Paralleling BMS protected Li-ion packs.

[stts]

Yea, Im having nothing but bad luck getting stuff right now. I just got an email that Battery hookup cancelled my battery order. It seems they had less than advertised. Just peachy. Oh well, I got plenty of machining on my bike parts to keep my busy for the time being. Next is angle aluminum to make a box on a hinge. 6 inch by 4 inch and 1/2 inch thick. Ill mount it to the center bolt on the handle bar. 8 inches wide should hold the computer, switches, temperture display, and whatever else I can cram in there with a rubber cushion. I can close and lock it to keep it dry and people out of it. So a vandal cant beat my controls to pieces. Seems people have been tossing the green rental scooters into the river. So we got haters around here. No point in giving easy access to the whizz bang stuff on my bike.

[AZguy]

Yeah I gave up on Ali - they wanted my passport and seemed like everything short of an anal probe

I do a lot of small circuits but have all the tools and usually can quickly identify the limitations of what should be attempted at the small scale - I don't mind spending some time/money on these "hobby" electronics but I have limited time for it

Whatever you choose to do good luck.. if looking for advice I've got plenty of background with this type of stuff... do it just about every day ;-}

[stts]

Well when I saw that circuit I figured everybody would go nuts over it, with all the worries of how batteries can be put together. I been thinking about it more and its not cheap to DIY entire circuits. Especially with the cheap prices the chinese charge for these things. So Ill end up doing something with this. Its just too sweet of an idea. And since BMSs are selling like hot cakes for cheap, Im thinking I could use those cheap for parts. One mos cost as much as some of those BMSs that are full of Nmos's. I may even just tap into a BMS as it is to use its battery cut off circuit. They all cut off the Negative with overcurent and undercharge. I could hyjack the gate drivers and put a small breadboard on top with the right side of that circuit suitably modified to handle higher voltages. That takes a bunch of work and expense out of actually working this out. But right this minute, ALIexpress wont even sell anything to me, so I got that problem to get worked out first. I looked at a bunch of other China websites but Ali seems to be the king of the electronics side. I may have to take a chance and start sending them wire transfers. Market places stateside are just too expensive these days.

[AZguy]

Not sure what you mean by the battery dying but P-ch MOSFET's in the above are disconnecting the bat+ and leaving the bat-'s connected together as a common ground which everything can use as a negative reference

When N-ch's are used typically they are disconnecting the bat- so no longer a common negative reference... not a deal killer but it complicates things to have to use a common positive reference (need to sense both batteries), especially at electric bike voltages, and the above can't handle that

It is possible with charge pumps to use N-ch's in the high side but that's an entirely different drive than what you all have in that circuit (again start from scratch) and will complicate things, especially at electric bike voltages


But hey I'm just an old guy that's been around the block for a long time and know there's plenty of new tricks this old dog can learn

[stts]

I found the circuit thats in those shunt box thingies. Its an Nmos switching out the ground. But only on one battery. The other is running thru the schottky. Not very ideal. But it does go to show that either power or ground is a snap to be switching out. I wont be buying into this circuit for $40. The schottky is wasting the battery. Ill be looking more for a diode free solution. Like in that awesome circuit above. Who knows. Maybe Ill just build it up eventually. An Nmos version. They are way cheaper at the higher currents.

[stts]

