Control direction of power flow and other things I don't know

revnarwhal

New member
Joined
Apr 8, 2019
Messages
9
I have this lab thing that I talked about in a past post.

Basic Power Configuration
  • Laptop Power Supply which goes through...
  • DC-DC Step Down Buckset to 16.8V which goes to...
  • BMS managing 18650 4S5P pack whichgoes to...
  • Power Switch which goes to the load
Load includes
  • DC-DC Step Down 5VVoltage Regulatorpowering aRaspberry Pi 3B+
  • DC-DC Step DownBuck setto 9V powering anArduino Mega
  • LCD Controller Board connect to LCD Screen
I am lacking fundamental electrical knowledge, and could use some feedback to get me on the path.

Here is the most generic version of my question:
How do I stop my battery pack from energizingthe LED on theDC-DC StepDownBuck that is supposed to be regulating the voltage from the Laptop Power Supply, and how do I do so without reducing the charging amperage greatly?

Attached is the closest thing I can provide as a drawing. Please never mind the markings on some of the listed componentsas they arein place to represent the type of component, not a specific version/model of that component. The diode I am usingis actually a 10A10 (which can allow up to 10A through it). Therelay I am usingis able to withstand coil voltage within the rangeof the Laptop Power Supplies I have as well as amperage higher thanwhat the Buck can withstand.

I've heard people say things like MOSFET and shunts and some other random assortments of letters... I'm learning as fast as I can andas I find things I don't know, but generally this just proves out how much I still need to know. Any advise or guidance about how to accomplish the simple goal

Thank you for your assistance


image_krvebh.jpg
 
I've even heard and experimented with the idea of using magnets to make the little electrobits go in only one direction.

What ways are there? How would you do it? What concerns might you have?
 
Learning is good :)

Does the small current the LED on the laptop to battery buck converter uses actually matter?
The other loads will add up to maybe 0.5A or more, so a few mA for the LED make any significant difference?

The simplest way would be to use a IN5822 Schottky diode (3A rated, low loss) connected like you've shown on the left side (right side is OK too but now ground of buck is raised voltage). You adjust the buck output a little higher to allow for the forward voltage drop across the diode (approx 0.3-0.4V for this type of diode). You could also use a regular rectifier diode like your 10A10 (drop ~0.6-0.7V). Voltages should stay "normal" on the battery side.
 
The LED only matters because it would be "always on" since there is no switch between the batteries and that buck, so they would be draining while the portable lab sits. The other concern was that I thought I read a warning about powering a buck in reverse, but not even sure about that.

Since you provided additional data about the loads, perhaps you can make sure I am not off base with my expectations.

This is what I have for the power supply itself:
Laptop PSU
Input: 100V-240V 2.5A-0.5A AC
Output: 90W 20V 4.5A DC

DC-DC Buck Converter
Input: 4-38V
Output 1.25-36V 5A
Conversion efficiency: 96%
Max Output wattage based on PSU: 86.4W (90W * .96)
Max Output wattage based on 16.8V: 84W (16.8 * 5 * .96)

This is what I believe is more or less the stats of the batteries:
4S5P (V*4 A*5)
V: 14.8V (12.8V - 16.8)
Ah: 12000mAh (2400mAh * 5) | 12Ah
C: unsure, but at least 1C, so 12A ???
Wh: 177.6Wh
Honesty: This is actually 5 separate 4S packs each with their own BMS, but essentially, they add up as shown.

This is what I thought was the max possible power need for the three load objects:
(RPi) In: 5V 2.5A DC (12.5W at 5V)
(Ard) In: 6~12V .285~1A DC (4.5W at 9V)
(Disp) In: 9~16V 2A DC (28W at 14V)
Total anticipated max power: 45W (W = V*A)

PSU/Buck should never need to provide more than 55% of their max power due to the "load" items alone (45W vs 84W).

4S5P should never need to provide more than 33% of it's max power (even at discharged voltage) due to the "load" items alone (45W vs 153.6Wh (12.8V * 12Ah)).

The 4S BMS is rated at 5A, so the charging of the 4S5P could max out the buck between the PSU and BMS (16.8V * 5A vs 16.8V * 5A * .96)

Not sure what would happen with these parameters:
4S5P depleted
Switch to load items in on position
PSU plugged in
I anticipate that most of the power from the PSU would be used charging the batteries, until they had sufficient charge (33%) to drive the load, but I really don't know.


Do I need to worry about "wave forms" and stuff like that? I mean, I have two different power sources inline, and one draws from the other sometimes. Also, in advance, I only know enough to put the phrase "wave forms" in this question, and also enough to realize that I am possibly "crossing the streams" as it were.


When it comes to the diode, I would need to get one rated at 5A or higher, no? Also, I am still learning about diodes, and trying to figure out what all the different versions are used for. I actually thought the 10A10 was a Schottky diode, so now I know I need to do more learning.
 
Re the loads, the RPi is not likely to be drawing 2.5A unless you have lots of gear plugged into the USBs & you're doing some hard work processing continuously. It might spike up a bit with CPU activity but they can draw only 0.5A or so running (mine does, RPi v3)
Display power seems high?
You could run both the display & the arduino from a 12V buck.
If you run the arduino with 16V, it'll probably pop.
Yes you're right about the diode, it needs to be rated for the bucks output current. You can usually find nice high current Schottky diodes in otherwise blown PC PSUs.... they have three legs & are a TOP220 package - look at 5V output section.
The 10A10 diode is definitely just a regular silicon diode, not Schottky!
The BMS I hope has a low voltage cut-off feature?
The loads would power up & draw their usual current when the voltage is right for them & basically this would be right down to flat for the batteries.
Nothing to worry about re "waveforms" it's all DC there.

Another way you could isolate the charger buck is by controlling a MOSFET from the input from the laptop & switching the output of that buck.
Ie like this:

image_qthxfa.jpg

The buck converter would only be connected when the laptop PSU is powered on
 
I based my amp figures on what was listed as "max" so that I would be building for worst case scenarios.

As I look at my RPi3B+ that is running right now, it is only drawing .27A, but while it boots, if it doesn't have access to more current you get that lightning bolt thing.
The items on the load side of the drawing I provided are just random things. My Arduino is behind a Step Down set at 9V.
The LCD Controller Board specs state it needs 2A+ and the seller told me that it would work fine direction from the 4S voltage range, so while I assume that it would not require 2A at higher voltage, that I still wanted to account for it "just in case".

I had previously been running the Arduino and display from a common Step Down at 12V (like you suggested) but separated them to have more individual power control, and so I could see how a device using the power in it's native form (12.2~16.8V per 4S) operated.

I have been recycling tons of parts and pieces, and I appreciate the feedback about looking into the PSUs. Just finding out that I could take a laptop display, buy a controller board for $35 or less, and end up with an HDMI monitor was amazing to me. I am working on a good list of where to source parts and what's worth while to scavenge and so forth.

The BMSs are all HX-4S-A01 (https://www.amazon.com/gp/product/B07HP754PQ/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1)
Features listed are: Over-discharge protection, overcurrent protection, overcharge protection, short circuit protection... sooo, nothing that is called "low voltage cut-off" specifically, but I believe that is what "over-discharge protection" equates to, but not certain.

Man, I don't yet understand all of your diagram, but I get the basic intent... I'm just annoyed that there are still so many things that I don't instinctively understand. In the case of your drawing, I see the resistors, and I comprehend why they are there, but I don't actually understand how they do the "why". Thank you for the drawing. I will try to grok it in fullness.

I spent about an hour last night playing around with rare earth magnets, nickle strips and 18650s because I am curious to know how they can and can't be used in quickly building battery packs, how they impact flow and how to make it safer since I've already had a few "exciting" moments.
 
revnarwhal said:
Could a DC capacitor be used to make the charge go in only one direction?

I believe you mean an "Electrolytic Capacitor", which has 1 leg Pos, the other Neg. These are normally connected in parallel with the Pos/Neg, and not in series one of the legs. And no, you can't. It's not a diode. Diodes and FETs are the only components really that can make current flow in one direction. FETs are basically Diodes with switches built in.
 
Electronics is interesting & fun to learn so don't stop, I'm still learning after "too many" years!

Re magnets, nickel strips, etc & using them to current flow, that's "fake engineering" at best & 100% B.S., sorry! Give up on that now, you'll save your eyebrows & nerves (from blast flash!) + it's damn dangerous!

re the BMS, "over-discharge" = "low voltage cutoff" so you should be fine there.

Capacitors store charge, they don't control it & you can't pass DC current through a capacitor, it just charges up like a jug filling with water & then current stops.

Think of MOSFETs & transistors as a two pin device where current flow is controlled by a third pin (gate or base).
With MOSFETs, gate volatge = Drain to Source current flow.
With transistors, base current (not voltage) = Collector to Emitter current flow.

Like Korishan says, MOSFETs have an inherent "reverse diode" built into them between Drain & Source

I put the resistors in that circuit because MOSFETs typically have a absolute max spec of 20V for the gate voltage & your power supply is already there.
MOSFETs are not tolerant of gate over voltage** so a little spike on connect/disconnect is likely to blow them.
MOSFETS typically have turned fully on (ie conducting, switch = "closed contact") with maybe 5V on the gate.
So the two equal resistors divide the 20V in half = ~10V on the gate = happy MOSFET.
The resistors (one anyway) also ensure the MOSFET is off if the laptop supply is disconnected by pulling the gate down to 0V.

** MOSFETs typically are tolerant of some D to S over-voltage & reverse voltage, transistors (C to E) are not
 
Thanks to you both for the feedback.

Korishan, yes, Electrolytic Capacitors. I learned enough about them to make a couple of poor qualityspot welders. I didn't learn enough to make a good enough spot welder to allow me to spot weld tabs to cells, so then I went down the road of magnets as a way of attaching cells.

image_lfdjay.jpg

From some of the applications I had seen with electrolytic capacitors it seemed like some people were using them as like an electric cistern... catching the magic electrobits and then letting them go out the bottom. It's likely I just didn't know what they were actually doing and made assumptions.


Redpacket, thanks for the more neophyte focused explanations. I'm gathering as much info as I can, but it often seems like my learning is based on the last thing I needed to know on a different project which then leads me to a new project before I finish the last project, only to return to the last project once I have played with the concept I just learned until I hit another gap in knowledge.

With regards to the magnets. I had made some nickle strips with magnets on both ends for connecting batteries quickly in series or in parallel and without soldering. My original version was sandwiching the nickle strips between the battery and the magnet. I kept the magnets in place with some superglue, but then found that almost every one of the superglued magnets still allowed conductivity, so I did more experiments. I even went as far as superglueing magnets directly onto both ends of a few cells I didn't care about to see if they would work as a quick way of attaching several in a stack (series) but that seemed to fail... someone somewhere told me that this would only work if the positive terminal of the battery was attached to the south pole of the magnet and the negative terminal to the north pole. I had been meaning to test this more, but got sidetracked by another project.

I think my hope was to have about 8 cells that I could rapidly fashion into any configuration I needed for a specific application or test.


For others who are trying to get as much information as possible as quickly as possible, let me direct your attention to an interesting place: Humble Bundle
Yes, often it's for games and such, but they do book bundles as well. One of them that is running right now is ELECTRONICS + 3D PRINTING BY MAKE
https://www.humblebundle.com/books/...medium=product_tile&hmb_campaign=tile_index_5
I've already picked up several bundles of books on different topics that I still have to find time to absorb.

When will I have that USB port in my head where I can insert PDFs and EPUBs directly into my memory?!?!


So a FET is sorta like a relay, except the charge that closes the relay also is the charge that passes through it?
 
revnarwhal said:
So a FET is sorta like a relay, except the charge that closes the relay also is the charge that passes through it?

Sorta, but not really. There are plenty of tutorials on youtube explaining how FETs work. The gate is triggered by voltage applied via the microcontroller. Altho, you don't "need" a controller to trigger the gate if you use a few other parts, like resistors and logic gates.
Some FETs only need, for example, like 5V to turn the gate on, but will allow 20V to travel through the switch. So they aren't "always" the same voltage. Other FETs act as variable gates depending on the voltage applied to the gate. For example, if 10V is fully closed, you could apply 5V to the gate and allow half as much current through the gate. so they could be used as current limiters as well.

I'm a bit rusty on the details, so the above may be slightly off. But again, there are a lot of YT videos that do some great explaining.
 
revnarwhal said:
With regards to the magnets. I had made some nickle strips with magnets on both ends for connecting batteries quickly in series or in parallel and without soldering.
....
When will I have that USB port in my head where I can insert PDFs and EPUBs directly into my memory?!?!


So a FET is sorta like a relay, except the charge that closes the relay also is the charge that passes through it?



OK that magnetic thing is fine, - I thought you meant something different. I thinkNorth or South pole matters onlybecause the cell cases are steel & magnets don't like samepoles facing each other. Eg if you stuck the say north pole of a magnet to a cell, it would "extend" the north pole to the other end of the cell. If you then tried to attachanother north pole magnet face at that end, they'd repel.

So yes a FET is sort of like a relay - the voltage on the gate "opens" (0V)or "closes" the relay (+10V, Nch type MOSFET).
Like Korishan says you can have in between partially on state but many power MOSFETshave a fairly sharp point from off to on, they're designed to be either off or on for switching circuit use.

The USB upgrade (join the Borg now)is no doubt coming soon :)
 
Back
Top