Buck converter 12v to adjustable 3.6V to 4.2V? high current

Announcement - Help us fight the BOTS! Please report all spam including stuff in your inbox!

fungus

New member
Joined
Jul 29, 2021
Messages
9
Hello. I got about 1000 18650 cells to make one 3.6V/4.2V big cell. I want to fast charge them using either a 12V car battery or other multiple 12V power supplies (in the 300W+ range each). I found these noname Chinese made 40A buck converters CCCV around $13-$25 ea. That means I need ballpark two of these buck converters for each power supply. If I use 5 power supplies, that makes already $125-$250 for the 10 buck converters for a measly 1500W charge capability. If I'm going to have to spend that much just to charge the batteries, I wonder if there is a larger amperage CC buck converters available for charging lithium batteries that might be more economical anyone has run across or give me some pointers. Thank you very much.
 

hbpowerwall

Administrator
Joined
Oct 7, 2016
Messages
1,984
I would exercise caution using 'cheap' buck converters around Lithium - Overcharging ends in fire...
 

MBF Dan

Member
Joined
May 14, 2021
Messages
35
What are you going to do with 1000 cells all in parallel? The current draw will be much higher than if you put then in serries groups to increase the voltage. Just wondering because you are looking at either some crazy long run time or some very high current and high current means bigger heat losses and bigger cabling.
I am looking at 3000 cells at 48 volts for a 100 amp load at 5 hours discharge time for a project so 1000 cells in parallel just makes me wonder what current and cabling you are looking at.
 

Korishan

Moderator
Joined
Jan 7, 2017
Messages
6,549
My first question is why are you trying to charge 1000 cells in parallel in the first place??? That voltage can't go very far before it's too low to be usable by many devices.
Second, I agree with HBPowerwall about using the cheap units. Perhaps sharing the link to the ones you are interested in can help us better help you.
Third, as MBF mentions, even if using these are they are, and using a Boost converter to push that power further, the Amp draw will be ridiculously stupid as you scale up. If you need just 100W, on that 4.2V battery, that'll be 24A! 100W is almost nothing for most loads, especially motors.

It would be better to build your battery pack at a much larger voltage, at least 24V (7s) and then use buck converters where you need 5V or so to power those devices. For one thing buck converters are waaaay more efficient than boost converters and if they do fail, they fail open, which means they won't blow stuff up due to a short. A boost converter can fail closed causing high voltage to surge through the system.

If the reason for going for a single large pack to get away from using a BMS, you technically should still use a BMS regardless, of some sort. Because of the size of the pack, you would need to really think about quartering the pack so you can detect flow from section to section to determine if there are any self dischargers. You'd also need a massive array of thermal sensors to detect odd heating.
 

fungus

New member
Joined
Jul 29, 2021
Messages
9
I did not know buck converters fail open. That good to know. Thank you whomever told me. That confirms using buck might be a good solution.

It is just the first unit for prototyping, but later goal make 7 or 14 of these to fully power the house for 3 days. and of course, the fun project to make an EV. Quick charge is an interest for mobile use like wheelchair. I do have access to 3-5KW power. I'm am using NMC cells. 4A charge, 20A discharge. I'm trying to not have to redesign parts as I scale up. I'm trying to parallelize the setup and use readily available parts like 12V chargers. Eventually, a 48V system., but each 1000 cell pack would/should be charged in the same manner. so then I would replicate the charger for each 1000 cell pack. Still that makes a hefty price tag just for charging. Eventually reducing the physical size of all the chargers would become an interest, but initially, just getting this small scale working inexpensively is the project and parallelize it for a larger scale.

The pack is built from 1x6 1s6p w/copper strips. electrically connected 2 strips to a very short double 20A (rated. 12awg) bus (rated 40A capacity), then 36 and 72 cell packs to a much bigger double busbar rated at 80A w/option to go 125 A. I do have accommodations to add extra cells to balance each 72 cell pack so Ah are all identical between packs. I could look for buck converters to charge each 6,12,or 36 cell packs, but I think the price ends up being more since I need more of these smaller buck converters. All the connection points go to the center of the packs to shorten distances within a bus. I would like to find the double wide series/parallel connected strips, but I can only find those in nickel and nickel plated tin.

