New Batteries for Thrift Store Neato Vacuums

The Neato XV-21 I found at a thrift store has bee joined by a XV-12 found by [Emily]. Both Neato robot vacuums found at thrift stores have degraded batteries. This was not a surprise as robot vacuums work their battery packs hard. Not just frequent charge and discharge cycles, but with a heavy power draw under use to run vacuum motors. This means both of these vacuums were retired by their previous owners when the battery no longer hold enough charge to perform its duties.

Neato XV-12 battery unhappy

Fortunately, this common failure also meant there’s a robust aftermarket for replacement batteries. Curiously, the economics of the markets are such that whole replacement packs can be ordered online for less than ordering individual cells and rebuilding the packs myself.

With previous experiments, I have gained confidence I can verify functionality of individual components using test mode accessible via USB port. And since this XV-12 was found with an official Neato charging dock, it’s time to install replacement batteries and test the full system.

The replacement batteries claim a capacity of 4000mAh which, on paper, is an increase from the original battery’s label capacity of 3800mAh. However, battery manufacturers play pretty loosely with these ratings so I expect the difference of 200mAh to be fairly insignificant in practice. When I took apart the original pack, I saw a thermister for monitoring temperature, an overcurrent protection fuse, and an overheat fuse. I assume the replacement pack has a thermister because the Neato computer can read it, but there’s no immediate way to tell if the overcurrent or overheat protection also exists on the new pack.

Neato XV-12 batteries old and new

With new battery packs installed, it’s time to put the robot up against the charging dock and verify the charging system works as expected. But before we do that, let’s take a closer look at this charging dock.

Cleaning Up a Thrift Store Neato XV-12

When I looked over my Neato XV-21, I thought it was far too clean to be a high mileage appliance. There were several possibilities:

  1. The previous owner was meticulous about keeping things clean (supported by the plastic protective wrap.)
  2. The Neato was barely used.
  3. A Neato is very good at keeping itself clean.

Thanks to a second Neato found in a thrift shop, this time a XV-12, we now know #3 is false. This one is well used, and very much looked it! The previous owner didn’t bother cleaning the vacuum before dropping it off at the thrift store. I have several loops of hair tangled up in the roller brush. The largest loop has even snagged what looked like furniture padding foam.

Neato XV-12 dirty brush roller

The largest loop has also tightened enough to damage the roller brush. Once I cut that loop of hair and debris off the roller brush, I see a slot cut into the rubbery blade.

Neato dirty brush roller ripped

There was also debris tangled on the motor shaft driving this roller that had to be cleaned off before the roller brush could function properly again. The cavity for the roller brush showed extensive wear from a life as a working household vacuum (above), whereas the XV-21 showed barely any wear (below).

Neato wear pattern

Those were problems found on the suction side of the vacuum. What about the exhaust side? The vacuum is generated by a fan which sat downstream from the dust bin filter, so its cleaniness would be a clue at how effective the filter was. Everything seen here are dirt particles that have made their way past the filter.

Neato XV-12 dirty vacuum motor

But even though this vacuum lived a harder life, its battery was actually in better condition and could run the vacuum computer for several minutes before fading out. As a result I was able to query all components individually using its USB port without having to hack a battery into this vacuum. All signs indicate that this robot vacuum is likely fully functional except for its battery.

And fortunately, with the arrival of replacement battery packs, we don’t need to drill holes and hack external battery packs for a full system test. We can install the new batteries into this XV-12.

 

Thrift Store Neato XV-12 Joins XV-21

I thought I was pretty lucky to find a Neato robot vacuum in a thrift store for $8. It didn’t work in the store and that’s why it was cheap, but I have since determined it was fully functional except for its battery pack. While waiting for its replacement battery pack to arrive, Player 2 has entered the game! [Emily] managed to find another Neato robot vacuum in a different thrift store. The new find is a model XV-12 and it included the charging dock for $11.

XV-12 And XV-21 A Pair of Neato

A little web research indicated that these two robot vacuums are contemporaries, both followed up Neato’s XV-11. The XV-12 is the direct successor that replaced the XV-11, and the XV-21 is a premium offering sold simultaneously with the XV-11. Aside from the cosmetic difference of purple plastic top on the XV-12, there are a few functional differences.

