Why isn't "node charging" through the balance leads more common?

fatty

10 kW
Joined
Nov 14, 2020
Messages
672
Location
USA
-The main pack leads could be hardwired straight to the controller, with no additional parallel connector hanging off.
-Only the balance plug would need to be plugged in to charge, not pack leads + balance.
-Every cell or parallel string would be independently top balanced on every charge for optimal capacity and longevity, reducing the need for failure-prone embedded BMS.

The commercial node chargers, BC168/CB86, UNRC UNA6/9 are all discontinued.

Why?
 
One of the custom builds wired up a custom harness with larger wire and charged each p-string with an individual 4.1V charger. He used a Pelican suitcase full of small chargers.

Teklektic 2WD Yuba Mundo?
 
spinningmagnets said:
One of the custom builds wired up a custom harness with larger wire and charged each p-string with an individual 4.1V charger. He used a Pelican suitcase full of small chargers.

Teklektic 2WD Yuba Mundo?

I remember reading that -- it was for LFP and he had to stagger the chargers to prevent overheating.
Certainly not an elegant solution, but that was many years ago, if I remember right.
 
I think Hyena may have done something like this on one of his bikes. JonesCG may also have experimented with this. But it was quite a while back. There was some discussion in one of the Fechter-Goodrum BMS threads about it, but I don't recall which one. It might even have been the beginnings of the design you've already found.


One issue is finding anything other than a bunch of individual PSUs to do this with, or else custom-designing a PCB (or set of) with that many electrically-isolated PSUs on it.

Another issue is getting *all* of the channel-chargers to be exactly the same, and to stay that way. Components do drift in value over time, so most of the really simple PSUs that do CC / CV will drift at least a little bit in end voltage (and current ability) over time, and require periodic checking (the older they get, the more frequent the checking), or else you may end up with packs that are no longer perfectly balanced, if that matters. I suspect this is part of why you don't find many COTS stuff that has done this (and why I don't see any available nowadays). If the PSUs don't have adjustments, then when they get far enough out of "true" you have to replace them to maintain that perfect balancing.

You should also check at least periodically for PSU failure--sometimes things (capacitors mostly) fail and leave the PSU in a state that allows too-high voltage on it's output, and it could kill cells (or even start a fire).

If you could find (or design) a multichannel charger that monitored all this for you, that'd be easier...but I don't know of one.

A further issue is an inexpensive, durable, reliable connector with enough pins but that is small enough to be reasonable, but can still handle the high current charge levels required. IIRC that was a big part of one of those threads I mentioned above. I think they went with a large D-Sub type of connector, but I don't remember how well it worked over time.


If you use well-matched cells of sufficient capability and quality that are *designed for EV type usage scenarios* to be much more than capable of the loads and conditions they'll see, then you may not really need a BMS *or* balancing via whatever method.

Otherwise...you need something to monitor under load, that can cut off discharge if a cell goes too low, if you want to ensure no low-voltage excursions due to whatever problem might arise in the pack. (Even if you don't have anything monitoring for charging (and just assume that none of the PSUs will ever fail in a way that allows too high a voltage)).


My EIG pack is one of those well-matched packs, and used wel-within it's capabilities, so far, but as it ages, cells are going to drift in capability, and it wont' remain balanced on it's own after that, and a BMS would be a good idea, at least for monitoring during pack usage, even if it is not always connected.

So what I've been pondering is instead of doing what you'd like to do, and instead of running completely BMSless and increasing my personal pack monitoring (whcih I don't really have enough time and energy for), I'd like to connectorize a BMS so it can be connected to the pack only when I am charging it, or when I am discharging it. Ideally I'd like to automate that with relays or something, but those are themselves sometimes unreliable, and/or may have significantly different timing from each other whcih might cause problems with the BMS. So using a connector that has different length contacts for certain things to ensure it all "powers up" in the right order (for that specific BMS) is the best compromise I've come up with so far.
 
Why not? cause its dangerous, simple answer.

Little more complicated answer, it would be a super long time to charge any good pack at the amperage that connectors can handle without overbuilding, reliably, and safely. Do not try to shove a 1C charge on a 10Ah cell down a 24g wire.. you WILL burn something.


If you are happy with teh < 5Amps that a 20g wire through a connector will give you... Yes YOu can charge one of my packs like this. I make lil terminal blocks for this.

