SMB Rev 3 Testing Revamp Project

[mtfunder]

No description provided.

[cwnicho2]

Earlier Work

• Redesigned for 48V input tolerance
  • several component swaps, often no change in footprint or layout required
  • significant redesign in 5V regulation now using I3A4 module
• Socket-able raspberry pi
  • moderate rerouting was required

Recent Work

• Test points added to SMB-V0.3, for key signals/voltages, consulted by @Djyankencsu
• Resistor values changes from prior sub version, implemented here
• Partsbox V0.3 project updated with proper designators and removed no longer needed parts from the BOM

Ready for higher-up review @mtfunder and ordering of at least a test batch

[cwnicho2]

Verified stock in real life and corrected partsbox stock. This is so that when we place an order we won't have incorrectly assumed what we have in stock already

[Djyankencsu]

On the schematic view, it is more readable to have all the test points together in one area and just specify each net where they connect. Probably most similar example on the schematic would be the Indicator LEDs section. Just add a Test Points section.

Also, talking with Mark, the test points do not need to be vias, and he was envisioning surface test points just exposed on the front copper. Also, he needs us to expose all the raw voltages going into the mux with dedicated test points. We should not rely on connectors on the board being used to measure voltages. With that in mind, it is probably also prudent to add a test point for the Pico 3V3 net.

He also asked me to do an initial double-check of the electrical properties and component selection. Checked U4, looks suitable. Maybe increase C20 to be 63-volt tolerant? Currently listed as 50-volt tolerant on a potentially ~48-volt net. D18 and D19 data sheet shows voltage drop between 0.3-0.4-volts, which given the Pico has another diode drop before its input at VSYS, it may be better to target slightly above 5-volts on the 5V Logic PWR/Relay nets if all components are tolerant.

Resistor values look good on the voltage dividers. All nets potentially carrying 48V have a 22:1 divider.

[cwnicho2]

Thanks Daniel,

• I reorganized the test points into a separate area on the schematic
• I added Pico 3V3 and all Analog Mux inputs as test points
• test points are changed to surface contacts instead of vias
• the capacitors C20 and C21 are 63V tolerant parts, the schematic was outdated, this is now fixed
• The pico diode I think could be removed, we now used D19 to prevent back current and D9 looks mostly redundant. This will land us on 5V
  Changes are pushed to github branch

[mtfunder]

These connectors will interfere with each other and maybe the capacitor, you may note, these are staggered on V0.2 because the plug overhangs the connector, as can be seen here:

[mtfunder]

Remove J5 & place J9 in its place (physically).

[mtfunder]

Move this via or add some copper here so it doesn't cutoff most of this large power trace for the Jetson:

[mtfunder]

Fix silkscreen for legibility, this is one particularly bad area but check entire board:
I think you had this fixed,, but something may have reset and stated showing the default values for each item in addition to the numbers you added.

[alexrallen]

Mark reviewing this design....

[cwnicho2]

Fixed the above comments on the board design and did a final review. Uploaded to Github and notified Mark over Slack to place an order

[cwnicho2]

Assembly Dec 2024

SMB_V0.3 was put into production.
15 boards were assembled and serialized SMB-V0.3-Y24-#
numbered 1-15
SMB-V0.3-Y24-1 and SMB-V0.3-Y24-2 were kept as testing models
It was found the Raspberry Pi was likely browning-out when closing relays which would cause it to restart
A capacitor on the input voltage of the Pi was used to fix this issue
All boards except 9 [15V not working] passed a hardware assembly inspection, and initial start up and power down test
Some connectors had to be removed and rotated to the proper orientation
They have been since cleaned with alcohol and dried