I am building a LiPo 7S bike battery charger. Rather than taking the conventional approach of charing the bank from its two ends and bleeding individual cell banks if they rise above the average, I want to charge all cell banks individually with a TP4056 or similar charge controller.
I am providing isolated/floating 5V for each controller by using 7 Wall wart USB chargers. This means I can connect them in series without shorting anything.
This works but - because of the large almost-flat voltage part of the charge curve - the charge controller is extremely sensitive to the sensed battery voltage and will throttle back the charging current greatly if even a tiny rise in batter voltage is detected. This is of course, doing what it's meant to do. However, if I try to charge through the balance charging leads, which are thin and high resistance, the controller takes forever trickling in charge, because these balance leads drop significant voltage for reasonable charge currents. They were never meant to carry charge current - only to sense voltage at no current - so again, all is as it should be. So I want to replace them with beefier low-resistance cables which can feed higher charging currents without dropping voltage - so the charge controller can sense each cell bank's charge state more accurately and charge as fast as possible.
BUT this approach is dangerous. If the chargfe controller fails short circuit then a cell bank is shorted out and the thicker wires will not blow so the whole battery could burn.
Fuses would be the obvious approach - a fuse in each charge line. The problem is that these drop significant voltage so we're back with very slow charging.
I suppose I could fit Hall effect sensors and relays and add a microcontroller to sense the current and turn off the relay if current becomes too high (or probably hold ON the relay as long as current is NOT too high) but this adds a great deal of complexity.
So my question is - is there an easier/cheaper/better way to provide this protection?
Any other comments on this proposal are also welcome.
ChrisXenon:
Fuses would be the obvious approach - a fuse in each charge line. The problem is that these drop significant voltage so we're back with very slow charging.
Interesting fuses you have. A normal fuse has a negligible voltage drop, until they blow.
Use 7 fuses. Clean the ends of each fuse and the fuse holders. Resistance should be much less than you can measure with a regular multimeter, maybe 50 milliohms for the fuse itself and 50 milliohms for the connections.
Design and build 7 copies of a MOSFET battery-protection circuit. They have to be mounted onto the batteries because they must protect the dangling leads when the battery is transported or in use. Therefore they are powered by the cells they are protecting and will eventually discharge the battery, even with very good low-power design.
The MOSFETS must be chosen with very low on-resistance. 18 milliohms is not too hard to find. But a MOSFET has an unavoidable "body diode" so you need two in series facing each other, so you are at 36 milliohms even before you connect it to the battery.
Plus there are more BMS issues that are simplified with a fuse...
wvmarle:
Interesting fuses you have. A normal fuse has a negligible voltage drop, until they blow.
Interesting normal fuses you know about. Any fuse relies on the heating effect of the current threough it to melt the fuse wire. The heating power = current squared x Resistance. My fuses are 2 amp and have a resistance of about 0.1 Ohms, so when passing 1 amp at nominally 3.5 volts they drop 0.35 Volts, which means the charge controller will see the battery voltage as 0.35 volts higher than it really is. That's huge and means the charge controller will basically trickle charge almost all of the capacity.
These forums are for helping people, not for making snarky comments. If you can't control your urge to make sarcastic comments to people you don't know who are asking for help then perhaps you should find another outlert for your agggression, because you're not just not helping - you're unpleasant.
Use 7 fuses. Clean the ends of each fuse and the fuse holders. Resistance should be much less than you can measure with a regular multimeter, maybe 50 milliohms for the fuse itself and 50 milliohms for the connections.
Design and build 7 copies of a MOSFET battery-protection circuit. They have to be mounted onto the batteries because they must protect the dangling leads when the battery is transported or in use. Therefore they are powered by the cells they are protecting and will eventually discharge the battery, even with very good low-power design.
The MOSFETS must be chosen with very low on-resistance. 18 milliohms is not too hard to find. But a MOSFET has an unavoidable "body diode" so you need two in series facing each other, so you are at 36 milliohms even before you connect it to the battery.
