to do is
add the pin holes (like a solder breadboard) so we can purchase and
solder them on.
We may use some of the connectors we ordered earlier this term, but I need to see them again to decide.
Q: Do we still need buffered power supply?
A: Yes, because our testing showed some fluctuations for voltage. To
reduce the noise, we will use the buffered power supply designed by the
previous team(s).
[NOTE:] There are
actually 2 issues with what we have seen:
Q: Can we replace the regulators with others (like buck
convertors)?
A: Not at the moment. Please ensure that the design created by the past
teams is followed.
We will need to use the buck converters to step down the 24 V rail from
the compute rack to a suitable 6.5 - 18 V range for the radars and media
converters. Based off what was procured, we can step down the voltage to
9 V with overhead to account for the voltage drop from the wires and use
an LDO for direct power and noise and ripple filtering.
Q: How many buffer circuits do we need?
A: 6 - each with 1A regulators this is to handle noise better on each
radar path.
Update: As of 2020, we are using two radar components each supported by a single media controller. Therefore, we will need 4 buffer circuit paths. Furthermore, we have procured fuses that will be able to accommodate the 0.7 and 0.3 A of the radar and media converter respectively. Further testing will be needed to ensure proper operation.
Q: Though-hole vs Surface mount?
A: Through-hole to handle tougher environments
Q: How will we route the wires? How are the radars mounted?
A: We will have the circuitry be placed near the compute rack and allow the power supply to output 6.5 - 18 V with enough offset to allow the LDO to remain in operation over the distance of the wire. The radars themselves will be mounted near the bumper of the vehicle which we will allow the wires to pass through the internal body into the drivers seat. From there, we will route the wires to the right side to reach the power supply.
Document generated by Confluence on Nov 28, 2021 22:40