irritating battery limits

[TheNumericDolfin]

pi circut.png (27.98 KiB) Viewed 2190 times

Here is my basic circuit. Works, except that the pi must turn off when charging due to the limits of rechargeable batteries (cannot charge and discharge at same time). I would like to be able to control robot while charging. Is there any way to accomplish this?

[DirkS]

Works, except that the pi must turn off when charging due to the limits of rechargeable batteries (cannot charge and discharge at same time). I would like to be able to control robot while charging. Is there any way to accomplish this?

Find a pack that can be used while charging. They do exist...
I have one, but that one's that is out of production.

E.g. Ravpower and Anker have units that can be used while charging

[Imperf3kt]

I am guessing the blue rectangle labelled 5v is a boost regulator?
And the 'irrelevant control circuit' is a motor controller?

You don't seem to have any battery balancing in place, just three Lipos connected in parallell. I would not recommend this, as it complicates matters.
Better to stick to a single Lipo or buy a 3s pack and use a buck converter instead of a boost converter.

[TheNumericDolfin]

How can i safely connect batteries in parallel, then? using one big battery large enough would be intolerably expensive. Also, I looked it up and google says that lithium batteries cannot charge while discharging. Am I not understanding something that would allow this to happen?

and yes, the labeled parts are what you guessed them to be.

[drgeoff]

I am guessing the blue rectangle labelled 5v is a boost regulator?
And the 'irrelevant control circuit' is a motor controller?

You don't seem to have any battery balancing in place, just three Lipos connected in parallell. I would not recommend this, as it complicates matters.
Better to stick to a single Lipo or buy a 3s pack and use a buck converter instead of a boost converter.

@imperf3kt

You'd better stick to areas where you know what you are talking about.

Balancing is for rechargeable lithium cells in series.

But it is preferable to use a single cell of the required capacity.

[TheNumericDolfin]

I am guessing the blue rectangle labelled 5v is a boost regulator?
And the 'irrelevant control circuit' is a motor controller?

You don't seem to have any battery balancing in place, just three Lipos connected in parallell. I would not recommend this, as it complicates matters.
Better to stick to a single Lipo or buy a 3s pack and use a buck converter instead of a boost converter.

@imperf3kt

You'd better stick to areas where you know what you are talking about.

Balancing is for rechargeable lithium cells in series.

But it is preferable to use a single cell of the required capacity.

A single cell is really not feasible cost-wise in this case. If i can, indeed, connect these batteries in parallel, then how might i keep it operational whilst charging? Perhaps there is a way I could only charge one cell at a time, and switch to the next when that one reaches full capacity? If so, how might I change exactly what each battery is connected two with a signal?

[PhatFil]

If using 3 x lipo cells in parallel stepping upto 5v why not also use 3 x tp4056 charge management boards and some additional circuitry to issolate cells from load 1 by 1 for charging.

[TheNumericDolfin]

If using 3 x lipo cells in parallel stepping upto 5v why not also use 3 x tp4056 charge management boards and some additional circuitry to issolate cells from load 1 by 1 for charging.

That is a very good idea, but i am a little unclear as to how exactly i might set up such a circuit. What would be the best way to set this up? Would i swap said circuitry with the gpios or is there some sort of circuit already designed to do this automatically?

[Imperf3kt]

@imperf3kt

You'd better stick to areas where you know what you are talking about.

Balancing is for rechargeable lithium cells in series.

You must be careful when charging cells in parallell as well, if the voltages aren't close you will permanently damage one or more batteries in the parallell setup.

I'm basing my answers off the several dozen results presented to me by Google, such as the following example
https://www.flitetest.com/articles/Para ... _Batteries

[salvato]

Could this article help ?
https://hackaday.com/2019/06/12/an-exha ... 650-packs/

[PeterO]

@imperf3kt

You'd better stick to areas where you know what you are talking about.

Balancing is for rechargeable lithium cells in series.

You must be careful when charging cells in parallell as well, if the voltages aren't close you will permanently damage one or more batteries in the parallell setup.

Again you've got it wrong ! With paralleled cells the voltage are by definition all exactly equal ! The only time care is needed is when initially connecting the cells together when you need to make sure they are all at the same voltage (i.e. state of charge). Once connected you can just treat them as a single cell of twice (or N times) the capacity. There is no need to separate them for charging, in fact that is a bad idea as you'll have to equalise the voltage again before paralleling them up again.

PeterO

[Imperf3kt]

I'm not sure if it's because I'm not using the correct words or what not, but that is exactly what I am referring to, PeterO - you must make sure the voltages are very close before parrallelling the batteries. I didn't mention disconnecting the batteries before charging them, nor did I suggest requiring to balance a parallell setup once connected.

As to cost, a single 10, 000 mAh battery from the manufacturer costs less than three 3300mAh batteries. Knowing how big a battery does OP requires, might help determine the best course of action.

[rpdom]

I think what PhatFil was suggesting was having three cells with separate charging circuits, so one cell could be switched out of the power circuit and put on charge while the others were supplying power. Keep rotating the cells one by one on a charge/power cycle.

I don't think that will work unless there is something to carefully match the voltages before switching a cell from charger to power mode.

Unless some clever circuitry with "perfect" diodes between each cell and the boost to 5v converter circuit.