The cleaning brush roller is different. The XV-12 uses a bidirectional design with flat flexible plastic blades. The XV-21 uses a unidirectional design with a combination of flexible plastic blades and bristles. The XV-12 brush can be mounted in either direction – note that geared teeth to engage the toothed belt on both sides of the roller. The XV-21 is designed to spin a specific direction – brushing debris towards the center of the vacuum – and only has geared teeth on one end of the roller because it won’t work properly if mounted the other way.

Neato brush comparison

The dust bin filters are also different between these two models. While the XV-12 uses a flat sheet, the XV-21 uses a pleated design for greater surface area. This partially compensates for the more restrictive filter used in a XV-21 that captures more particles from the vacuum air stream.

Neato filter comparison

The XV-21 was sold as the upgrade model. Its roller brush with curved flexible blades and bristles combine to pick up more dirt, and its pleated filter keeps more of that dirt in the dust bin instead of passing it into the exhaust. These two differences reportedly improve vacuum capability at the cost of greater power consumption which translates to shorter run time on battery.

In addition to the design differences between the XV-12 and the XV-21, there are additional differences between these two specific thrift store finds. The XV-21 was surprisingly clean hinting at a very low usage in its previous life, but the XV-12 shows signs of a well-used robot vacuum.

 

Battery Replacement Options for Thrift Store Neato XV-21

With a successful all-up system test of Neato XV-21, running on hacked-up battery packs through a full vacuum cycle, we have confidence everything works on this thrift store bargain except for the battery pack. So focus returns to these NiMH battery packs.

Neato XV-21 battery pack

The default option is to buy some battery packs specifically built and sold as replacement batteries for a Neato robot vacuum. A straightforward Amazon search for “Neato battery pack” will return this “Amazon’s Choice” item at $30 for 4000mAh NiMH packs.

A tempting choice is to purchase some lithium-ion batteries that would also fit in the Neato battery bay. However, battery packs sold on Amazon are typically designed for very high draw applications like multi rotor aircraft. Batteries designed to survive such use tend to have lower energy density. Thus lithium packs that fit within the bay’s dimensions actually don’t have much significantly higher capacity than 4000mAh.

Another concern with installing lithium batteries is the fact the Neato’s on board charging circuits are designed for NiMH battery cells, which has a different charging profile. Charging lithium cells as if they were NiMH cells risks damaging the lithium cells and possibly lighting them on fire.

There is, however, this item which are lithium replacement packs intended and designed as an upgrade option for Neato vacuums. Not only are they sized appropriately to fit, they also have an integrated battery management system. Presumably, the circuit will make the Neato brain think there’s a NiMH battery pack installed but treat the lithium cells properly. At 4400mAh it offers 10% higher claimed capacity relative to default option, but at over double the cost, the cost/performance ratio is poor.

There is also the option to purchase new NiMH battery cells and rebuild the battery pack myself, reusing all the associated parts from the existing pack. I had expected this to be the lowest-cost option by supplying my own labor, but when I searched for NiMH battery cells in 4/3A form factor with 4000mAh capacity, I was surprised to find that they were selling for more than $3 per cell. There are 6 cells per pack and a Neato requires two packs, for 12 total cells. This means buying raw cells and rebuilding them myself would cost over $36, more than just buying a prebuilt pack from Amazon!

With this little bit of research into lithium upgrade option and rebuild option, it looks like the default $30 option is the best way to go. But before that replacement battery pack arrived, my Neato got a friend!

Mounting External Batteries on Thrift Store Neato XV-21

Using small batteries I hacked into existing battery bays, I was able to query sensor status and command individual motors. However, those small batteries were not powerful enough to run multiple motors simultaneously, which meant a full system test would not be possible until larger batteries are installed. At this point I could order some Neato-specific replacement batteries and have decent confidence it will work, but I would like additional confirmation before I spend money.

But I didn’t have any batteries that were more powerful and still small enough to fit within Neato vacuum’s existing battery bays. This meant batteries had to be installed externally. And if they’re going to be external, I might as well go with the biggest batteries I have on hand: those I purchased for Sawppy the rover.

But where could I install those batteries?

Neato XV-21 with two big power packs

The most obvious place: is to put them on top of the dust collection bin. This is a popular location for Roomba battery retrofits, because it preserves weight balance and does not hinder Roomba operations. However, it won’t work for a Neato: this position blocks the line of sight for the laser distance scanner.

Neato XV-21 with two big power packs on top of dust bin

Installing the batteries on the front bumper would be out of lidar line of sight, and might help improve vacuum performance because their weight would help push the dirt brush roller into the ground. But this would change the behavior of front bumper which might confuse vacuum mapping algorithms, and using fragile batteries as your front bumper is never a good idea.

Neato XV-21 with two big power packs in front of bumper

If we put batteries off to the sides, it preserves front clearance but now it interferes with side clearance. The vacuum would no longer be able to follow along walls closely to vacuum near walls if the batteries are attached to the sides.

Neato XV-21 with two big power packs along sides.jpg

Installing behind the vacuum disturbs the weight balance – battery weight is now trying to lift the brush roller off the ground. The back is not flat, making installation difficult. And that circular shape is there for a reason – its clearance helps the robot turn in tight spots. Batteries install behind the robot would get bashed against obstacles as the robot turned.

Neato XV-21 with two big power packs behind

And obvious we don’t have any way to mount batteries on the underside of the vacuum, as there is only about 5mm of ground clearance. That exhausts all the potential directions we can go for external mounting.

Time to get creative.

Rethink top mounting, we recognize the problem is that we can’t block laser distance scanner’s line of sight. It looks out all around the top of the vacuum, except for one place: the raised protective cap above the lidar housing. Batteries mounted on top of that cap would not block laser scanner’s line of sight.

Neato XV-21 with two big power packs above lidar

Wires were run from battery compartment, which required drilling two small holes. To keep the power wire out of sight of lidar, the wires were routed through existing grooves on top of vacuum body and then followed existing  support pillars for lidar housing. This way, the wires should not introduce additional blind spots.

Neato XV-21 wire routing for two big power packs.jpg

With this hack, the robot vacuum can run on my pair of Sawppy rover batteries, which are plenty powerful enough to run all vacuum systems simultaneously. Now this little Neato is alive and can run through a full vacuum cycle, verifying that all systems worked.

Neato XV-21 up and running with two big power packs

Obviously, this won’t be the final long term solution. For one thing, Sawppy wants its batteries back. For another, the additional height on top of the robot hampers its ability to get under furniture for vacuuming, and the exposed wires are vulnerable to tangling on protrusions. What we need next are batteries powerful enough to run a Neato and fit within existing battery bays.

Sending Commands to Neato XV-21 Via USB

I’ve managed to establish a serial connection to my Neato XV-21 robot vacuum’s USB port, putting it into test mode and querying sensor status. The next step is to start issuing commands to see if individual components respond. We left off on querying control panel user interface buttons, so the next test is to see if I can draw my own user interface on that control panel screen. Typing in help setlcd retrieved the following information:

SetLCD - Sets the LCD to the specified display. (TestMode Only)
BGWhite - Fill LCD background with White
BGBlack - Fill LCD background with Black
HLine - Draw a horizontal line (in foreground color) at the following row.
VLine - Draw a vertical line (in foreground color) at the following column.
HBars - Draw alternating horizontal lines (FG,BG,FG,BG,...),
across the whole screen.
VBars - Draw alternating vertical lines (FG,BG,FG,BG,...),
across the whole screen.
FGWhite - Use White as Foreground (line) color
FGBlack - Use Black as Foreground (line) color
Contrast - Set the following value as the LCD Contrast value into NAND. 0..63

This is enough to draw simple test patterns, but it isn’t enough for me to put up legible text prompts for users. This was not a surprise as I understand this mode to be a diagnostics console and not intended to facilitate a completely new UI like what I wish to do. Still, I can probably convey some simple if cryptic information by drawing horizontal and vertical lines. The valid range is 0-127 inclusive. And while vertical lines across that entire range are all visible, horizontal lines beyond 124 seem to go under bottom lip of display bezel and not visible.

setlcd bgwhite
setlcd hline 20
setlcd vline 50
setlcd vline 100

Neato Line Art

I’ve already played with turning the laser distance scanner on and off. The next most important thing to get running on this chassis are the drive wheels. If I can’t make the robot move using this interface, nothing much else matters. Here are the commands listed under help setmotor:

SetMotor - Sets the specified motor to run in a direction at a requested speed. (TestMode Only)
LWheelDist - Distance in millimeters to drive Left wheel. (Pos = forward, neg = backward)
RWheelDist - Distance in millimeters to drive Right wheel. (Pos = forward, neg = backward)
Speed - Speed in millimeters/second. (Required only for wheel movements)
Accel - Acceleration in millimeters/second.
(Used only for wheel movements. Defaults to 'Speed'.)
RPM - Next argument is the RPM of the motor.
Not used for wheels, but applied to all other motors specified in the command line.
Brush - Brush motor forward (Mutually exclusive with wheels and vacuum.)
VacuumOn - Vacuum motor on (Mutually exclusive with VacuumOff)
VacuumOff - Vacuum motor off (Mutually exclusive with VacuumOn)
VacuumSpeed - Vacuum speed in percent (1-100).
RWheelDisable - Disable Right Wheel motor
LWheelDisable - Disable Left Wheel motor
BrushDisable - Disable Brush motor
RWheelEnable - Enable Right Wheel motor
LWheelEnable - Enable Left Wheel motor
BrushEnable - Enable Brush motor
SideBrushEnable - Enable Side Brush Motor motor
SideBrushDisable - Enable Side Brush Motor motor
SideBrushOn - Enable the Side Brush
SideBrushOff - Disable the Side Brush
SideBrushPower - Side Brush maximum power in milliwatts

The first few attempts – using just individual commands – were met with either errors or no movement.

setmotor lwheeldist 100
No recognizable parameters

The first successful command that triggered movement was one where I specified three parameters on a single line: left wheel distance, right wheel distance, and speed.

setmotor lwheeldist 200 rwheeldist 200 speed 100

The robot moves! This is great news, but we still have two more motors to play with. First up is the brush roller: I had expected the motor to be a digital on/off affair, or maybe something I can command with a particular motor power level. I was quite pleasantly surprised it can be commanded to a particular RPM because it implied a more sophisticated closed-loop feedback control system.

setmotor brush rpm 750
Run Brush Motor @ 750 RPM

The final motor control for vacuum suction fan was a little frustrating. I had no issue turning it on with setmotor vacuumon and then back off with setmotor vacuumoff. But I haven’t yet figured out the right syntax to command it to spin at a partial power level.

setmotor vacuumspeed
Value required for option VacuumSpeed
setmotor vacuumspeed 50
No recognizable parameters
setmotor vacuumspeed 50%

Invalid Command: 'setmotor vacuumspeed 50%'

I’ll update this post if I ever figure it out, but for now it is enough that it spins.

With this set of tests, I have determined that all individual components are functional, but only in a limited context. The limitation is because I could power just one device at a time with the small batteries I have installed. What I want next is a full system test. And for that, I will need to install more powerful batteries.

Query Neato XV-21 System Status Via USB

With the experimental batteries in place, the computer boots up and stays up for longer than three seconds. I tried to run a vacuum cycle but that was too much for these batteries to handle, so for now I’ll stick with digital exploration starting with the standard UI to dump revision information on components inside this particular Neato XV-21 robot vacuum.

Neato XV-21 info

After that, it’s time to go exploring through its USB serial port diagnostics!

I went straight to the most interesting component: the laser distance scanner module. First I had to put my vacuum into test mode with testmode on. Followed by setldsrotation on to turn on the motor that sweeps the laser. And finally, fetch a set of laser distance data with getldsscan. This gave me a large set of numbers. According to the legend given in the first line of the response, a distance and intensity number per degree of sweep. Partial excerpt below:

AngleInDegrees,DistInMM,Intensity,ErrorCodeHEX
0,2378,82,0
1,2397,88,0
2,0,88,8021
3,2494,85,0
4,2505,82,0
5,2509,12,0
[…]
355,2289,99,0
356,2303,103,0
357,2303,99,0
358,2315,94,0
359,2367,91,0
ROTATION_SPEED,5.10

Future experimentation will determine what range of distance values are typical, similarly for intensity values. A quick web search failed to find a reference guide for error hex codes, but hopefully there’ll be a small set that I can infer from context. More exploration to come!

In the meantime, I repeated the getldsscan command to verify the values are getting updated. They were – I received a similar but slightly different set of numbers. Hooray! the most important sensor looks like it is functioning.

Next item on the exploration list: querying analog sensor values with getanalogsensors.

SensorName,Value
WallSensorInMM,63,
BatteryVoltageInmV,16120,
LeftDropInMM,60,
RightDropInMM,60,
LeftMagSensor,-2,
RightMagSensor,-3,
UIButtonInmV,3317,
VacuumCurrentInmA,0,
ChargeVoltInmV,132,
BatteryTemp0InC,34,
BatteryTemp1InC,31,
CurrentInmA,622,
SideBrushCurrentInmA,29,
VoltageReferenceInmV,1225,
AccelXInmG,-8,
AccelYInmG,-12,
AccelZInmG,1092,

I see three distance sensors – wall and two drop sensors, one in each front corner. It’s not immediately clear what “Mag” in left and right “MagSensor” referred to. Do they detect magnetic fields? Or are they measuring magnitude of something? Bottom three lines indicate the vacuum sensor suite includes a three-axis accelerometer which can measure in units of thousands of G, implied by a Z value nearly 1000 in this vacuum sitting on the floor. Remainder of the values are related to power management.