I've got circuit designs for a parallel to series converters on the output, so I could auto-convert the parallel to series through the circuit board if that is the path I want to take. I could then convert the 1000 cells into 7s or 14s. Problem is with that circuit, I can't charge and use at the same time, so that makes justification to stay with smaller individual buck converters to allow for simultaneous charge while using. Other idea was to use a 2 stage boost converter with MPPT, but now I read that when boosts fail, they could spike in voltage, so I want to rethink that part.

I was going to use thermal FLIR camera to periodically check for problematic cells, but putting some thermal sensors later sounds like a good idea for earlier warning. I'm not there yet. I'm not sure how I would put thermal sensors for that many cells tho. Maybe I'd need to put some aluminum sheets in a strip and rely on thermal conduction along the strip to detect higher temps.

I didn't find the link to the $13 one, but this one seems to look identical as one of those 40A buck converter I was looking at but costs more at $20.
 

fungus

New member
Joined
Jul 29, 2021
Messages
9
What are you going to do with 1000 cells all in parallel? The current draw will be much higher than if you put then in serries groups to increase the voltage. Just wondering because you are looking at either some crazy long run time or some very high current and high current means bigger heat losses and bigger cabling.
I am looking at 3000 cells at 48 volts for a 100 amp load at 5 hours discharge time for a project so 1000 cells in parallel just makes me wonder what current and cabling you are looking at.
I'm looking at solid copper strips to cable, blades to bar. All the DC runs are designed to be real short, then I'm going to grid tied micro inverter at the battery packs to 240VAC for the long runs. There are no long runs using DC. Even between battery packs will be on AC.
 

MBF Dan

Member
Joined
May 14, 2021
Messages
35
For that application I would definitely look at designing for 14S 48 volts. That is the standard for most commercial inverters you would want to power. You could make standard packs of x cells and then put mutual of those in parallel later to upgrade the system. But the power you are talking for a full home you will want to step up the cell voltage on the DC side. Also you will want to think about the fire risk. I am using military surplus ammo cases for my project as they will contain a fire.
Oh I personally wouldn't look at fast charging with used cells and no thermal management. My goal is to design for 5 hours run time 0.2C both for charge and discharge. I don't want to push salvaged cells.
 

fungus

New member
Joined
Jul 29, 2021
Messages
9
I would exercise caution using 'cheap' buck converters around Lithium - Overcharging ends in fire...
Hi. I think I saw one of your videos. I thought I'd reduce the risk of fire and increase efficiency by parallelizing the cells. Less need to be critical on balancing by focusing on regulating the voltage during CCCV charge to never exceed the 4.2V. I am wary of Chinese electronics suck. But an overvoltage protection circuit would be a good idea if I don't have confidence in the Chinese stuff. yea, what a waste of money sometimes getting the bad stuff.
 

Korishan

Moderator
Joined
Jan 7, 2017
Messages
6,549
It would be far better to just go ahead and jump straight to 14s 48V system instead of building a 1000cell pack at 4.2V. Use those 1000 cells to make a 14s string first. Then get another 1000 cells and duplicate it. You can always parallel the strings on the output and share a common charger/inverter.
By making a huge pack and using a charger specifically for that one, then making another one, and another one.... You are basically wasting time, money and energy. Because you'll have to reconfigure the whole system in the end anyways, and you now have all these chargers that you no longer can use on the larger string.

With 1000 cells, you can make a 14s70p (actually comes out to 980 cells, 14s71p is 994 cells). If the cells are 2400mAh, then the 70p will yield 168Ah, and in 14s would give 8.7kWh. Plenty to run most small to medium items.
I didn't find the link to the $13 one, but this one seems to look identical as one of those 40A buck converter I was looking at but costs more at $20.
as far as this is concerned, I wouldn't trust it to run at 40A for longer than a few minutes before it becomes too hot to function properly. It can surges of 40A for a few seconds, maybe. It probably can run safely around 20A without getting too hot. Even with that heatsink/fan on it.
 

fungus

New member
Joined
Jul 29, 2021
Messages
9
It would be far better to just go ahead and jump straight to 14s 48V system instead of building a 1000cell pack at 4.2V. Use those 1000 cells to make a 14s string first. Then get another 1000 cells and duplicate it. You can always parallel the strings on the output and share a common charger/inverter.
By making a huge pack and using a charger specifically for that one, then making another one, and another one.... You are basically wasting time, money and energy. Because you'll have to reconfigure the whole system in the end anyways, and you now have all these chargers that you no longer can use on the larger string.