However, my Powerlabs does not support balance wire charging officially. Maine Charge/Discharge leads must be hooked up.

I use Dsub headers and mil spec crimps and a matching terminal block on the batt/charger.

My Dsub can take 5 Amps reliably per pin, weatherproof, cheapo, repeatable, have won wars with those Cannons. Up to over 50pins. I use the 9, 15, and 25, 37, alot. I have them all here, and back shells.

5A is all Im gonna push with a 20g wire.

I have 13A Molex in the works too. However that increases cost. The Cannons I can get for like 25-30$ for 100 pins and the shells.

Higher quality than the Molex too, but not if you buy retail.. then 3-4x the price of the Molex stuff.

Good luck charging through junk silicone 24g. Danger, ranger. Get hot on an amp. Then still take 15 hours to charge your 15Ah pack. / cell..

I can charge up to ( how ever many the charger can do) 8s at a time like this, even thought I have ganged up 22s powerlabs, you have to do it one set at a time or else the channels are not isolated from each other and you may blow the charger. Im still working on the logistics and implementation. Each set of 8 or less in the charge cycles is three hours at 5A for my 15Ah pack.

But yea I got some of this inthe works. I definitely have a terminal block balance connector, connected to each pack that is much stronger than the JST i used to build with. Then plug the harness into that for the balance assy.


24 0.51 0.020 0.20 88 133 3.5 2 1.6 1.4 1.2 1.0
22 0.64 0.025 0.33 52 80 5.0 3 2.4 2.1 1.8 1.5
20 0.81 0.032 0.50 34 53 6.0 5 4.0 3.5 3.0 2.5
 
DogDipstick said:
Why not? cause its dangerous, simple answer.

Little more complicated answer, it would be a super long time to charge any good pack at the amperage that connectors can handle without overbuilding, reliably, and safely. Do not try to shove a 1C charge on a 10Ah cell down a 24g wire.. you WILL burn something.

Sure, good point. But plenty modify packs or build our own packs, and (up)sizing the balance leads correctly is trivial.
It's a trivial implementation challenge vs what can be a fundamentally advantageous method of balance charging (depending on requirements and system design).

I'm just surprised that despite the ubiquity of battery packs in enthusiast use (across applications), node charging never took off.
 
I expect that is partly because even "enthusiasts" may not want to actually deal with anything more than just plugging the battery in and letting it do it's thing. This *can* work with individual group charging (all groups simultaneously), but the issues with that are noted previously....so it's "easier" to just use a common battery that has a BMS on there to do all the "work" and then just plug in a "charger" to the main connection.

And since the typical BMS/bulk charger is what is already sold as a unit, most people will use that, and unless they reach a problem point with it they are unlikely to even put any thought at all into it, much less change it. Since it would take effort to find and setup a different system, most people won't do that. ;)

If the individual group charging had been commercially developed along with monitoring/shutdown boards for overdischarge protection, and as widely available and inexpensive as the typical balancing BMS and bulk charger, then there'd likely be a lot of people using that method, simply because it is what "came with their battery".


Oh, one other set of problems I forgot to add previously:

--if a PSU should fail shorted, or in any way that no longer provides voltage to the group and then can drain power from it, it could kill a group just like a failed BMS balance channel can.

--If a PSU should simply fail to output, and not charge the group (or to a voltage below the ohters), an imbalanced pack has now been created, but without any kind of monitoring board in place, there's nothing to tell you this happened, and also nothing to prevent overdischarge of that group once it's being discharged without cell-level monitoring. If the pack was already empty when put on charge, and the PSU on a group didn't charge it at all, then you'd almost immediately begin reverse-charging that group, which is one of the things that is very very bad to do to a cell. The results of that vary but can definitely include fire.
 
amberwolf said:
I think Hyena may have done something like this on one of his bikes.
If you could find (or design) a multichannel charger that monitored all this for you, that'd be easier...but I don't know of one.
...
If the individual group charging had been commercially developed along with monitoring/shutdown boards for overdischarge protection, and as widely available and inexpensive as the typical balancing BMS and bulk charger, then there'd likely be a lot of people using that method

This came up in a previous discussion and I was going to wait until I had first-hand experience, but Hyena's was the BC186, a 6s x 8A/each node charger which can now charge to 4.35V.

His thread also gave the Turnigy P606 as a "successor" to the BC168, although that was a small 50W 6s charger.

