Hacker News Reader

Comment by StavrosK | Hacker News Reader

After getting into drone racing (where you need to take good manual care of your batteries because they're basically a bunch of cells soldered together), I realized that every battery we have everywhere is just some standard cells in some standard configuration.

With that knowledge, you can replace the battery in any gadget (small ones are usually 3.7V, or 1S), if you know how to care for it. Replacing the battery in a MiniDisk like this, I would maybe not trust the internal MiniDisk charger and charge by hand, but it should otherwise work identically to the factory battery.

View on HN

This might work if you're the kind of person to continue manually checking and caring for the battery pack, but it's not good advice for the average hobbyist.

The crucial missing detail is that battery packs are assembled from identically matched cells, which age together as a unit. As soon as you replace a single cell with a new cell of the exact same model, or a new cell of a different capacity and charge characteristics, that cell will charge differently than the rest of the battery pack. This results in an unbalanced charge situation.

If the battery pack has a battery management system with leads in between each cell to balance the pack, this might work. The BMS can drain off excess charge from the unbalanced cells and convert it to heat.

If the battery pack does not have a BMS, the unbalanced charge situation can lead to pack failure, with varying degrees of smoke and flame depending on the battery chemistry.

When in doubt, play it safe and just get a new battery pack.

generally in these situations you'd be replacing all the cells at once, so they are reasonably matched. Nobody tears apart a battery and resolders everything just for one cell, that's a waste of time.

and if you're really concerned about it, you can match cells for impedence yourself, but it's generally not a problem if you replace all the cells at once with cells of the same kind/batch/etc.

I was actually wondering how balance chargers work, do they really burn off excess voltage with resistors? I always assumed they applied extra voltage to charge the low cells, rather than discharge the high ones.

Yes.

Yes, usually they use ADCs to read the voltage in between each cell in the series. If individual cells have higher voltage than other cells, a resistor for that cell is switched on to burn off excess charge.

More advanced controllers can use flyback transformers to move charge from one cell to another. This is vastly more expensive than just using a resistor, though, so it's only used in applications where energy conservation is key, like solar projects or where heat is a constraint. The LTC3300 is a good example: https://www.analog.com/en/products/ltc3300-1.html#

That's good to know, thanks!

Balance chargers for small packs, like you'd have for a radio controlled aircraft, have separate wires for the positive and negative plus in between each cell. So, a 6-series configuration would have 7 wires and the charger uses them to charge each cell separately to the correct voltage.

The problem with that is that each wire has to be able to handle the full charge current, but at a low voltage. So, if you're dealing with, say, an electric car with dozens or hundreds of cells and thick copper cables the size of garden hoses it's no longer practical. Instead, you charge it the simple way (by applying a large voltage to the whole series string through a positive and negative lead) and use a battery management system that measures voltages between cells and drains any high-voltage cells gradually through a resistor. It's kind of wasteful, but it's fairly simple and if your batteries are well balanced the BMS shouldn't have to do much at all.

Using the high cells to charge the low cells would be a nice feature; I'm not aware of any EV BMS that does that, but I'm not an expert and I'm really only aware of what's going on with DIY conversions. I don't know what the state of the art is for BMSs in commercially manufactured vehicles.

On a tangent, but... do you think it's possible to build a battery that discharges as a pack (in series), but charges each cell individually?

I imagine it will require some creative wiring, but charging efficiency should then go through the roof, right? I image I went e-biking and returned to my car to recharge - having individually chargeable cell should reduce the overall charge time a lot.

A battery management system handles this.

You still charge the pack in series, but the BMS will discharge cells that have a higher voltage than others. This results in all cells having similar charge levels.

It doesn't change the charge time, though. You're putting the same amount of energy into the pack either way. 14.8V @ 100mA applied to 4 cells in series is the same energy as 3.7V @ 100mA applied to 4 cells individually. Can't escape the laws of physics.

If power efficiency is critical (solar applications, for example), the BMS might have fancy switching circuitry to re-route energy from overcharged cells to undercharged cells. This is expensive, though, so most BMSs just use resistors to burn off extra charge from overcharged cells.

> You still charge the pack in series

I understand that's how it's done, but can we charge faster if we DECIDE NOT to charge in series? I take it the answer is no?