With 1000 cells, you can make a 14s70p (actually comes out to 980 cells, 14s71p is 994 cells). If the cells are 2400mAh, then the 70p will yield 168Ah, and in 14s would give 8.7kWh. Plenty to run most small to medium items.

as far as this is concerned, I wouldn't trust it to run at 40A for longer than a few minutes before it becomes too hot to function properly. It can surges of 40A for a few seconds, maybe. It probably can run safely around 20A without getting too hot. Even with that heatsink/fan on it.
Good point. If China says it is 40A, design it as a 20A. How many time I bought products from them and their specs were just overinflated.

My original goal was if there was any buck converter available with higher current capacity available that might be a cost savings to me. The reason to not go 14s 48V at each pack is to eliminate the BMS and extra wiring required for every pack. Long run, I want to have variable voltage output or use PWM with some setups. The whole setup is field serviceable with no tools required. Basically the cells are the consumables, and electronics are fixed. Even the idea of being able to swap out the whole battery pack of cells for a freshly charged set while keeping all the electronics stationary. Instead of swapping out the whole cells plus electronics, I wanted to be able to just plug and play cells be it a tiny single 3.7V cell or a large 3.7V cell (for longer runtime) utilizing other technologies to manage power. The idea like replacing only the batteries in a remote control instead of having to replace the whole remote with fixed batteries inside. If I have to eventually have a BMS, I can stack them in parallel to achieve a desired balancing - like a BMS plug in module, but I hoped to minimize the need by bottom balancing the packs beforehand. I can balance the packs to within 0.5Ah. I have space to add extra cells into the packs so all packs are pre-balanced to the same capacity. like the idea of having a car wheel, and then adding lead weights to fine tune the balance.
 

Korishan

Moderator
Joined
Jan 7, 2017
Messages
6,549
As soon as you start to go series connection, you'll need some kind of bms, regardless. Because each pack in the series needs to be balanced against each other. Otherwise they'll get unbalanced eventually. A bms would also keep from over charging the cells, or under discharging them. So no need to have them bottom balancing.
And between the two, top balancing is way more efficient of the usage of the cells than bottom balancing as you can pack in more capacity. There's more capacity in a cell between 3.9-4.1V than there is between 3.2-3.4V, for example. Kind of think of it like a barrel of water with a hole at the bottom. When the barrel is at the top 100% and drain to 90%, the stream coming from the hole is shooting out a good distance and a certain amount of water will flow in a few seconds. But run that same barrel with only 50% down to 40% water and that stream will be much lower and the flow will be slower, and the amount of water given in the same amount of seconds will be significantly less. This is because the pressure is different.

Just something to consider. No everyone does top balancing and some do do bottom balancing. Usually this is done with cells that aren't very well naturally balanced with each other, though.

But the bms is pretty much a must. You can still have your modularity and have everything else too with the bms and equipment. The only issue is that you can't remove 100% packs from the string. But, you couldn't do that anyways no matter the configuration unless you bring the whole system offline.
If you want to be able to service the packs/strings while in service, always plan to have at least 2 modules connected in parallel for the string. That way you can pull one out, and the other can still handle the load. Just make sure that the second pack can handle the amps required while under load during this process. This is a term called quite often, N+1 configuration.
 

fungus

New member
Joined
Jul 29, 2021
Messages
9
The BMS is only be needed in the charge cycle and cells being in parallel will efficiently self balance (unlike those older BMSs that has power loss). I want to get out of the business of having to search for efficient BMSs or knowing what technology the BMS has. My original plan was to use multi stage variable voltage output boost converters or a parallel to series circuit. That won't need a BMS. I still agree is good to add a plug in BMS as an extra safety factor, but not required in this design. the packs are all balanced to the same capacity, the max current the BMS will need to handle will be minimal compared to a 14s system.

Plus the advantage of keeping the electronics outside of the cell packs.

The whole unit will not need to be put offline as long as the grounds in the charger are isolated from each other at critical points -- which is doable, with care and attention. I'm thinking of color coding the packs that have to be isolated from each other's ground - just more work.