[MarkR]

If you have a smart enough charge controller, it will be able to provide power while charging.

Or you could use a separate boost converter to power the Pi from the cells all the time, there is no problem doing this while charging.

Mobile phones obviously manage to do this.

As far as connecting cells in parallel- it is ok, provided they are close to the same voltage before you connect them. Don't try parallel connecting a nearly flat cell to a fully charged cell, a large current will flow and it might end badly (there is ~ 1.0V difference between a full and empty lipo)

[PhatFil]

I think what PhatFil was suggesting was having three cells with separate charging circuits, so one cell could be switched out of the power circuit and put on charge while the others were supplying power. Keep rotating the cells one by one on a charge/power cycle.

I don't think that will work unless there is something to carefully match the voltages before switching a cell from charger to power mode.

Unless some clever circuitry with "perfect" diodes between each cell and the boost to 5v converter circuit.

Something like that, but it was late and i overlooked the issues you raise, whoops..

[rpdom]

I think what PhatFil was suggesting was having three cells with separate charging circuits, so one cell could be switched out of the power circuit and put on charge while the others were supplying power. Keep rotating the cells one by one on a charge/power cycle.

I don't think that will work unless there is something to carefully match the voltages before switching a cell from charger to power mode.

Unless some clever circuitry with "perfect" diodes between each cell and the boost to 5v converter circuit.

Something like that, but it was late and i overlooked the issues you raise, whoops..

No worries. The idea was good and could work with a bit of extra circuitry.

[fanoush]

Also, I looked it up and google says that lithium batteries cannot charge while discharging. Am I not understanding something that would allow this to happen?

The idea is that to charge and also run it at the same time you need to provide more power than what your robot needs. So you don't charge and discharge at the same time, you power the robot from the charger directly and only extra current goes to charging whole battery pack which is unused and not providing power, only charging.

[amcdonley]

A subject dear to my heart, this is - how to recharge a robot without shutting it down? (In my case because the bot has no wakeup timer function.)

Carl Recharging On His Dock.jpg (238.12 KiB) Viewed 1815 times

How I solved it for my GoPiGo3 robot:
1) Eight 2800mAh NiMH AA-cells (series connected)
2) Twin pigtails in parallel off the battery, one to bot and one to charging contacts (or charger)
3) GoPiGo3 board has diode protected 5v regulator (and power switch, and low voltage shutdown protection)
4) Tenergy 1025 6-12v smart charger in 1A mode
5) Logged and profiled vBatt() readings in charge mode and trickle charge mode
so bot could know when to get off charger autonomously (at start of trickle mode)
and could recognize unsuccessful dockings (docked but not charging, or docked but low_voltage after time due to early trickle_mode)
6) Ripped the guts from a Roomba charging dock

There is an issue with this method though - bot contacts are not short protected yet.

Here is the voltage profile:

battLife-3-Charge-Discharge-Cycles.png (70.09 KiB) Viewed 1813 times

(There is a 1-2 minute trickle at the peak of the charge cycle, which bot detects and gets off dock.)

This is my bot getting off and onto the charger:
https://vimeo.com/333207248

[TheNumericDolfin]

OK, the regulator should(In theory) make sure that the voltage does not vary too much. As long as at least one battery is providing 2.1 amps to the pi, would there be any danger of burning out the pi or accidentally shutting it down? As long as all batteries do not run out at once and provide at least 2.1 amps, then should this work? And if so, by what mechanism could I actually swap what circuit the batteries are connected to? Ideally, something like an electronic version of a train rail switcher controlled via GPIO. I have tried to find what this would be called, but I do not know what magic words to ask google.

PS. sorry was disconnected from the internet for a while, could not respond. Also, Carl is god dammed adorable.

[amcdonley]

Have you looked at this phenomenal UPS https://lifepo4wered.com/lifepo4wered-pi+.html ?

I have the earlier version running a desktop RPi continuously for six months at a time. I like to update the OS periodically and sometimes that requires a reboot.

[TheNumericDolfin]

OK, so i took a look at that specialized battery, and it definitely looks intriguing. The problem lies in that this solution would only work for the pi part of the robot, and would not allow function for the rest whilst charging. Ideally, I would need a circuit that could take a signal from a GPIO pin and change what an individual battery is connected to. Unfortunately, trying to search for specific electronic components online is a complete nightmare when one does not know that exact name. If somebody could tell me the name of this type of component I would be very grateful.

[amcdonley]

OK, so i took a look at that specialized battery, and it definitely looks intriguing. The problem lies in that this solution would only work for the pi part of the robot, and would not allow function for the rest whilst charging. Ideally, I would need a circuit that could take a signal from a GPIO pin and change what an individual battery is connected to. Unfortunately, trying to search for specific electronic components online is a complete nightmare when one does not know that exact name. If somebody could tell me the name of this type of component I would be very grateful.

It is often the case that a robot logic needs a lower voltage than the robot's motors. The use of a switching supply step-up buck on the output of the battery circuitry will give you two separate voltages without needing any switching. Pololu is a great source for voltage regulators - https://www.pololu.com/category/132/ste ... regulators.

You should not be trying to design a power supply. Stick with commercially available circuits from professionals!