search

waterloo-rocketry / daq

Data Acquisition schematics, boards, datasheets, docs, and software
5 stars 3 forks source link

Briefcase power board #13

Closed BluCodeGH closed 2 years ago

BluCodeGH commented 3 years ago

This change is Reviewable

BluCodeGH commented 3 years ago
  1. Done (also realized that lower currents make webench use a bigger inductor so used a 2A max instead of 3A)
  2. Nope, on the recommended layout that net is a big fill
  3. The way it works is that it turns on above 19.3 and off below 19.2, the goal is to protect lipos so there's no resistor divider trying to hit within that range or anything. Nominal input is 24V on DC power but could range from 19.2 - 25.2 with lipos.
  4. 18.2 is a standard 1% resistor multiple.
  5. Yeah 10nF works fine, I don't see a reason to be picky about soft start time.
  6. 1.7Mhz roughly lines up with a PSRR peak of the LDO
  7. Webench just recommends the 10uF input capacitor so I have C1 as optional if we need to add a second.
  8. Added a note with what webench asks for
  9. Done
  10. Done
  11. Yeah we should probably be in that 80-85% efficiency range. Even if it's only 60% efficient at those low currents we're not dissipating a lot of power.
  12. Yes diode is in stock (but only 137 of them), luckily webench gives us tons of alternatives if it goes out of stock.
BluCodeGH commented 3 years ago

Ah yes, mounting holes, that should be no problem *furiously redoes layout*. I like the idea with the test points too, thanks!

BluCodeGH commented 3 years ago

Alright first pass at routing complete, I really have no idea what I'm doing when it comes to grounding things properly to avoid noise so reviews / knowledge would be greatly appreciated!

rdakhil commented 3 years ago

Looks great! Just a few comments:

  1. Nitpick but the datasheet link is missing for the buck converter on the schematic
  2. I would recommend moving L1 a little closer to the SW pin as it's a bit far which could cause some noise
  3. C2 should be as close to VIN on the buck converter as possible
  4. Not sure if there is a reason for it, but C9 is daisy chained to C12 and has no other connection to GND image
  5. Depending on how much current you could be running at, you might want to add a fuse to your battery connector and make any traces connected to +BATT thicker (eg. R3, R7) or alternatively you could use a +BATT polygon
BluCodeGH commented 3 years ago
  1. Done
  2. I couldn't get the layout to work out nicely with L1 closer, do you think it will be a significant issue? The reference design from webench has it fairly far away.
  3. Done
  4. This is a kelvin sense connection, it lets the LDO get an accurate voltage reading irregardless of high currents flowing through that capacitor.
  5. We're only looking at 1A max so the trace widths should be plenty. Also R3 and R7 are just for sensing so will only have a few milliamps at most.
BluCodeGH commented 2 years ago

I really want to get this ordered in the next couple days so DAQ can be ready for cold flows before I need to leave Waterloo, please review if you have some time!

BluCodeGH commented 2 years ago

  1. Don't worry these are big fat screw terminals, also this board is not going to be directly connected to a battery it will go through the power distribution of the briefcase first.

  2. Done

  3. Don't worry I used 0805 footprints, note was just wrong

  4. Datasheet said to /shrug, added a note

  5. Backup is the current pancake boards

  6. Wilco.

  7. 1.7Mhz has the highest PSRR on the LDO

  8. It's there (under the datasheet field).

  9. That's the amperage rating of the entire board not those terminals.

  10. All the tiny traces are just carrying signals, don't worry I made the actual power traces big. R12 is part of the feedback from the output voltage to the LDO.

  11. But fixing this would be a lot of work and messes up my symmetry :frowning_face:

  12. Moved a bit further.

  13. Since it's a super low noise power supply.

BluCodeGH commented 2 years ago

Hardware has been ordered, merging.