It make no sense what you say. When the battery dies, "grounds" are meaningless. That circuit the way it is works outstandingly for low voltage packs. And its a great schematic for making a higher voltage circuit. It could work well by simply changing all the parts to run with higher voltages. And it would certainly work if it was changed to swap the grounds, rather than the power lines. Then both power lines would be common, and only the ground from the best battery would be switched to the output. All at microsecond speeds. The supervising circuits would be hooked to the battery side at all times so they can monitor continuously the pack voltages. Packs used hard and driven below cutoff, would recover while the other pack took over. When the second pack tires, the first pack now has more use to give up. With the monitor circuit switching either the power or the ground, depending on the mosfets that are chosen. I just dont see any downside, other than the heat lost in the power mosfets. And that part is in question depending on the real world Vds the circuit operates within. This will run with one or both batteries in any useful state of charge. Each having different charge voltages. This circuit will use whatever charge there is. Its beautiful. The end of wondering if the packs are charged similar. Deal with it when its more convenient. The only stipulation is that the packs have to be the same voltage range. But they could even be different chemistries. Lead acid for one and lithium for the other. 4 lead acids to best match up to 48v lithium packs. This circuit would manage the power of both till they both are sitting at the cutoff voltage. The lithium would by far hold up longer than the lead acid, but you could use LifePO4 to have more equal performance. Its so hard to go wrong with this circuit. Each battery completely unaware of the other. One could actually get cut off by the BMC, and the other would seamlessly take over to keep you going. You could add charge indicator to the circuit, so you can see if a battery is dead. Right off the reset line, or off a nand gate. Just a red led, then you see on your display how much you have left on the battery with the green light. If the shut box thingy has all this in it, then I likely would add indicator lights to it. Im always modding on circuits to add features.

[AZguy]

There are very good reasons you *never* (and I mean never) see PMOS in electric bike current paths

Breaking the negative side is not as trivial as it may seem since now you lose your reference for voltage sensing with anything remotely resembling that circuit above

I guess I'm just saying that I'm of the notion the circuit above brings next to nothing for electric bike applications... that trying to make it work will take more time than just starting from scratch...

But go ahead and prove me wrong! I'd truly enjoy it

[stts]

Its still not that big a deal. Breaking the positive or the negative side is like flipping a coin. We are not talking about the whole bike circuit. Just whats being presented to the bike at the T60 connector. And they would still be controlled by the tiny Nmos drivers that are driven by the supervisor circuit. Thats easily protected from higher voltages. In fact I wonder if the circuit is just fine with the Pmos drivers. There's hardly no Vds of big size across those devices. The battery is always being pulled down by the load. Its not like it is in a regulated supply. So Vds is always low. So in actuality, I bet low Vds devices could get by fine yielding much lower Rds. Those that build these circuits could plug various Pmos's into simulator software to test that out. Or they could simply CAD up an Nmos version that may be cheaper to build anyway. Nmos being cheaper than Pmos. And it counts when currrent capacity jumps up.

One thing is pretty clear. Those shut box thingys are pretty efficient, all closed up in a plastic box. So efficient that the makers are not worried about the heat output. So its either a big scam going on in there, or a low Vds reality is making for a highly efficient circuit design.

[AZguy]

I think the greatest issue is the P-channel MOSFET's doing the heavy lifting which isn't practical for bike battery voltages and currents

Once you replace them with N-channel the whole rest of the circuit goes out the window for the most part since now the negative side is what's getting broken which is entirely non-trivial

Oddly enough I've got a station with battery back up in my shop right now (just finishing this failure analysis project today ironically) that was toasting batteries and the problem turned out to the the LVCO disconnecting the negative (which is what N-channel's will do) and another connection to battery negative causing the failure of the LVCO to disconnect the battery and over-discharging the battery...


Regardless, these are reasonably simple circuits at the fundamental level... it's just the details that make them challenging...

[stts]

