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 Ben_Affleck » February 26th, 2019, 8:00 am

Hi all,

Has anyone reverse engineered the electronics in the LithiumPowerInc 7200mAh LiNMC pack?

My NiMH in my D85 has just died and I was thinking to "upgrade".
I am hobbyist with some knowledge on electronics and microcontrollers, with some free time and a whole stack of reclaimed 18650 Li-Ion cells still in good form.

I explored the LiFePo4 mod on the forum, but this is not satisfying as I do not own any LiFePo4 cells, they are hard to find and shipping costs more than the cells, so it doesn't make sense for me.

I understand that LithiumPowerInc's pack is a good replacement for the original 12V NiMH that is why I would like to know what solution have they found for interfacing with the D85's original NiMH charging circuit.
Ben_Affleck
 
Posts: 12
Joined: February 26th, 2019, 7:39 am

Re: Neato lithium ion battery revisited

Postby glnc222 » February 26th, 2019, 10:45 am

The problem with lithium ion cells for the Botvac is the cell voltage, which does not seem to divide well into the 12v nominal Botvac system voltage so that a pack would be either too low or too high.
Maybe you can find a spec for the voltage on the lithium ion replacement pack for the Botvac.
The lower LiFePo4 cell voltage divides well. Nothing wrong with trying a pack too low, but over-voltage might stress some component, don't know enough to tell (voltage varies quite a bit when running, along the discharge curve, up to 2v drop; when initially full the NiMh cells are a volt over the nominal, about 11-13v). Also with the Botvac the SetConfig BatteryType USB command is not available to engage the lithium ion charging procedure, so you would need the heat-simulating adapter similar to the LiFePo4 project (DIY version the XV project, else re-using a balancer board). . I think the newer Botvac's with OEM lithium batteries run on the higher 15v of the XV models (see the replacement battery specs) Those LiFePo4 cells should last a lot longer than any lithium ion cells, according to literature.
With lithium ion cells an under-over voltage protection cut off board is essential due to the fire hazard should anything fail in the charging circuitry; LiFePo4 is more robust, or at least would only damage the cells instead of the house....
There are other types of "safe" alternative lithium chemistries, not as durable as LiFePo4, you could check those voltages. LiMn etc.

When developing heat-simulating adapters for the NiMh charger, a vulnerability of Neato system boards emerged: with lithium ion, going full resistive at the maximum voltage, failure to stop the charger pushes the Neato constant-current charger to its maximum voltage, from the dock plates. On the XV at least, when the internal components go to 21v some system board components can burn out, not rated for that not normal in any NiMh operation as NiMh continues to draw charging current when full and dissipates it as heat. The lower 12v Botvac system could be less vulnerable as the dock supply is lower, unknown. So for the XV experiments I developed a protection board interrupting the charger instead of the cells, quite a mess.
glnc222
Robot Master
 
Posts: 4913
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby Ben_Affleck » February 26th, 2019, 5:53 pm

Project 12V NiMH Botvac D conversion to 14.4V LiIon
Recruits welcomed :)

1. Charging plan
For charging I'm going to use a dedicated BMS, so at full charge the pack (4s2p) will be balanced at around 16.8V.
As I want to connect to the robot's battery pins with my own circuitry it would help alot if someone already determined:

- What is the maximum voltage provided by the Botvac motherboard when charging (I need the voltage at the battery pins, NOT at the wall base charger plates)?

2. Charging know-how
The base specifies on its sticker that it may be used for the NiMH 12V Botvacs or the 14.4 LiIon Botvacs.
This tells me that the base voltage is further converted on the internal board for charging the battery.

Thinking as a manufacturer it makes sense to have as less design variations as possible on the motherboard layout (and modules).

- Did someone already compared the charging/PSU circuitry on the motherboards of both the 12v NiMH and the 14.4v LiIon Botvac versions?

I expect/hope for marginal differences.

3. Charging complete signaling
For the charging to stop I am going to simulate the drop in resistance by circuitry. Heating the thermistor with the BMS resistors is not ideal, nor precise. I want better control over this. I want to build a circuit able to signal the ending of the charging process by lowering the resistance over the thermistor pins.
Question regarding this:

- What is the thermistor resistance threshold at which the bot decides the charging process has finished?