Imagine a series of 14 cells. I could charge them as a whole, but there is normally some limit on how much power I can pump into it. Now suppose that I disassembled the series and now I have 14 cells that I can charge independently before re-assembling them back. If I did that, would I gain anything? If nothing else this should prevent unbalanced charging, reducing overall heat of the pack and allowing higher current overall?

I can see that some packs advertise 5a charge current and other packs of the same capacity offer 10a or even 15a. Wonder why is there such difference.

It would be marginally faster (best case probably a percent or two) for a much more expensive charging circuit. The reason for this is that you can discharge batteries much faster than you can charge them.

Also, it seems like you're imagining that you can charge batteries with more power if you charge them in parallel instead of series. This isn't true since you're just trading a higher voltage/lower current for a lower voltage/higher current.

No free lunch for me then. Thanks for explaining.

Not a problem, just expensive. Wires running to each individual would have to be larger than ones used just for sensing or small charge bleeding.

I was considering a design where there was a PCB on every cell with an individual charger, but it's just more cost and parts to break, plus no redundancy (unlike having two big chargers in parallel).

How about a middle ground then?

Let's make a few 3-series batteries and charge them. When done charging we stack four of them together to get a single 12-series to power an e-bike. Would that allow for much faster charging?

No. Charging in Series or in Parallel still requires the same amount of power.

I don't think it's very expensive. Nearly all hobbyist RC drone batteries that are have 2 or more cells in series have a big discharge lead with all the cells in series, plus a smaller parallel "balance port" that provides access to each individual cell. Balance ports are used for charging, and can also be used in flight to wirelessly report the voltage of each cell. There are plenty of common (and apparently somewhat reliable) balancing chargers for under $30.

A popular $22 balancing charger for hobby RC drones: https://hobbyking.com/en_us/turnigy-accucel-6-50w-6a-balance...

A $15 sensor that plugs into the balance port and sends telemetry data about each cell: https://alofthobbies.com/frsky-sp-flvs-smart-port-lipo-volta...

A similar sensor that beeps really loudly when any cell falls to 3.5 volts (these are very popular for people without telemetry systems): https://outofdarts.com/products/lipo-voltage-alarm-1-mini-si...

An article explaining balance ports in RC aircraft batteries: http://www.tjinguytech.com/charging-how-tos/balance-connecto...

What problem are you trying to solve? Wasting a tiny bit of power in resistors?

I’m not sure what you mean. I believe the problem is that battery cells wired in series will tend to discharge at slightly different rates due to slight differences in their composition from their manufacture, and that cells can be damaged when their voltage goes outside a certain minimum or maximum range, so it’s important to monitor each cell individually while discharging and charging.

Absolutely, and this is not unheard of in custom e-vehicle wiring.

You have a discharge circuit, in series, and a charge circuit, wired separately to each cell. You're only using one at a time, as long as everything is properly grounded they don't get in each other's way.

It's not really about charge time though: any vehicle battery is going to discharge faster than it charges, so it's practical to just dump a bunch of current across the pack if you're charging using the same circuit.

What it does, is takes good care of each individual cell, and if you log some metrics you can detect underperforming cells and replace them.

I will say that for most applications, charging off the series wiring, and adding an overvoltage module that shuts off charge to each cell when it's topped off, is going to be much simpler and good enough.

But it can be done.

I guess any fast charge pack has this, cars in particular. Pushing large currents thru strings of high impedance empty batteries must be dangerous and energy inefficient, while switching them to parallel configuration would make a world of difference. I'd be surprised if there weren't power routing circuits already available from major IC manufacturers.

The only difference between charging them in series or parallel is that the Volts/Current ratio changes.

The actual power required does not change..

Doesn’t it improve efficiency? Squeezing electrons through a string of exhausted batteries require current over a longer time and therefore more losses.

Even though cells may exist with the same specs, it's hjighly recommended to create a battery pack from a single supplier and standardize on that. For example when I used to do professional laptop repair, I discovered the '18650' is only a size rather than a catch-all for all 12V cells, and isn't recommended to daisy-chain the various good ones together if they only share the same voltage. There's over a dozen different colors that indicate these batteries were built by various suppliers with use cases such as fast discharge rate, higher peak voltage, greater capacity, or a cost-effective battery that survives fewer discharge cycles. Creating a pack with mismatched batteries could cause the load balancer or device to operate outside of spec- don't cheap out on your power source!