Plus there are more BMS issues that are simplified with a fuse...
Thanks for your comments.
The new "beefy" cell bank charging leads will terminate in a Molex socket which recesses the wires well, and I would 3D print an endcap, so I'm not to worried about them shorting the battery when away from the charger.
And I'm thinking worst case is that the TP4056 will fail short circuit, and I'm wondering how much current that will pass until it destroys itself and becomes open circuit. So I'm not sure how big the risk is, though I do understand how bad the consequences would be. However, the battery is always charged when attended and stored with the bike in a concrete shed.
Surprised your having such significant issues with the voltage drops across fuseholders and the ballance leads at such a low current.
I was surprised too. At first the balance leads themselves were the bottleneck. Only when I replaced one on a test rig (two single cells in parallel) did I see the fuse problem. Maybe the cheap fuse holder is a significant factor so I will look at that. Thanks for your comments.
ChrisXenon:
Interesting normal fuses you know about.
Have a look at e.g. mouser web site, look for 2A fuses. The typical resistance of such fuses is 30-50mΩ, which at 1A would give a 30-50 mV drop. In other words, the fuse takes <1% of your supply voltage. That's simply negligible for most practical purposes. Even your wrongly calculated 0.35V will not cause any issues charging a LiPo battery with a 5V charge supply.
It sounds like you're doing something wrong indeed. It's quite certainly not the fuse itself.
Mayself I would not entrust a mess up of TP4056 20p chargers to charge what is probably an expensive Lithium Polymer battery.
There are plenty of chargers out there that will reliably charge Lithium batteries an far faster than TP4056 chargers are capable of, such chargers are not expensive.
Bear in mind , if you are using a multimeter , it won’t read such low resistance with any accuracy .
As everyone has said the fuses won’t have that high a resistance and should not be a problem .
Fit one in a circuit , pass 1A and measure the voltage each side if you won’t further confirmation
Well thanks again for your time folks. Sorry about my glaring maths error. 0.1 Ohm x 1 amp = 0.1 Volt of course, not 0.35. What I actually measure with the fuse in and out of the circuit is a drop of about 0.2 Volts, which does dramatically lower the charging current. I will accurately measure the fuse/holder combinaiton (not using a multimeter - as suggested) and find exactly what is going on.
To address other comments, I currently charge either with a simple end-to-end power brick - 29.4 Volts, non-balancing and then balance once a week using an IMax B6 - I have split the battery into two with a centre tap to allow S4 and then S3 balance charging. Most of the charge time, the balance charger is pushing in less than 200mA because one cell bank is up to 4.2 volts and the others can't be charged fatser than that bank can be bled. That is why I have looked for another solution. Granted an S7 charger would be more convenient - only one charge cycle not two - but I assume it would still spend the bulk of its time trickle charging.
I also have a balance charger of the kind usually fitted inside the battery. My tests showed this to be worryingly unreliable and in the end I wasn't willing to seal it inside my battery and walk away from it.
And I wanted to explore the challenge of building a parallel charger.
wvmarle:
Have a look at e.g. mouser web site, look for 2A fuses. The typical resistance of such fuses is 30-50mΩ, which at 1A would give a 30-50 mV drop.
The resistance of a fuse when carrying a substantial fraction of its rated current can be a lot higher as the fuse
wire warms up. Its a balance between the heat generated in the wire and the heat lost through conduction
along the wire. Fuses normally use wire whose resistance increases with temperature to guarantee proper fusing action (thermal runaway).
Contact resistance in a cheap fuseholder could be surprizingly large too. Spring steel contacts and riveted (rather than spot-welded) construction have a lot more resistance than decent phosphor-bronze or beryllium-copper fittings.
Other options for the OP is to use a soldered-in fuse, or a fuseable link using a loop of copper wire that will carry the current needed, but melt when that value is exceeded. Then there are silver fuses, etc.