Author Topic: Feasibility of my power design?  (Read 1054 times)

0 Members and 1 Guest are viewing this topic.

Offline adj8shTopic starter

  • Newbie
  • Posts: 3
  • Country: us
Feasibility of my power design?
« on: July 09, 2019, 02:08:55 am »
Hi,

I'm a researcher trying to power multiple instruments for a UAV payload using a singular LiPo battery. I'm struggling to solve all my issues with my current design linked here: https://imgur.com/a/epNa0qz. I'm currently stuck with several problems.
  • From what I understand, the current draw of this system (~6.5 amps) greatly exceeds the discharge capacity of any LiPo battery. At best I've been able to source a 50C 10000mAh, I would need something like a 10000mAh 65 or 70C.
  • My current solution (independent batteries for each device, excluding device 2 on the schematic) is to use an independent battery for each, which adds a total payload of ~900g. The goal of my design to reduce the weight by at least 300g, anything more than that would be an awesome bonus.
  • I'm also uncertain on how to physically handle the wiring splits; most voltage regulators I see have only one positive and negative lead on the step-down end. I need one +/- line to go to the 12v-5.5v stepdown, and two additional lines going to each 12v device. Would I just solder multiple lines of wire of the same gauge together?

Any advice or input would be greatly appreciated! I'm far from an electrical engineer, and while this task is a little daunting to me at the moment, learning how to do this is pretty fun :D

Edit: I just realized I had a unit conversion mistake; so I think I'm good for question 1. A 50C 5000 mAh battery should be able to discharge 250 amps, so total current load should not be an issue with this design.
« Last Edit: July 09, 2019, 03:29:04 am by adj8sh »
 

Offline adj8shTopic starter

  • Newbie
  • Posts: 3
  • Country: us
Re: Feasibility of my power design?
« Reply #1 on: July 09, 2019, 04:01:19 am »
Some more questions. I'm not having much luck finding a proper step-down converter that goes from 14.8v to 12v while also having a maximum current draw that is > 6 amps. So far all I have found is this (https://hobbyking.com/en_us/linear-regulator-4s-to-12v.html). Are there other ways to drop down the voltage while supporting a large current draw? Or am I not searching in the right spot for what I need?
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 12864
Re: Feasibility of my power design?
« Reply #2 on: July 09, 2019, 07:06:34 am »
Its probably easier to run the 5V regulator direct from the battery (after a fuse) rather than running it from the output of the 12V regulator.  That reduces your 12V load current requirement to under 3A which makes it a lot easier to find suitable buck converter modules,   You'll have to look at the max. current draw of the  Pi + other 5V loads and the weight of the buck converters that would be required to see if it would also be worth splitting off the device 3 load to a separate buck converter module.

Deciding where and how to splice wires can get complicated when every gramme counts.   In general a crimp is probably going to be lighter, more compact and more reliable than a properly done soldered splice.   
« Last Edit: July 09, 2019, 07:08:33 am by Ian.M »
 

Offline adj8shTopic starter

  • Newbie
  • Posts: 3
  • Country: us
Re: Feasibility of my power design?
« Reply #3 on: July 12, 2019, 01:51:19 pm »
Thanks for the feedback Ian. After more searching and tweaking, I ended up with this design: https://imgur.com/a/lLmf6Sk

The YEP-20HV (https://hobbyking.com/en_us/yep-20a-hv-2-12s-sbec-w-selectable-voltage-output.html) seemed custom tailored for my application (especially with two sets of leads). Two quick questions:

  • Would it be safer to have two separate fuses placed before each buck converter? I.e., a 5A before the 14.8->12 stepdown, and the 10A before the 14.8->5 stepdown? I'm not sure if that's necessary or overkill, but I definitely don't want to compromise damaging any of the instruments/devices.
  • Another naive question, on the LM2596 (https://www.addicore.com/lm2596-module-p/ad281.htm), would I just solder the appropriate +/- leads to each of the four marked holes on this circuit? Or does it connect a different way?

Thanks again for the help!
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 12864
Re: Feasibility of my power design?
« Reply #4 on: July 12, 2019, 03:34:32 pm »
The purpose of a fuse is to prevent a wiring fire or tracks burning off a PCB if a semiconductor fails or some other short-circuit occurs.  In general, semiconductors blow before fuses do.  Don't expect a fuse to prevent damage to your electronics, although the odds of the board being repairable are much higher if the circuit is appropriately fused.

Unless you are using excessively thin hookup wire that cant handle 10A, or the PCB has seriously undersized tracks, you probably only need one fuse as close to the battery as possible.

Re: soldering wires to a LM2596 module or similar.  Yes, that's how its usually done, unless you use pins for a direct board to board mounting, but soldered stranded wires without strain relief are unsatisfactory in high vibration environments and tend to fail by fatigue where the individual strands enter the solder joint .  e.g. if you are mounting it in a RC helicopter you'd probably need to crimp terminals on the wires intended for direct PCB mounting, that also grip the wire insulation and solder the terminals in the holes to prevent any solder wicking up the wires.   

Caution: many LM2596 modules cant deliver their rated current.  Don't expect more than 2A without additional cooling, and if its got a fake LM2596 chip, you'll be lucky if it can do much over 1A.   Fakes that operate at ~50KHz instead of the 150KHz of the genuine chip are extremely common, with the problem being that the inductor and capacitors aren't big enough for the lower frequency.
« Last Edit: July 12, 2019, 03:53:41 pm by Ian.M »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf