Neato lithium ion battery revisited

News and information about the Neato XV-11 Robotic Vacuum. All discussion and troubleshooting questions go here.

Re: Neato lithium ion battery revisited

Postby glnc222 » March 17th, 2019, 9:27 pm

When you switch to lithium the absence of self-discharge will stop the repeated charging cycles when left on the dock, probably triggered by a certain voltage drop which won't happen with lithium. Yet another reason the lithium will last a lot longer, but the LiFePo4 is supposed to last very much longer than regular lithium, maybe 10 times longer -- life of the robot. These cells are used in some of the robot lawn mowers which are large enough to accommodate the higher number of cells and larger bulk of the LiFePo4 with lower energy density.
glnc222
Robot Master
 
Posts: 4966
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby Ben_Affleck » March 18th, 2019, 11:44 am

I am currently browsing ebay for a ready made BMS, as it doesn't make sense to reinvent the wheel if it is already available cheaply:

I found 2 options, both at the same price point:

1. This one seems to be doing balancing and also overcharge/overdischarge protection:
https://www.ebay.com/itm/4S-30A-12-8V-w-Balance-3-2V-LiFePo4-LiFe-18650-Battery-BMS-Protection-PCB-Board/263691717493?hash=item3d65404b75%3Ag%3AULcAAOSwuvla~BNv&LH_BIN=1

2. This one seems to be doing only balancing:
https://www.ebay.com/itm/4S-12-8V-14-4V-LiFePo4-LiFe-Battery-Charger-Module-Balance-Board-3-2V-4-Cell-4S/322616103362?hash=item4b1d6b11c2%3Ag%3AT3QAAOSw6M9bvw9U&LH_BIN=1

Based on the technical specs an obvious choice would be no. 1, but I've made a cardboard template with the provided dimensions (47 x 56mm) and I didn't find any way to make it fit in the battery compartment of my D85.

A solution for this maybe would be to place the balancing board inside the bot, but once again I don't know if there is anymore space inside it anywhere. Maybe someone already explored this "finding more space" issue and can give me a tip..

Another question for both boards is what is the maximum rated input voltage they can use, because I didn't find any specs for this anywhere. I know that it should be at least 14.8V, the stated charging voltage used, but I don't know if it can support the maximum 18.5V Botvac can supply. Again, if anybody already played with these modules, please give me some input on this
Ben_Affleck
 
Posts: 17
Joined: February 26th, 2019, 7:39 am

Re: Neato lithium ion battery revisited

Postby glnc222 » March 18th, 2019, 5:32 pm

The only other user who did this posted about his Botvac battery in thread on that topic
http://www.robotreviews.com/chat/viewtopic.php?f=20&t=19149
There is some discussion of the balancing board.
One would expect a balancing board for the particular series of cells to tolerate the Botvac charging voltage as this will be similar to that for such battery sizes.
The temperature charger terminating circuit must signal the charger at a voltage below the over-voltage protection level of any such circuit used, maybe 0.2v less.
glnc222
Robot Master
 
Posts: 4966
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby glnc222 » March 18th, 2019, 9:51 pm

I recall the balancing board used in the Botvac project was modified by changing the resistors on it. It appears to match the specs of ones suggested above. The problem with these boards is insufficient balancing bypass current to heat the boards up sufficiently. They specify 58ma balancing current -- at least with the supplied resistors -- maybe not enough.
The problem with XV experiments found in Germany was newer balancing boards, with dedicated IC's made for balancing control of the mosfets, supported insufficient bypass current, compared to older boards discontinued. A custom board was even manufactured by the hobbyist supplying these DIY batteries there (roboter-forum thread), when the part on the original board was discontinued.

I noticed the protection boards work by inserting ground side mosfets on the batteries, controlled by voltage sensing IC's, specialized controllers specifically for this function. The current capacity spec is a matter of how heavy duty are the mosfets used. If making a voltage-to-temperature adapter board with an ICL7665, the extra side could be used to implement the typical over-voltage protection with a mosfet, maybe the FDP8800 I used. An alternative might be to have a mosfet open a bypass circuit through a resistor which would work similar to the balancing circuits, not for heating but to satisfy the Neato charger expecting less than infinite battery resistance to charging voltage. The engineering needs refinement.

In one XV experiment I put things in the dust bin for space.
glnc222
Robot Master
 
Posts: 4966
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby Ben_Affleck » March 19th, 2019, 3:53 pm

glnc222 wrote:An alternative might be to have a mosfet open a bypass circuit through a resistor which would work similar to the balancing circuits, not for heating but to satisfy the Neato charger expecting less than infinite battery resistance to charging voltage.