The RadioLink CB86Plus is a 6s x 6A/each x 8port node charger and can charge to 4.23V, but packs larger than 6s must be separated without any external electrical connection. otherwise "the charger or batteries will get burnt."

The UNRC UNA6 and UNA9 are 6s and 9s x 8A/each, but only the 6 is available, and only via Aliexpress. It's not clear if the more recent Plus and Plus+ models can charge higher than 4.2V.
 
I would say just from a reliability standpoint, a single bulk charger plus BMS is an order of magnitude more reliable than an individual charger per cell group (I.e. node charging).

A charger is a relatively complex circuit with many points of failure, whereas a BMS usually has relatively few components *per cell group* (couple of resistors/capacitors and a transistor per channel).

Also WRT accuracy, usually a BMS will use 1 (or 2 for larger series connections) ICs to monitor and switch the cell group balance circuits. The voltages are usually measured via multiplexing through a single ADC, so relative accuracy of voltage measurement between cell groups is accurate, repeatable, and stable.

A single charger for each group introduces greater errors in terms of measurement (unless each charger circuit output is controlled by feedback by a single brain), more effort to calibrate each channel, and more environmentally effected hardware to drift out of spec.
 
amberwolf said:
I expect that is partly because even "enthusiasts" may not want to actually deal with anything more than just plugging the battery in and letting it do it's thing. This *can* work with individual group charging (all groups simultaneously), but the issues with that are noted previously....so it's "easier" to just use a common battery that has a BMS on there to do all the "work" and then just plug in a "charger" to the main connection.

Well, I guess I'm thinking more of the members here, some who spend literally many thousands of hours and dollars on the hobby, designing and/or testing components, building frames and/or packs from scratch And I'm sure the same occurs in the many times larger RC hobby. In these cases, plugging in 3 balance plugs rather than disconnecting the main discharge plug seems.. trivial.

Not saying node charging is appealing to many, but those running without a BMS (especially easily-imbalanced RC LiPo), would seem to be a large enough market to get better than a $50 failure-prone node charger. Although I guess we could say that about the larger high-voltage charger market as well...
 
serious_sam said:
I would say just from a reliability standpoint, a single bulk charger plus BMS is an order of magnitude more reliable than an individual charger per cell group (I.e. node charging).
I don't agree -- every second or third thread in this subforum is for a misbehaving (or incorrectly-spec'd) charger or failed BMS.
However, I do get your point...

amberwolf said:
--if a PSU should fail shorted, or in any way that no longer provides voltage to the group and then can drain power from it, it could kill a group just like a failed BMS balance channel can.

--If a PSU should simply fail to output, and not charge the group (or to a voltage below the ohters), an imbalanced pack has now been created, but without any kind of monitoring board in place, there's nothing to tell you this happened, and also nothing to prevent overdischarge of that group once it's being discharged without cell-level monitoring. If the pack was already empty when put on charge, and the PSU on a group didn't charge it at all, then you'd almost immediately begin reverse-charging that group, which is one of the things that is very very bad to do to a cell. The results of that vary but can definitely include fire.

serious_sam said:
A charger is a relatively complex circuit with many points of failure, whereas a BMS usually has relatively few components *per cell group* (couple of resistors/capacitors and a transistor per channel).

Also WRT accuracy, usually a BMS will use 1 (or 2 for larger series connections) ICs to monitor and switch the cell group balance circuits. The voltages are usually measured via multiplexing through a single ADC, so relative accuracy of voltage measurement between cell groups is accurate, repeatable, and stable.

A single charger for each group introduces greater errors in terms of measurement (unless each charger circuit output is controlled by feedback by a single brain), more effort to calibrate each channel, and more environmentally effected hardware to drift out of spec.

I think these points get to the real reason: not that a node charger can't be reliable, but that it can't be reliable while using the questionable design, garbage components, and amateur assembly that characterizes the rest of the hobby market. We can (and do) say the same thing about bulk chargers and BMSs, but as you say, with simpler design and far fewer components, it takes a bit longer for those problems to manifest.
 
fatty said:
every second or third thread in this subforum is for a misbehaving (or incorrectly-spec'd) charger or failed BMS.
Please stop exaggerating and perpetuating the BMS myth. There's no particular reason that BMSs should be any less reliable than any other electronic hardware. Think about it logically. For every 10 "failed BMS" stories, I'd bet 9 are user error, or just cheap shit cells or packs, prone to failure anyway. Or like you said, incorrectly spec'd. After all, most people have a problem with jumping to conclusions to suit their preconceived notions. 1 out of the 10 is maybe a legit failure. And for that 1 failure there's another million BMSs in successful silent operation, just doing their thing, making no news.