The 3-axis accelerometer values above from getanalogsensors partially overlap with results of getaccel. As the name indicates, this one is completely focused on accelerometer readings and includes results of additional calculation in the form of pitch and roll in degrees and a sum of acceleration vector.

Label,Value
PitchInDegrees, -0.64
RollInDegrees, -0.53
XInG,-0.012
YInG,-0.010
ZInG, 1.079
SumInG, 1.079

The list of digital sensor values from getdigitalsensorsis shorter, and for whatever reason, labelled in all capital letters.

Digital Sensor Name, Value
SNSR_DC_JACK_CONNECT,0
SNSR_DUSTBIN_IS_IN,1
SNSR_LEFT_WHEEL_EXTENDED,0
SNSR_RIGHT_WHEEL_EXTENDED,0
LSIDEBIT,0
LFRONTBIT,0
RSIDEBIT,0
RFRONTBIT,0

Here we can see whether the dustbin is installed, whether either or both wheels are at full extension, and it looks like there are four switches associated with the front bumper. The top line shows if a charging plug is connected, but that’s only one of many parameters around power management. Most of the rest are in getcharger.

Label,Value
FuelPercent,88
BatteryOverTemp,0
ChargingActive,0
ChargingEnabled,0
ConfidentOnFuel,0
OnReservedFuel,0
EmptyFuel,0
BatteryFailure,0
ExtPwrPresent,0
ThermistorPresent[0],1
ThermistorPresent[1],1
BattTempCAvg[0],33
BattTempCAvg[1],31
VBattV,16.03
VExtV,0.16
Charger_mAH,0

If I were to create my own control logic for driving my Neato around, the most important parameter here is FuelPercent. I’ll have to be responsible about not draining the battery too far, beyond that I can leave the rest in the capable hands of Neato’s existing battery management system. Speaking of power consumption, the biggest drains are the motors, and I can monitor them all with getmotors.

Parameter,Value
Brush_RPM,0
Brush_mA,0
Vacuum_RPM,0
Vacuum_mA,0
LeftWheel_RPM,0
LeftWheel_Load%,0
LeftWheel_PositionInMM,-1
LeftWheel_Speed,0
RightWheel_RPM,0
RightWheel_Load%,0
RightWheel_PositionInMM,-1
RightWheel_Speed,0
Charger_mAH, 0
SideBrush_mA,28

Not very exciting at the moment, because all motors are at a stop. I imagine these numbers will get more interesting once the robot gets underway.

One final discovery in inputs: when in test mode, the user interface buttons on the top of the robot no longer trigger their usual functions. I could check for status of all five buttons via getbuttons which implies I could use these buttons in my own projects without worrying that I’ll trigger the default vacuum behavior. Cool!

Button Name,Pressed
BTN_SOFT_KEY,0
BTN_SCROLL_UP,0
BTN_START,0
BTN_BACK,0
BTN_SCROLL_DOWN,0

But if I want to actually have my own user interface using those buttons, I would need to also display my own information on the LCD. Which brings us to phase 2 of playing over USB: start sending commands to see if components follow orders.

USB Serial Communication with Thrift Store Neato XV-21

My Neato XV-21 robot vacuum now boots up and stays running on a pair of old remote control helicopter batteries. This is a vast improvement over its comatose state when I found it in a thrift store. I pressed the start button to see if it’ll actually vacuum, but spinning up the motors drew too much current and aborted. Looks like these batteries are only good for probing electronics, which is still more than what I had before. And a good incremental step forward.

A few web searches on Neato technical details all pointed to posts on “Neato Robotics” forum on robotreviews.com. I guess this is where all the Neato robot tinkerers hang out. From this forum I learned of a tool for Neato maintenance that can help communicate with the vacuum as well as uploading firmware updates. Unfortunately, this forum also shared an update that Neato has taken these tools off their website. Without them, plugging the vacuum into my laptop running Windows would only result in a device without a driver.