What your talking about is trivial. Higher voltage rated parts are a dime a dozen. The current carrying capacity is where the money will go. And this isnt a regulator, its a switch. So easy dividers can turn this into a much higher voltage monitor. And zenors can protect the low voltage circuits from the high voltages its monitoring. Being a switch, this will switch in the pack with the best voltage. When it drops below the threshold programmed by the divider, this switches the other pack in. As the spent pack does recovery, the monitor circuit resets to be ready for the other pack to trip its programmed under voltage. In this way both packs are used and rested to be able to use both to the last nub of their minimum. All while not causing the motor control to shut down even once until both packs are totally spent to their consistent minimum. Its an awesome circuit. Everything you could want from a parallel pack setup. And the Chinese got nothing better to do than to amp this circuit up to handle beefier bike packs. They already make every kind of buck and boost supply we could ever want, I got no reason to doubt that they already have a circuit like this ready for bike packs. Im not inclined to build this myself. I got tons of other things Im doing. But I can spot what I want in a barrel of circuits, thanks to this cool schematic. Somebody needs to post an opened up shunt box thingy so we can see the insides. And I love the idea of parallel packs. When in doubt of the range of one, slip another one in next to the first. No more doubts. Especially when your in a rush.

[AZguy]

I don't think it is terribly useful for bikes

The back to back MOSFET's are P-channels and would need to be changed to N-channels (at much higher Vds) to get reasonable Rds(on) and everything on the left would need to get swapped around (inverted) to deal with that

The logic is far too low of voltage and looks made to detect only whether not a battery is present and above 4.38V... it would be possible to regulate a lower voltage for the NAND gates and use dividers on the threshold detectors I suppose but there is still the issue that now the logic needs to drive the MOSFET's relative to bat+ and that will throw the voltage detect all out of kilter

I suppose it could be done but personally, I'd just start from scratch and create something optimized for bike batteries


One of the really big challenges with doing something like this for bike batteries is going to be getting all the heat out of the back-to-back MOSFET's

[stts]

Well this exact circuit is not enough for bike packs. But the components can be upgraded to work with bike packs. Its just a reference for the ideal parallel bike pack circuit. With this, you can look at a chinese circuit and tell pretty quick if its doing the same thing. I wish they showed what was in that shunt box thing. With good pics, I can tell if its doing what we need. But it may just be an expensive "Y" connector. Maybe thats why I dont find any open.

[AZguy]

FWIW that circuit doesn't appear to be designed for electric bike batteries

It's for low voltages (even 12V batteries would likely be too much) and would get very hot with just a few amps

[stts]

Now this is a great thread. The world has circuits for everything to be done in every way. And the world has the Chinese that, until recently, have nothing better to do than build every circuit idea they can steal and sell it back to us. Im such a shameful criminal buying their stolen stuff like that. Hah. Anyways, here is a circuit that takes care of all the head aches of all the things we figured could go wrong. It basically keeps both batteries seperate, but keeps the power flowing till both batteries are spent. This is the circuit I want.



Im just wondering if all this stuff is inside that Shunt Box thingy people posted about previously. Has anybody opened the Gizzmo to see whats inside? I see them costing $35 on AliExpress and they are good for 40 amps. Seems like it would have these Mosfets devices and all the goodies to control them.

The thing is, to just hook 2 batteries in parrallel, you get the 2 wasting their energy charging each other. And this circuit prevents that. So you can add a fresh battery next to the half used one and they both get used with no waste at all. So its all brainless and you dont have to keep track of anything. You can even use a boost circuit to crank 36v hoverboard packs to 48v and then plug into this circuit. So both packs are used fully as if you have two 48v packs. Or even use 24v packs, if there any of those out there. American circuits are crazy expensive, but the chinese make them for pennies on the dollar. I remember the car stereo world was so much fun. Getting preamps, amps and boosters and they all hooked together so easy. Now they got compatibility modules that cost as much or more than the head units you want to power up. I cant even run my factory radio in my identical spare car that came with no radio. It says the vin number doesnt match. Hah.

[AZguy]

Not really

A quick look at the numbers in the datasheet and:

15A through 1 diode you will be dumping somewhere around 10-15W (figure 1 in data), 30A closer to at least 25W and dropping voltage at least 0.5V @15A (figure 7) and approaching a full volt at 30A

It will get hot, hot, hot and the heat is just wasted battery power - at 30A the temperature will depend on the heat sink (without one will just fry, they don't even specify the thermal impedance to ambient)