I've got ways to make the cells hot swappable. Plug and play.

My real interest is to see if I can find a higher current Buck converter as in the title of the thread. not the BMS, but thank you for that info.
 

Korishan

Moderator
Joined
Jan 7, 2017
Messages
6,549
the max current the BMS will need to handle will be minimal compared to a 14s system.
On the contrary, the current required would be much higher. On a 14s system to power 500w load, you'd only need 9.6A @ 52V. To power a 500W load on 1s, you would need 131A @ 3.8V!! Essentially, you would need 14 times the current to go with 1s compared to 14s.
The lower the voltage, the higher the amps. It's a law, and it can't be avoided or ignored.
Sure, you could have a lot of wires hanging off this pack to power a lot of boost converters, but that would be a nightmare to troubleshoot, increase chance of shorting something out, and wire management would be a pita.

The BMS is only be needed in the charge cycle
The bms needs to monitor not just charging, but also discharging. Because you want the bms to disconnect the battery before it gets too low.

My real interest is to see if I can find a higher current Buck converter
Once you go over about 30A for a boost converter, you are starting to get into drastic diminishing returns. The cost of the units go up drastically compared to the output current. 3.8V * 30A = is only 114W.

If this is a hobby build and are quite aware of the how much this is going to cost you, by all means continue forward. But it doesn't seem like it when you state your first pack is 1000cells and plan to add more. You will quite literally be spending at least 5 times the money to build a 1s setup with nothing but boost converter(s). A decent BMS that can do 7s or 14s will be in the range of around $100-200USD. For just the boost converters you are looking at well over $300USD to be able to get a good amount of power available. And then there's the heavy gauge wiring that must be used with low voltage/high current.

I'm sure a lot other members here will agree, you really need to re-evaluate the build, materials, and more importantly cost and time of this build. Sure, it can be done. But at what cost, and compared to more practical ways? Can you fly from LA to NY in a bike-plane? Sure can. Is it practical? Nope.
 

fungus

New member
Joined
Jul 29, 2021
Messages
9
Yes. I understand less voltage is more current. That is the law. I'm not avoiding it. It is within my specs. I'm using straight copper. Not tin or brass or nickel. The 3D orientation of the cells allows for shortest dual DC paths available and the busbars can handle the currents. I'm not concerned about current. I'm concerned about power.

Power out is the same. Irrespective of 1s or 14s.

I've got low voltage shutoff handled so I don't need a BMS to do it for me. So there is still no need.

The question is about Buck converters, not boost converters. This thread is asking about low cost high current buck converters. 12V to variable 4.2/3.6V. If you are aware of any, I'd appreciate a link.
 

Korishan

Moderator
Joined
Jan 7, 2017
Messages
6,549
I didn't find the link to the $13 one, but this one seems to look identical as one of those 40A buck converter I was looking at but costs more at $20.
I doubt you will find a decent high current buck converter for less than $20, or even less than $50. And I got confused as to what you were looking for because you posted a link to the Boost Converter in that same post, not a Buck Converter.
 

fungus

New member
Joined
Jul 29, 2021
Messages
9
I doubt you will find a decent high current buck converter for less than $20, or even less than $50. And I got confused as to what you were looking for because you posted a link to the Boost Converter in that same post, not a Buck Converter.
I had mentioned I couldn't find the link but the buck converter looked like the one in the picture. Buck is in the thread title.

The target is less than $125 and becomes interesting. If in the $125 to 250 range is something still consider, but less desirable.

It becomes economically feasible since there are only going to be a small number of chargers and they are independent of the cell packs and the number of cell packs will scale rapidly. The current charger in the $125-150 range is too small too slow. for 1000 cells

I basically need to charge a large number of cells 1s. Buck is the cheapest soo far at small scale. It parallizes 20 buck converters using adapters to the 36/72/144 cell pack. It works fine except for the low power long charge times. I thought instead of buying another set of small chargers, I could reduce the volume of the charger by using a bigger capacity buck converter to charge each 36/72/144 cells (~100/200/400 Ah). And like someone mentioned, less wiring mess adapters needed. It was a one time setup, no problems, no debugging, works fine tho.
 
Top