I already know that at room temperature it is at 10-11ohms and with rising temperature the resistance goes lower, but how low does it need to go in order for the robot to consider the charging process finished?

4. Baterry usage / discharging
I know that 16.8V is over the threshold for the original NiMH pack which should be around 14.5V when full, so I will probably use a buck coverter to lower the voltage, if needed and if the motherboard does not support 16.8V as an input.
Maybe someone already explored this:

- What is the maximum voltage supported by the robot as battery input without stressing/frying the board?
Not the theoretical one, the practical one :)

5. Recharge signaling
Usually the BMS cuts the output at a very low-voltage, something like 2.7V/cell. For the whole pack this would mean 10.8V, which I think is lower than the voltage at which the bot decides to go back to base for recharging, so the BMS should not interfere with the return to base protocol and the pack would provide enough juice for the road back.

So my buck converter should keep the voltage just over the "return to base" threshold in order to obtain the maximum operating time from the battery. As the pack would go under the threshold, the buck regulation would fail as well and the bot would be signaled to go to base for recharging.

It would help if someone already determined :
- What is the threshold voltage at which the robot decides to stop and return to base for charging?

6. Safety
Charging : The BMS should manage the charging process in safe way. For added safety I will still use the original bi-metal switch in seriers with the BMS/battery.
Discharging : The robot will return to base long time before the cells would over discharge, but if something unexpected happens the BMS should cut the output, and if temperature rises from the discharge, the bi-metal switch would cut the connection also.

7. Miscellaneous
Helpful know-how, if already determined by someone else:

- How does the robot change internally the status of the battery pins from output (when charging) to input (when it runs on battery)?
- A reverse engineered schematic of the charging circuit / internal PSU of the Botvac would be highly appreciated.
- The same goes if someone opened the LithiumPowerInc's pack and reverse engineered the schematic as it seems to be a 3s3p pack, so under the required voltage levels. High level I would guess a BMS + boost converter and some microcontroller supervising signaling?
Ben_Affleck
 
Posts: 12
Joined: February 26th, 2019, 7:39 am

Re: Neato lithium ion battery revisited

Postby glnc222 » February 26th, 2019, 9:01 pm

The Neato NiMh charging procedure is explored in detail earlier in this long thread, complete with description from a Neato Robotics representative. There are two NiMh procedures in practice. Common single cell (usually four at once), open-to-the-air chargers, as for AA batteries, use a small drop in charging voltage which occurs when full, when a constant-current charging circuit operates (a standard type of voltage regulator, either linear components or a buck voltage converter or switching power supply). Before this voltage blip the NIMh cells will dissipate heat near full, and with enclosed battery packs the heating is excessive, so the charging is stopped on the basis of a temperature rise to get ahead of over-heating. So these packs contain thermistor sensors, with the Neato ones falling in resistance as the temperature rises. The temperature read is reported over USB and can be observed continuously with the Neato Control program for PC (other thread). Details of these thermistors, tables etc., are earlier in this thread.
The algorithm used is not a temperature level but the rate of increase: charging terminates when the temperature rises faster than 1C/minute -- apparently a standard in NIMh equipment. So a simulating circuit need only impose a sudden step increase in temperature in order to trip this detection circuit. There is a temperature limit I think 60C where the Neato system declares an emergency shutdown, for possible fire hazard I suppose; there are also temperature tripped fuses in the NiMh battery packs. To simulate this step increase it is necessary only to parallel shunt the thermistor with a transistor to lower the pair resistance, which is in series with a fixed resistor supplying an analog voltage input to the cpu -- this voltage simply needs to be lowered. A voltage-to-temperature adapter circuit is shown earlier in this thread where the ICL7665 voltage sensor with internal voltage standard for convenience was found useful. A German hobbyist making lithium replacements for the XV brought that part to my attention. Some ultra-high precision 0.1 per cent resistors may be needed to calibrate this precisely.
The earlier method, used in Germany, was to use the heating of components in balancer circuits, as the voltage rises, to thermally simulate the NiMh heating and thermally couple to the thermistors.
Incidentally the best source of the particular thermistors Neato uses -- there are many types with different characteristics -- is salvaging from discarded Neato NiMh batteries. However, we did identify a standard part which appears closest to the Neato ones, earlier in this thread. One could just dispense with them and use a fixed resistor in principle, abandoning emergency temp detection.
Commercial lithium replacements all use a direct voltage-to-temp simulator while DIY Neato owners used balancing circuits. The problem there was that newer balancing circuits became more efficient and reduced their own heat dissipation, so it became hard to find any ready made balancers suitable. The Botvac LiFePo4 project shows one that appears to work. Such circuits are specific to the type of cell chemistry, LiFePo4 different from Lithium Ion etc., as it is all voltage dependent and the cell types differ in that.
With direct voltage-to-temperature signaling, care must be taken to coordinate the trip voltage with that of any over-voltage protection circuits added so as not to confuse them.

A complication in NiMh charging is additional phases after initial termination: a "topping off" procedure is used with charging at a lower current, only 1/2 amp, for fixed time period instead of temperature controlled. This phase can proceed in Neato's after the indicator light turns solid green indicating full. The NiMh heating can terminate the charging before the full capacity of the battery is reached.
These follow on phases present requirements for balancer type adapters and so on, which must handle the current involved and what not. Apparently signaling temperature rise again can terminate these phases.

Without empirical proof I assume there is a risk in powering the 12v Botvac system with a four series lithium ion pack going to 16.8v max, possibly damaging some components. Expert engineers would have to opine.
I seem to recall what Lithium Power did was use a lower voltage three series with a large capacity to compensate, so that the capacity at the upper end of the discharge curve is sufficient for the Botvac, while the bulk of the capacity at lower voltages is never used. This strategy vs LiFePo4 better voltage fits is one disadvantage of this product, with a sort of misleading capacity spec. -- and higher than necessary cost, for all that unused capacity.
The Neato system has step-down voltage regulators for a 5v supply used for sensors and lidar, a common logic component supply, as well as a 3v lower voltage used for cpu components (as they have gotten lower power over years). These may be linear voltage regulator components, and if supplied with too high an input voltage could over heat -- they are basically forming a resistor network and dissipating heat in the reducing resistance. Again engineering expertise -- or empirical results -- are needed to know exactly what will happen. Of course these low voltage supplies do not support high current drains like the motors. The components on the system board have not been identified so the specs are not known.

Neato Charging (and standard procedure) is a constant current variable voltage regulator, standard type of circuit, supplying a higher voltage than the unconnected cell to push in the charge, increasing the battery terminal voltage as the cells fill. This regulator is on the Neato system board and may be a buck voltage converter instead of linear component (typical for the high current used, 2amps in all Neato models -- battery cell replacements must accommodate this charging rate, a separate spec from discharge capacity). Lithium cells charge to a specified maximum voltage and must be protected from applying higher voltages. Additional charge is sometimes pushed into lithium cells with a constant voltage phase, ending when current falls to zero. Neato does not use this 2nd lithium charging phase, and depending on the cell type this phase may only add another 10 per cent of charge.

As to the power distribution circuit, this was traced for the XV by a senior engineer posting here, but not for the Botvac. Basically some Mosfet transistor switches and diodes arrange the power path according to system state, charging or running etc., under cpu control. The system is powered by the dock when charging. A problem with a diode in this circuit, D14 in XV models, causes a leak of voltage back into the charging contacts from the battery when running. The system detects connection to the dock from this contact voltage, so the software gets confused and stops running because thinks it is in charging mode (other thread on this). The diode appears to have been under-specified for load and larger capacity replacements cure the problem.
glnc222
Robot Master
 
Posts: 4913
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby glnc222 » February 26th, 2019, 9:39 pm

Caution fabricating Lithium Ion packs connecting cells. Cells with vendor supplied spot welded soldering tabs are preferred, and strong warning will be enclosed against soldering directly onto such cells, as heat is what sets off the runaway self discharge fire hazard (though the newer Lithium Nickel formulation, in power tool packs, is less dangerous -- NASA crush test). The German's soldered directly using high grade constant temperature irons for very fast soldering to prevent heat build up, but skill is required as well as equipment -- and maybe a safely prepared environment in which to work... (I made an automatic battery cycling measurement device with the battery set inside a cook pot).
Also cover exposed unused terminals while working to prevent short circuit accidents also fire hazard.
glnc222
Robot Master
 
Posts: 4913
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby glnc222 » February 26th, 2019, 9:49 pm

Regarding the dock specification as covering both 12v and 14.2v systems, it is not clear that there is only one mode of powering the dock plates. These docks can communicate with the robot, via voltages detected and what not, and have power management IC's in them with unknown functions. Testing is needed to know how these docks behave. The docks for the original Botvac, compared to the newer models with lithium batteries, could be different.
glnc222
Robot Master
 
Posts: 4913
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby glnc222 » February 27th, 2019, 1:20 am

Caution I think the complex "topping off" charging phase of the NiMh charger presents a complication for adapter circuits for Lithium Ion. The adapter circuit boards from Lithium Power can be quite complex being factory made and may have a microcontroller to deal with the problem as well as mosfet switches. If additional charging current is delivered by the charger after maximum voltage for lithium is reached, the charger voltage will rise and trigger a cut off from the over-voltage protection board -- which Neato will detect as a battery failure and issue an error msg.
What happens with the balancing board method is the balancer simply shunts the low current charging flow around the battery, presenting a phantom load for the charger to operate on. A similar feature may have to be added to direct voltage-to-temperature signaling. I think the way I handled that (it's been a long time) with the LiFePo4 XV model experiments was to terminate the charging at a little below the full capacity so there was room for the additional "topping" off charge. Perhaps another way to handle it is with a balancer but not used for thermally coupled signaling. I can also imagine a diode directed circuit which presents a simple resistor shunt bypassing the battery to the charger once the battery reaches the maximum voltage, i.e. a voltage controlled mosfet switching circuit which loads only the charging current, not discharge from the battery. There can be a problem fitting all this circuitry into a Botvac without a factory made adapter board for compactness. Some balancing boards ready made may have the over-under voltage protection wanted in them, not sure, but this would save some space. I think I once commented in the German forum that their balancer could be supplemented by a direct thermistor simulator. Balancers are also recommended for lithium ion packs more than NiMh. Some cell types including LiFePo4 are sometimes described as "self balancing". The balancing function insures all the cells in a pack are charged to the same voltage, compensating for variations in wear and capacity loss across the multiple cells, as packs as a whole wear as the weakest link. Supposedly they add lifetime to the packs.
glnc222
Robot Master
 
Posts: 4913
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby glnc222 » February 27th, 2019, 4:17 pm

Re Buck Voltage Converter step down from 16.8 etc.: most of the smaller boards I see listed are rated for 3 amps but some larger ones are rated 5 Amps, and that is the kind of peak load which occurs from motors in a Neato robot. The average running load can be a little over 3 Amps. A customer commented on one 5 Amp unit that it could go to 5 Amps only with a cooling fan, and you will notice the higher rated models all have large heat sinks, maybe a problem for fitting inside a Botvac. This could be why Lithium Power/ AnewPower did not use this option in their product. One also has to wonder why they would not have simply used a 14.7v (four series) battery if that worked in the Botvac.
glnc222
Robot Master
 
Posts: 4913
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby Ben_Affleck » March 2nd, 2019, 7:35 pm

Physical space for fitting everything in seems to be a problem.
I might need to design my own board with everything neatly arranged.
Some needed larger caps and inductors might make stuff bulky anyway..