Agreed (except you probably mean 3.7 V cells and not 12 V). Your cells should be as similar as possible to prevent uneven wear.

I don't disagree, and building battery packs for battlebots was a fun experiential way to learn this, but it is important to match chemistries. Chargers for nickel-metal-hydride batteries won't correctly charge lithium-ion batteries, or worse nickel-cadmium batteries. Not to mention that some battery packs (the ones in the Thinkpad come to mind) have a "battery monitor IC" actually built into the battery pack so it helps to start with an existing pack and then "re-manufacture" it with cells rather than just wire together the same number of cells.

Perhaps to most critical element in these "smart" batteries is the temperature sensor which dials back the current when they start getting hot to avoid thermal runaway. (source of MANY battery fires in laptops from counterfeit battery packs).

As for matching the cells, in our battlebot we would take each "unit" (which in our case was 20 NiMH batteries) and condition it by draining it and charging it in cycles while monitoring voltages to level out the batteries and "sync" their charge levels. We still ended up with a couple of fires but that was because we were charging them right at the limit of what is reasonable to do because the time between bouts could be pretty short.

> I would maybe not trust the internal MiniDisk charger

Why wouldn't you? It has same voltage and even same capacity (weird considering evolution, but maybe choice was deliberate). I dunno, but maybe MiniDisk charger even has temperature sensor.

Do you just reuse the battery management system (bms) that’s already in the cell?

It is a single-cell battery so no onboard bms would typically be present - a look at the terminals confirms this. The charge and protection logic is built into the device which charges the battery itself.

edit: http://forums.sonyinsider.com/topic/29187-lip-4wm-battery-re...

A review of the original pack teardown reveals it has a small pcb with what looks like some voltage cutoff or thermal fuse.

No the batteries he's referring to (4-6 cell LiPo's) don't have any electronics in them. They are as bare bones as you can get for the racing factor (lightweight and more capacity).

For many people integrating batteries into their projects, this is true and you can simply take most off-the-shelf cells with the right voltage profile and put them into some series/parallel config.

At least with lithium-ion technologies, for larger applications, this ignores a lot of the degradation phenomena and electrochemistry of the battery.

Even off the shelf 9v batteries are typically 6x1.5v AAAA cells in series. (i.e. they actually are batteries)

They can also be stacks of plates though, so buy a sample of a particular brand before you decide bulk-buying them is a great/cheap source of cells!

How does one can get into drone racing? Where should I start?

You buy a cheap controller (RadioMaster TX16S is the best value for money right now, at $120ish, but you can do quite well with a $30 FlySky), some goggles ($40 for some Eachine EV800D, I forget the exact model number), $150ish for a low-end starting 5" quad, and $30 for a 4S battery.

I might also recommend getting a tiny whoop instead, those are very small drones for inside the house, they are super fun and you can fly them whenever you want, vs having to make a trip to a suitable location.

That should get you started. Nowadays I much prefer acrobatic wings, though, I made a Flitetest Versa from Depron foam for like $30 and it's amazing fun.

I'm not having any luck finding the TX16S for that price, any suggestions?

Here is one, though it's a bit more expensive, I think maybe I used points to get the price lower:

https://m.banggood.com/RadioMaster-TX16S-Hall-Sensor-Gimbals...

Don't order from Banggood. It takes weeks to arrive and is likely to be a knockoff. There are several US retailers like RaceDayQuads.com, GetFPV.com, etc.

  very small drones for inside the house, they are super fun and you can fly them whenever you want

House yes, anywhere with a fire suppression system _NO_. The little glass bulbs used in most of them don't react well to having a drone bump into them. Buddy of mine had this happen where he worked not long ago, you really don't want to be that person...

You watch quadmovr [0] and google all the weird words in the video titles :D

0: https://www.youtube.com/watch?v=U6wsa7StHaY

After buying a transmitter/controller, Liftoff (https://store.steampowered.com/app/410340/) is a great simulation to learn how to fly.

Shoot me an email. I have a couple radios (one new and used) I can part with.

Reply on news.ycombinator.com