This is my plan for now
Ben_Affleck
 
Posts: 17
Joined: February 26th, 2019, 7:39 am

Re: Neato lithium ion battery revisited

Postby ztshopping » March 28th, 2019, 12:20 am

question for XV-11: why not add a control to the charge station that will cut off the charge current when the charge voltage is raised to 16.8 V before the BMS is triggered? this will mimic the Neato Li-ion charge regardless the setConfig battery type value. or even skip the BMS on the cells and let the charge station do the cut off ?


Neato will be fine when the charge current is completely cut off, the battery may not be fully charged but it should be fine with Li-ion cell. just use a little bigger capacity cell. 2P of 2600maH cell will make it 5200maH total, and even 80% full will make it about 4100amH.


The other thought is simply add a step down converter between the 24V power supply and the Constant Current circuit to limit the max voltage to 16.8v and max current to 2.0A, this will force Neato to stop charging or maybe with very small current when Li cells are fully charged without triggering the BMS bypass to avoid error 002, or simply remove BMS completely and use a heat-fuse like that used in the NiMH pack to protect over heat.
ztshopping
 
Posts: 3
Joined: March 26th, 2018, 5:42 pm

Re: Neato lithium ion battery revisited

Postby glnc222 » March 28th, 2019, 7:54 pm

Some clarification may be needed here. In my LiFePo4 experiments I did use a charger cut off of the kind described above for over-voltage protection, vs the common battery disconnect, to protect the system board from high voltages besides just the battery. However, any such cut off, caused by the charger to stop properly, driving voltage high on a constant-current system, will cause an error condition declared by the system software. There actually is a system failure in not terminating charging properly. Vulnerability of the system board may be less in the lower voltage Botvac system compared to the XV models. Whatever circuit, battery or charger disconnect, is easiest to construct with say, the ICL7665 comparator, could be used. The battery disconnect is easiest to attach to the battery. It would only activate in some abnormal failure of the normal system operation -- or problems in constructing a DIY battery and adapter, design and construction errors.

In Germany one engineer making his own adapter circuits did use a bench power supply instead of the dock to supply a lower charging supply voltage for protection.

You still have to signal the NiMh charger to stop with a rising temperature signal, synthesized from voltage input or thermally created by the heat in a balancing circuit in contact with the pack thermistors. (Only the high voltage side pack thermistor in the dual pack series system is monitored for charging control.)
glnc222
Robot Master
 
Posts: 4966
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby ztshopping » March 28th, 2019, 10:41 pm

@glnc

when I say cut off I really meant to cut off the power supply at charge station, not on battery, so there will be 0 v on the charge station plate. It is like during the normal charging process I forcefully pull Neato away, or un-plug the charge station power. Neato will likely do the left-right maneuver after it detects no power on the charge station. This should not do any harm to the Neato circuit. The point here is that cutting off the charge station power at 16.8v or even lower using something simple like Z-Diode to avoid BMS by-pass that cause Neato 002 error.

Or use a buck step down converter to limit the charge station power to 16.8v and current to 1.5A , such step down converter will replace the Neato CC board. This change indeed change the charge station to a Li-ion charge station: CC at 1.5A and max voltage at 16.8v or 16.5v to avoid Li-ion BMS over voltage protection to kick in.

This work at charge station side is much simpler than the work on the battery side

when u need make it a Ni-MH charger, adjust CC to 2 A and max voltage to 24V. when u need Li-ion charger make it CC 1.5A and Max voltage 16.8v.

use this step-down converter.

https://www.ebay.com/itm/400440723717?_ ... EBIDX%3AIT
ztshopping
 
Posts: 3
Joined: March 26th, 2018, 5:42 pm

Re: Neato lithium ion battery revisited

Postby Ben_Affleck » April 2nd, 2019, 3:32 pm

I've just had a brainfart.
The chinese balancing board opens a circuit through a resistor when the cell is fully charged, right?
So basically that resistor receives current (and voltage) only when the cell has reached full charge. We can consider the battery pack charged when all individual cells have reached the threshold value and the circuit through the resistor has been opened, so when all of the bleeding resistors are in a logic HIGH state.

What if I would be using the point where the bleeding resistors are connected as inputs for a logic AND gate with the output activating the thermistor signaling circuit (mosfet, resistor..)?
There are many 4 inputs logic gates available cheaply and in this manner I would be also sure that the pack is balanced properly before ending the charge.