Not an easy challenge!
Ben_Affleck
 
Posts: 12
Joined: February 26th, 2019, 7:39 am

Re: Neato lithium ion battery revisited

Postby glnc222 » March 3rd, 2019, 8:21 pm

I am not sure what is gained going to a four series higher voltage because the Neato charge usage cleaning is programmed based on the OEM battery, and higher capacity does not increase run time. The XV's would use only about 2ah charge cleaning before recharging. So if a 2 parallel three series of LiIon has 4ah capacity, it might well provide the full run time in the upper part of the discharge curve. I did not see capacity specs for your cells in the posts so far.
It is also not clear how you would charge the higher voltage four series, as the internal charger on the system board is calibrated in software for the OEM battery voltages, regardless what the dock supplies -- which is also still a question.
If you could fit in 9 cells for 3P three series surely you would have quite a large capacity -- unless you have really small LiIon cells.
glnc222
Robot Master
 
Posts: 4913
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby Ben_Affleck » March 12th, 2019, 6:37 pm

The Li-Ion adaptation for the 12V system is indeed a rabbit hole. Now I understand why the big guys struggled for some time before bringing to market the Li-Ion replacement battery pack..
I am not giving up on the idea, but for now I am really missing my bot cleaning the house (my wife also..) so I am going the LiFePo4 route.

I found these cells:
http://www.candlepowerforums.com/vb/showthread.php?452999-Test-review-of-Enerpower-18650-3-2V-1800mAh-(Blue)
The spec results from the tests performed by these guys seem right for my application, so I ordered 8 cells (2p4s), they should arrive in couple of weeks.

I will be using an ebay 4cell BMS LiFePo board for charging.

For charging termination I want to go on the simulation circuit route, not on the heating resistor route.

If I understood correctly, the charge termination signaling is just a quick drop in the resistance over the thermistor pins.
That should be easy enough to simulate, right?

If a more linear resistance decrease is needed I am thinking also at a voltage controlled resistance circuit with a J-fet or even a vactrol (even an inhouse build with an led and a photoresistor should do the job). But is this the case or the sudden drop in resistance is enough?
Ben_Affleck
 
Posts: 12
Joined: February 26th, 2019, 7:39 am

Re: Neato lithium ion battery revisited

Postby glnc222 » March 12th, 2019, 9:20 pm

The voltage-to-temperature adapter I used on XV model (with higher voltage than Botvac) is in this post above in this thread with schmatic
http://www.robotreviews.com/chat/viewtopic.php?p=124772#p124772
The ICL7665 voltage comparator has two sides where one is used for the temperature and the other for an over-voltage protection with a mosfet charger interrupt instead of the usual battery interrupt (because over-voltage in the charging circuit once burned out a system board component -- on the XV not Botvac with lower 12v system and charger voltage).
The ICL7665 by itself can ground a resistor in parallel with the thermistor to lower the resistance and raise the temperature a few degrees C, simulating the increase faster than 1C/minute terminating NIMh charging.
You can experiment with resistance to calibrate the temperature by reading it over USB with the Neato Control program.
There is still a question of what happens to the charging algorithm with its lower current extra timed topping off cycle after terminating according to temperature. The lithium set up will simply deliver another temperature signal. Whether this is sufficient wil have to be tested in practice, monitoring the charging behavior, looking for any oscillating behavior. The balancing boards provide an automatic bypass of this lower charging current without disturbing the battery.

Cell capacity is critical in one respect, supporting the charging rate, often delivered as a multiple of the Capacity, C, and the cells used in the old XV experiment were not as good as those in the Botvac project. The LiFePo4 also seemed to have a lower charging rate than discharge rate maximum spec -- they are very good at high load discharging but may not support fast charging the same as other lithium.
glnc222
Robot Master
 
Posts: 4913
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby glnc222 » March 12th, 2019, 9:25 pm

You must not raise temperature reading too high as there is an emergency fire shut down protection I think around 60C in the software.
glnc222
Robot Master
 
Posts: 4913
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby Ben_Affleck » March 13th, 2019, 7:26 am

I skimmed through the datasheet and ICL7665 seems like a good choice for the application.
Altough it seems to be an 1 and 0 solution. In this case the resistance would remain unchanged throughout the charging cycle, only to be switched lower at the end of charging.
Isn't Neato expecting a linear / exponential decrease in resistance throughout the charging cycle, like it receives with the thermistor? If it doesn't complain, than this is the easiest route.
Ben_Affleck
 
Posts: 12
Joined: February 26th, 2019, 7:39 am

Re: Neato lithium ion battery revisited

Postby glnc222 » March 13th, 2019, 7:37 pm

The Neato temperature algorithm does not appear to be very sophisticated, and is just monitoring temperature over time for a rate of increase over a minute reaching the trip value 1C/minute. The step function increase used with the ICL7665 worked to charge lithium batteries in the XV model.

You should have over-voltage protection in case it does not terminate the charger, an under-over protection board if not a balancer. The LiFePo4 has no fire hazard like lithium polymer but the cells can be damaged. They are also described as "self-balancing", though they should be assembled in balanced condition. I had an external separate-cell LiFePo4 charger for this purpose. Maximum voltage a little less than lithium polymer 4.2v/cell, check the specs.
glnc222
Robot Master
 
Posts: 4913
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby Ben_Affleck » March 16th, 2019, 9:21 am

Some charging routine findings on the Botvac D85:

These results are based on:
- Botvac D85 with original NiMh 3600mAh battery with some life left in it (on a fresh charge it still operates for around 45min before asking for a recharge)
- values from Neato control software over USB
- tested with battery completely emptied through Neato Control -> New Battery procedure (Neato Control reported ~14% fuel left, and about 10.5V under load)
- tested with normal recharge routine when the bot asks for it (Neato Control reported ~25% fuel left, and about 12.5V under load, then it requested to be recharged)

Going back to base to recharge:
~25% battery (~12V - 12.5V under load)

Initial charge:
- The bot charges with 1900 mAh
- Stops initial charge when the temperature rises with +1C in less then 1min15sec (>1min and <1min20sec). The battery voltage when it stops charging is between 15.5V and 16V

As a note: Just before it stops charging Neato Control reports anywhere from 60% to 75%. Immediately after it stops charging it jumps to 93-94%. I don't know if this is because of the bot firmware or the Neato Control software. It never ever shows 100% (keep in mind that for the test I am using a NiMh battery that is far from new..)

Topping off (doesn't matter how many top-offs, the behavior is always the same for every top-off):
- The bot starts charging when the battery voltage drops through self discharge under ~13.5V (Neato Control reports 89% battery)
- It starts charging with ~400mAh for about 10sec, afterwards it switches to charging with ~1900mAh and stays at this value until the end of charging occurs.
- Stops top-off charge when the temperature rises with +1C in less then 1min15sec (>1min and <1min20sec). The battery voltage when it stops charging is between 15.5V and 16V, Neato Control reports 93-94% charge (at least in my case)
Ben_Affleck
 
Posts: 12
Joined: February 26th, 2019, 7:39 am

Re: Neato lithium ion battery revisited

Postby glnc222 » March 16th, 2019, 6:58 pm

I am not sure you have the "topping off" term correct. Regular charging is at constant 2 amps. The "topping off" is a half-hour timed cycle at about 1/2 amp, trickle charge, not governed by temperature because none expected, though it might terminate on heating -- has to be tested. With the XV, these cycles were tracked with graphs from Germany shown earlier in thread, and some I made. Neato Robotics rep mentioned these extra low current cycles, which come from standard industry procedures for NiMh. I noticed they occur after the indicator light turns solid green, so that indicator does not mean complete end of charging. This extra cycle fills in a little bit of capacity remaining when the main charging cycle terminates from heating which occurs before full capacity is reached. Open air single cell chargers stop differently, on a small drop in terminal voltage, from changing resistance, at the full point. Enclosed packs get too hot to wait for that.

Repeated charging cycles as the battery discharges are at the full 2 amps.

An untested aspect of Neato design is whether the system draws from the battery when on the dock, causing repeated charging cycles wearing out the battery, or whether it is fully powered by the dock when parked. The system will run on the dock without the battery installed.
Even powered by the dock, self-discharge of NiMh can cause a repeated charging cycle at some point to maintain full capacity for cleaning on demand.
Lithium batteries have much lower self-discharge rates, avoiding shelf-life problems etc.
glnc222
Robot Master
 
Posts: 4913
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

Re: Neato lithium ion battery revisited

Postby Ben_Affleck » March 17th, 2019, 4:32 pm

I think you are right, it's not really topping off or trickle charging, because no extra energy is being added, but replaced.
It is more like a new charge cycle based on reaching a threshold (~13.5V) when self discharging.

In my case the self discharge is accelerated because the bot doesn't enter sleep mode while it is connected through the USB interface to Neato Control on PC.
This specific self discharge while connected to the PC is reported by Neato Control as battery discharge and it is about 90mAh, altough the bot stays also connected to the dock.

Test setup:
- the bot was connected to the dock the full time
- the bot was connected through USB to a PC with Neato Control the full time
- a software was used to take a printscreen every 5 seconds, the printscreens were automatically saved in jpeg form with the timestamp as a title
- after 2 days of monitoring in these conditions, thousands of printscreens were processed, looking for changes in temperature, voltage, charge percent and charging current.

Test timeline:
The bot was used normally until it requested to be recharged at the dock.
When connected to the dock and Neato Control, the bot started charging from 25% (~12.5V) until ~15.75V with ~1900mAh and stopped charging when the temperature raised with more then 1C in less then 1min15sec.
In about 2 hours it discharged to 13.5V while the bot was kept on by the USB connection (Neato Control reports ~90mAh discharge from the battery)
When 13.5V was reached the bot started charging with ~400mAh for 10sec and then it switched to charging with ~1900mAh for about 10min until the temperature raised with more then 1C in less then 1min15sec.
This type of cycle (discharge until 13.5V, recharge until +1C in less then 1min15sec) was repeated indefinitely.
No other type of discharge or recharge was observed. No trickle charging at a lower current, etc

I've done the same test starting with a fully discharged battery with the help of Neato Control feature of rejuvenating the battery and reseting the battery on the bot with the "New battery" feature. Besides starting the charge from a lower point, everything else happened in the exact same way as above.
Ben_Affleck
 
Posts: 12
Joined: February 26th, 2019, 7:39 am

Re: Neato lithium ion battery revisited

Postby Ben_Affleck » March 17th, 2019, 5:02 pm

Btw, I have 2 D85's, both with NiMh batteries:
- One is used with the return to base feature at it was meant. It cleans, it returns to base, recharges, continues work or waits on the dock for the next scheduled clean cycle.
- The other one is manually charged when needed. I'm placing it on the dock when I want to charge it, right before I want it to clean the house. When the light on it goes full green, I take it off the dock and I activate the clean house feature. After it cleans the house, if it still has juice I don't charge it. I will still use it for further cleaning cycles until it asks for a recharge. I don't let him dock and I keep it in this state until the next planned cleaning cycle, when I am going to manually place it on the dock again. So these are more like full-to-empty battery cycles, and the bot stays on the dock only while charging from empty-to-full.

The first one is the one with the battery that just died.
The second one is the one that can still work for 45min before it asks for a recharge.
Both started their life in the same time.

So I would say that keeping it always on the dock stresses the battery more, because of all the small recharges that compensate for the self discharges.
Ben_Affleck
 
Posts: 12
Joined: February 26th, 2019, 7:39 am

Re: Neato lithium ion battery revisited

Postby glnc222 » March 17th, 2019, 9:22 pm

Note that the maximum battery terminal voltage for NiMh can be a couple volts higher than the actual battery voltage when full, to push in (or through to heat) charge. i.e. on XV with 14.7v nominal pack, charging terminates at 18.5v but the battery falls quickly to 16.5v starting point of discharging curve after charging current stops -- the drop from 18.5 is not discharge, just stabilization after charging. The 12v Botvac will be similar but proportionately lower. The trickle charge topping off cycle might be delayed until this stabilization is complete, would have to test; or also until the temperature falls to a normal level.

Hope you can post schematics of any adapter circuits you develop.
glnc222
Robot Master
 
Posts: 4913
Joined: January 23rd, 2012, 9:19 pm
Location: North Carolina, U.S.

PreviousNext

Return to Neato Robotics

Who is online

Users browsing this forum: clone1008, fufu, jdong, johnnyh and 173 guests