Personally, I wouldn't use a BMS that is not programmable/bluetooth. You get a bluetooth BMS and you know what it's doing and when. And you can connect from your phone and keep an eye on cell levels and events. At least that way you're not blindly relying on some hardware without knowing if/how/when it's actually BMSing.
 
Be lots of 12g wire connecting each Parallel string to charge. Makes things a bit bulkier in terms of size.

You'd have to unplug each P string for charging.
10S3P
@ is (nothing/empty) is only there to line up negative conductor connecting bus.
+ 1/2/3/4/5/6/7/8/9/10 -
+ @@@@@@@@ -
+ 1/2/3/4/5/6/7/8/9/10 -
+ @@@@@@@@ -
+ 1/2/3/4/5/6/7/8/9/10 -
+@@@@@@@@@-
^main + connector @@^main - connector

If your charging a string of 10S3P, then each 3P would have to be disconnected in some way to charge. Disconnect every P so they are isolated from each other. Voltages need to be the same before reconnecting in parallel. Voltages must be the same when connecting in parallel otherwise you get a surge from one parallel group to another.

However if one cell in a 10S1P string is lower/higher then the other 9 series cells, then its best to make it the same as soon as possible. If that cell is weaker then the others, then it might be best to get rid of it and replace it with a healthier cell, lots and lots of work if its 18650 in holders and tab welded. Different job completely if its Lipo pouches like Hobbyking.
 
markz said:
Be lots of 12g wire connecting each Parallel string to charge. Makes things a bit bulkier in terms of size.
Nah, the chemical-resistant wire I like to use is rated at 8A for 20g.
The bigger problem is the JST connectors are only rated for 3A, and certainly not for the mating cycle count, but it seems fairly common to hardwire the leads directly to the charger board, as well as to replace the battery-side connector.

markz said:
You'd have to unplug each P string for charging.
If your charging a string of 10S3P, then each 3P would have to be disconnected in some way to charge. Disconnect every P so they are isolated from each other. Voltages need to be the same before reconnecting in parallel. Voltages must be the same when connecting in parallel otherwise you get a surge from one parallel group to another.
This is not at all required with node chargers -- they are internally isolated. See the product links above.
 
fatty said:
markz said:
Be lots of 12g wire connecting each Parallel string to charge. Makes things a bit bulkier in terms of size.
Nah, the chemical-resistant wire I like to use is rated at 8A for 20g.
The bigger problem is the JST connectors almost certainly rated for that current nor plug cycle count, but it seems fairly common to hardware the leads directly to the charger board, as well as to replace the battery-side connector.

markz said:
You'd have to unplug each P string for charging.
If your charging a string of 10S3P, then each 3P would have to be disconnected in some way to charge. Disconnect every P so they are isolated from each other. Voltages need to be the same before reconnecting in parallel. Voltages must be the same when connecting in parallel otherwise you get a surge from one parallel group to another.
This is not at all required with node chargers -- they are internally isolated. See the product links above.

I did not know that about node chargers, I learn something new. I was just thinking of how I would try to charge each string with a regular charger. I would figure out the amps I would want to charge at and find corresponding wire gauge then probably go down one gauge. Then figure out the interconnecting connector I'd use to parallel each string back to a normal pack. When I was putting a pack together I'd put each string together then charge each string individually to the max voltage of my charger so I didnt have to pay to close attention to voltage #'s, just let it max out to the appropriate voltage. Then triple check voltages when I was soldering them in parallel. I did something similar with the 5S2P Makita's, except charge each brick (5S2P) to the same voltage then connect together. I'd buy solid core wire from Home Depot, house wiring had 3 conductors per cable, I'd chose the one with 10g inside, then usually one wire would be smaller 12g.
 
fatty said:
Nah, the chemical-resistant wire I like to use is rated at 8A for 20g.

" Rated". I cannot believe that "BC168 Balance Charger" shoves 8000mA through 24g wire. thats craxy . Lipo even. How idd they not burn errytag? It does tho.


Now yall nknow I cut corners... try this, burn that, figure it out. I am telling ya sub par wire gauge in a a charge system is not good.