Thoughts?
Ben_Affleck
 
Posts: 17
Joined: February 26th, 2019, 7:39 am

Re: Neato lithium ion battery revisited

Postby glnc222 » April 2nd, 2019, 9:49 pm

When speaking of a logic AND or other gate component, these are standardized to either TTL logic (old fashioned, more power hungry), or modern CMOS logic standard voltages, CMOS being the full 5v rails of the power supply for them (you do need a standard logic chip power supply to use them, whatever the input voltages...). You won't get that out of the voltages on the balancing current resistors in the balancer I suspect, though I am not completely sure how those work. What you usually would have to do in order to perform logic operations on non-logic part voltages is use voltage comparators which effectively do parts of the logic on arbitrary voltages, by converting the arbitrary inputs into pure logic level outputs -- takes a lot of parts in some cases, depending on how complex a relation you want. It would seem to be simpler just to use a comparator to detect the maximum charging voltage on the battery instead of differentials across cells used by the balancer controller IC.

The original German balancer is analyzed at the beginning of this thread, using an individual cell circuit with a simple comparator no longer made, LBA1. This balancer was made to be stackable, with a separate small board on each cell in any size pack. The lithium batteries using this for the XV models were made by member Hausman of Roboter-Forum.com, thread Lithium Ionen Upgrade
https://www.roboter-forum.com/index.php?thread/5804-neato-lithium-ionen-upgrade/
A custom board was made when the original comparator ceased to be made.
Technically the controller IC for balancers monitors differentials between the cells and not the overall battery voltage. Balancing current presumably flows at any time, not just when the battery is full, as it tries to keep the voltages on all the cells equal at all times, I think. Balancers incidentally, seem to be mostly used in external battery chargers, instead of inside battery packs. There can be separate pack balancing cycles performed after main battery charging cycles, is my impression -- more expertise needed.

I once posted on the German forum some idea of supplementing the balancer with a direct voltage to temperature signal. There may have been a way to do this with the old LBA1 version but I would have to check, and it probably would not work on the newer, controller IC types. Maybe you would know this engineering better -- I'm just self-taught hobbyist.

Incidentally the LiFePo4 cells are described as "self-balancing" requiring less of the balancing circuits used with lithium polymer to extend the life of battery packs with varied aging of the cells.

The common protection boards, for both over and under voltage (excess discharge as with short circuits), are to protect the battery from failures in the attached electronics including charging and discharging. These are more critical with lithium ion than LiFePo4 because of a fire hazard with the former, besides damage to the cells. If such failures in the Neato system board are considered unlikely, there is less need with LiFePo4. Neato has protection against over-discharge in software in that it shuts down when voltage is too low to run the system, besides normally keeping around 30 per cent charge unused as a buffer against damaging deep discharge, and reserve for returning to base.

Incidentally the German guys were professional engineers who just played with the Neato as entertainment.
glnc222
Robot Master
 
Posts: 4966
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby glnc222 » April 2nd, 2019, 10:04 pm

As to how to cut off the charger in emergency, whether to shut off the dock, the connection to the Neato, or the battery connection to the Neato (the standard protection circuit method), it seems more complicated to signal the external dock to shut off as some additional connection outside the Neato would be needed. The common protection board method seems to be simplest in being contained within the battery pack without modifying anything in the Neato itself -- and is available ready made commercially (though whether it will fit is another matter, and it did require modifying the commercial board with different resistors if I recall that project correctly, posted on this site -- but that was for balancing, not protection).

Incidentally there is a 5v logic power supply in the Neato used to operate the IR proximity sensors and mag strip sensors at least, and a CMOS logic circuit will demand only miniscule power.
glnc222
Robot Master
 
Posts: 4966
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby Ben_Affleck » April 3rd, 2019, 6:24 am

I am still waiting for brand new LiFePo4 cells and components to arrive.
I will experiment with all of the above proposed scenarios and come back with conclusions.
Ben_Affleck
 
Posts: 17
Joined: February 26th, 2019, 7:39 am

Re: Neato lithium ion battery revisited

Postby Ben_Affleck » July 1st, 2019, 2:58 pm

My quest is finished.

After experimenting with every possibility, the best, safest and cheapest route you can take in order to mod a Botvac Li battery pack is LiFePo4 cells + balancing + terminating charge signaling circuit.

I've decided that the detailed description of the project would better fit the LiFePo4 thread, so you can find it here:
http://www.robotreviews.com/chat/viewtopic.php?f=20&t=19149&p=152415#p152410

Thanks glnc222 for your insight
Ben_Affleck
 
Posts: 17
Joined: February 26th, 2019, 7:39 am

Previous

Return to Neato Robotics

Who is online

Users browsing this forum: funnyfsdfsdaf-123, im1kissfan, jdong, johnnyh, leonicholson, NoJimForHome, raidencmc and 397 guests