Could you link me some of this 20g wire that takes 8 amps reliably? I would not feel safe doing 8A through connectors with 20g wire.

I can only find typical 20g. Like I said earlier. I am good only shoving 5A through that 20g. I dont like shoving more through. The connectors I have are designed for 5A max, and even if you could manage 8A, that is still a minuscule charge IMO.

( literally 1/2mm^2 wire) .

Single core is listed 6A. Solid. Less with more strands. @ 43 strands of more, this moves down to 2.5A on some charts.

(https://www.engineeringtoolbox.com/wire-gauges-d_419.html)

Some charts dont recommend going over 1.5A in " 20g Bundled power transmission wires" ( however, chassis wire, not 100% DC, can be more)
(https://www.powerstream.com/Wire_Size.htm)

5 Amps or else you are getting hot. Lemme tell you, you dont want hot wires in a battery.
(https://en.wikipedia.org/wiki/American_wire_gauge)

Yes all these "ratings" are based on certain KcmIl / A. You can bend and fold the "rules" a little but not much and it will catch up wi you eventually. I expect my cut short corners to fall on me some day. For sure. I chose 5A / 20g cause I have seen alot of DIY kits that bundle power wires in to DSub style pins... and all those experts say a 20g on a typical pin takes 5A.


Here is a pic of my "Pole Access System" that I have been trying out. It is nice to have a easy view of the packs voltages. Plus charge at will @ desired current. Please critique. BMS shares a harness that brings them all together in 21pins to BMS, Monitor, problem? Disconnect and have clean access to the cell voltages to know the problem or not. This three plug harness is not shown. That my lil terminal access block. Please critique.

I do have some pcbs that hold cells that I was thinking about drastically modifying. Lol. Bars and plates. Still, I can only follow what I see in OEM, and for some reason ALL the OEM use a certain gauge for the balance system.. I I decide tehy may be right and I dont really try to mess with that. IF Chevy feels 20g is ok for their pack BMS, I copy as seen fit, for I am uneducated.

You see alot of 20g in OEM balance harness... for 17Kwh-60Kwh-90Kwh packs. So they must be some kind of usable. That gauge. Either that or they are not able to srtuff more through typical quality connectors at the rates that would not compromise cost.
 

Attachments

  • BATT3.jpg
    BATT3.jpg
    75.6 KB · Views: 982
  • BATT2.jpg
    BATT2.jpg
    82.2 KB · Views: 982
DogDipstick said:
" Rated". I cannot believe that "BC168 Balance Charger" shoves 8000mA through 24g wire. thats craxy . Lipo even. How idd they not burn errytag? It does tho.
Definitely crazy. And they sure do burn :(

DogDipstick said:
Could you link me some of this 20g wire that takes 8 amps reliably? I would not feel safe doing 8A through connectors with 20g wire.

https://www.mcmaster.com/7304K12/
Chemical and abrasion resistant, flexible, and easiest wire I've ever soldered. Not saying I would cut it right at the spec with 20g, but 18g is rated for 12A and would be fine for short leads on the pack itself. Certainly not 12g.

Note that these ratings are for insulation integrity only -- everything else (voltage drop) is application-specific.
 
DogDipstick said:
Here is a pic of my "Pole Access System" that I have been trying out. It is nice to have a easy view of the packs voltages. Plus charge at will @ desired current. Please critique. BMS shares a harness that brings them all together in 21pins to BMS, Monitor, problem? Disconnect and have clean access to the cell voltages to know the problem or not. This three plug harness is not shown. That my lil terminal access block. Please critique.
So the BMS connects through the 21pin DSub? Wasn't amberwolf looking for something like that? Interestingly, not to keep with the pack full-time, but rather to charge through?

I like the replacement of JST with DSub. Being shielded for signals, it still probably isn't ideal, but definitely a step in the right direction.
Based on familiarity, I'd use an automotive connector, something from the Delphi (now Aptiv) GT 150 family, rated 15A.

DogDipstick said:
Still, I can only follow what I see in OEM, and for some reason ALL the OEM use a certain gauge for the balance system.. I I decide tehy may be right and I dont really try to mess with that. IF Chevy feels 20g is ok for their pack BMS, I copy as seen fit, for I am uneducated.

You see alot of 20g in OEM balance harness... for 17Kwh-60Kwh-90Kwh packs. So they must be some kind of usable. That gauge. Either that or they are not able to srtuff more through typical quality connectors at the rates that would not compromise cost.
Proper automotive packs are better matched (and stay matched), and they can balance continuously rather than only while charging, so probably just no need to waste the cost or weight.

Those are very nice packs. Care to detail the design and construction?
 
Regarding 20AWG wire current capacity, it can be given as around 10A, depending on the information source. One chart, as example:
https://www.powerstream.com/Wire_Size.htm
However, that chart gives different current ratings for use as "chassis wiring" (11A), and another for use as "power transmission" (1.5A). I didn't look to see how they define each of those.


Regarding the D-Sub and shielding: It *can* be used as a shielded connector, but it is often not used that way. It depends on the specific cabling used. However, if you are using it in a system with a lot of electrical noise (like if the motor was running), it might be good practice to use that shielding at least for the outer cable jacket, to help reduce induced noise in the balance wires that may reduce momentary accuracy of readings.

Depends on impedance of the circuits in question, which in this case should be very low at the battery end, but very high at the monitoring end. (the higher the current flow in a wire is, the less other small induced currents will affect the voltage on that wire, but there should be virtually zero current in the balance wires when only using them for monitoring).
 
amberwolf said:
Regarding 20AWG wire current capacity,


Regarding the D-Sub and shielding:

Depends on impedance of the circuits in question, which ing them for monitoring).

I like to think of a wire in "Kcmil" and "Kcmil/Amp", ( not generic charts), and the deduction of the capability through "adiabatic testing".

No need for shielding for what my limited knowledge would entail: there are no signal wires there, just power wires. I aint hitting a baby ( sensitive circuitry) with a stick, It (battery) is the stick. I understand the notion that no current travels there: I digress, for he balance wire does the current transfer? No? Lemme get back to this. I gotta go ATM. I do have more to say.

I would shield the sensitive wires from the power, not the power from the sensitive. I ain't paying for fancy Dsubs in shiny metal shells for shielding. nosiree.

Imma gonna make a me a ESR meter from a DIY chinese kit,: I do want to know this figure.

IDk.
 
amberwolf said:
Regarding 20AWG wire current capacity, it can be given as around 10A, depending on the information source. One chart, as example:
https://www.powerstream.com/Wire_Size.htm
However, that chart gives different current ratings for use as "chassis wiring" (11A), and another for use as "power transmission" (1.5A). I didn't look to see how they define each of those.
Well, those charts are better than nothing if you have unknown wire and nothing else to go on, but as above, it's better to use wire specifically rated according to insulation material -- then there's no question. This also reduces the chance of getting mislabeled undersize wire -- I got cheap 22AWG-labeled Chinese ebay wire that was actually 24g (Google search shows this is quite common).

amberwolf said:
Regarding the D-Sub and shielding: It *can* be used as a shielded connector, but it is often not used that way. It depends on the specific cabling used. However, if you are using it in a system with a lot of electrical noise (like if the motor was running), it might be good practice to use that shielding at least for the outer cable jacket, to help reduce induced noise in the balance wires that may reduce momentary accuracy of readings.

Depends on impedance of the circuits in question, which in this case should be very low at the battery end, but very high at the monitoring end. (the higher the current flow in a wire is, the less other small induced currents will affect the voltage on that wire, but there should be virtually zero current in the balance wires when only using them for monitoring).
This is a good general explanation, but not really relevant for that application. Shielded connectors and cable wouldn't do anything by themselves -- the EM the shield picks up has to go somewhere. BMSs have no provision for connecting the shield, and as we know, charger grounds are not isolated, so you'd be dumping that EM right back into the negative conductor.

Also, there are no temporal signals on balance leads, so to interfere with the BMS/charger, you'd actually have to induce a voltage into the balance leads -- not a problem unless you wrap your balance leads in phase wires, haha.
 
As we discuss implementation, it's becoming clear why node charging isn't more common:
At the very least, the 3A JST balance connectors would have to be replaced.
Even if the chargers were hardwired with 8A-rated conductors (they aren't), the battery-side balance leads all the way back to the cells should also be upgraded.
Building the charger right would be expensive, so.. they're not built right.

Frankly, selling these to the general public with 24AWG leads and JST connectors is a liability. No wonder node charging hasn't taken off...
 
BMS's are the cause of a pack death "often enough" that it's a reasonable risk to run a battery pack with no BMS.

If someone was interested in that, I would suggest..

1) Use a bigger pack than needed.

2) bulk charge to 4.0V per cell, and drain to 3.3V per cell. Even with an occasional unbalanced condition, no single cell would go below 3.0V, or above 4.2V

3) assemble a balance charge system like Teklektik. A 14S pack would have 14 small 4.1V chargers to individually charge each cell (or possibly only 7 chargers, with two separate 7S harnesses on the pack)

4) if you add 14 tiny cheap digital voltage readouts, you can clearly see when you need to take the time to do a "balance charge". It might only be once a month...
 
DogDipstick said:
Yes all these "ratings" are based on certain KcmIl / A. You can bend and fold the "rules" a little but not much and it will catch up wi you eventually. I expect my cut short corners to fall on me some day. For sure. I chose 5A / 20g cause I have seen alot of DIY kits that bundle power wires in to DSub style pins... and all those experts say a 20g on a typical pin takes 5A.

To add, here is additional discussion on 45A through DSub-15 (3A each) showing somewhat unfavorable contact resistance and temp rise. And I'd want to put 48A through a 7pin and 144A through a 21 pin...

Although DSub from reputable manufacturers is rated at 5A/pin, I'd suggest de-rating any Chinese eBay/Amazon connectors. And better yet, consider dedicated automotive-type power connectors.
 
fatty said:


Although DSub from reputable manufacturers is rated at 5A/pin, I'd suggest de-rating any Chinese eBay/Amazon connectors. And better yet, consider dedicated automotive-type power connectors.

I am an amateur flingDinger. Yes I watch any "system" I make like a hawk for problems.

My Dsub backshells are NOT shielded at all. They are just plastic.

I do worry myself sometimes, not being formally trained, if I am doing something Danger-Ranger. I have only ever seen power transmitted through Dsub with DIY kits, not the ( Tv, computer) low power applications you see in production... but, I did ask an Electrical genius, in person, and he said it was OK. I do know a few of those. Guys.

That batt ( above) is a 30Ah 72v pack built with Chevy Volt cells. IR is 0.0008mOh / group (2p) on the datalog. Tested in at 30.5Ah. Aluminum angle and banded for corner protection, and airspace. 100% copper. It is split into two 10s 30Ah w/ 6g wire in between ( main discharge/charge). That battery is gonna last a VERY long time.

Too bad it was stolen off a porch right it arrived at the destination I shipped it to. Lol. Bah Hummbug.

Above, I said, I begin ( just got a kit) use Molex ML-XT ( DISCONTIN). (13A.) 18g wire ( can take up to 14g) in the plans. MX150 is another i have considered. Molex Picoblade is better than JST.

Also, the GMWeatherpack line, I can get a 788~ pc kit for about 140$ ? I think. DIYAutotune.com is the best place for those, those guys are great. I have a 49' 3100 and a '55 Apache I am "contracted " to building the harness for now. https://www.diyautotune.com/product/755-piece-weather-pack-kit/

I really want to know the name of those LIL TINY POSITIVE INTERLOCING WEATHERPROOF plugs on the BOSCHE bikes..... they are tiny and cool.. very solid... and I have no idea as to the name of the product line.

Those "Aptiv GT 150 Series 7 Position" are expensive AF and not common.. Lol @ 7$ for a 7 Pin connector.. I need a system that has every piece, every pin config,available for much less than that. This is why I look for discontinued lines of connection products. Cheapo.

Cannon does make ( called high power cannons, ComboD) Double density, high density, Combo-D, all those. They do make weatherproof too. All kinds of BOB terminal headers are available, Im just like a big kid playing with legos.

Plus the Cannon / AMP/ Tyco dsubs cause they cheap. Everywhere. Kinda foolproof.

I personally cannot really tell the difference tween 24g and 26g.. by eye. Have to read the print. Kinda scary thinking that might be fake too.

So I got a ton of the DB25. Convenient for the size I need. @ 20s, that is max 5A to the battery.....@ 5A / line,. ... Through all those tiny wires, all acting like tiny resisters. I do say, when you desolder the junk JST from a BMS and resolder in your own, positive interlocking, reliably connected 200-500 (rated) connection cycles... +.. harness, You do feel better ( than the junk JST there alone).

(https://www.mcmaster.com/7304K12/) ..That IS the wire in my pic. I get it for 4$ / lb.
 
Back
Top