Neato mini USB cable connection to laptop.jpg

On a lark, I rebooted my laptop into Ubuntu Linux and plugged in the vacuum. There were never any Neato drivers for this operating system, and I was curious what I could see via Linux tool dmesg.

[10547.714901] usb 1-1: new full-speed USB device number 25 using xhci_hcd
[10547.866228] usb 1-1: not running at top speed; connect to a high speed hub
[10547.876232] usb 1-1: New USB device found, idVendor=2108, idProduct=780b
[10547.876235] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[10547.876237] usb 1-1: Product: Neato Robotics USB v2
[10547.876239] usb 1-1: Manufacturer: Linux 2.6.33.7 with fsl-usb2-udc
[10547.881256] cdc_acm 1-1:2.0: ttyACM1: USB ACM device

Well, that was more straightforward than I had expected. The ACM in ttyACM1 here stands for abstract control model. The operating system sees a communication device where all control is handled by the device, and all I had to do is treat it like a serial port. It’s not a true serial port, but close enough the technical differences aren’t important right now. What matters is the fact that I can run minicom --device /dev/ttyACM1 and issue a simple call for help. We are in business! The channel has been opened to talk to a Neato vacuum brain and see what it says in return.

help
Help Strlen = 1792
Help - Without any argument, this prints a list of all possible cmds.
With a command name, it prints the help for that particular command
Clean - Starts a cleaning by simulating press of start button.
DiagTest - Executes different test modes. Once set, press Start button to engage. (Test modes are mutually exclusive.)
GetAccel - Get the Accelerometer readings.
GetAnalogSensors - Get the A2D readings for the analog sensors.
GetButtons - Get the state of the UI Buttons.
GetCalInfo - Prints out the cal info from the System Control Block.
GetCharger - Get the diagnostic data for the charging system.
GetDigitalSensors - Get the state of the digital sensors.
GetErr - Get Error Message.
GetLDSScan - Get scan packet from LDS.
GetMotors - Get the diagnostic data for the motors.
GetSchedule - Get the Cleaning Schedule. (24 hour clock format)
GetTime - Get Current Scheduler Time.
GetVersion - Get the version information for the system software and hardware.
GetWarranty - Get the warranty validation codes.
PlaySound - Play the specified sound in the robot.
RestoreDefaults - Restore user settings to default.
SetFuelGauge - Set Fuel Gauge Level.
SetMotor - Sets the specified motor to run in a direction at a requested speed. (TestMode Only)
SetTime - Sets the current day, hour, and minute for the scheduler clock.
SetLED - Sets the specified LED to on,off,blink, or dim. (TestMode Only)
SetLCD - Sets the LCD to the specified display. (TestMode Only)
SetLDSRotation - Sets LDS rotation on or off. Can only be run in TestMode.
SetSchedule - Modify Cleaning Schedule.
SetSystemMode - Set the operation mode of the robot. (TestMode Only)
TestMode - Sets TestMode on or off. Some commands can only be run in TestMode.
Upload - Uploads new program to the robot.

Replacement Battery Test for Thrift Store Neato XV-21

My Neato XV-21 vacuum couldn’t power on when I found it in a thrift store, and the problem (or at least, the first problem) was its battery pack: it could only deliver enough power to boot up the onboard computer and run it for about three seconds before its voltage dropped and the computer shut down again.

In order to assess the rest of the vacuum, I need to find a replacement power source. Since the actual condition of vacuum components are still unknown, I’m not inclined to spend money buying batteries specific to a Neato robot vacuum. The next experiment, then, is to wire up something out of batteries already on hand. I looked through my pile of batteries to see which would satisfy the following criteria:

  • Delivers roughly 7.2 volt power of a six-cell NiMH battery.
  • Fits within Neato battery compartment.
  • A pair of identical batteries, one for each bay.

The winner was a pair of batteries I had for my E-Flite remote control helicopter. The aircraft itself had long since been crashed and trashed, but I kept its batteries. Two cell lithium polymer batteries have nominal voltage of 7.4 volts, within operational range of 6-cell NiMH batteries. Physically they were small enough to fit within the battery compartment, but at 800 mAh their power capacity is too low for full vacuum operation. They were designed to deliver high power to fly a small helicopter toy though only for a few minutes. Ideally they can run the vacuum motors briefly to verify their operation, but mostly I just need them to run the Neato computer for more than three seconds.

To connect these batteries, I first disassembled the old tired original battery pack. Before disassembly, I noticed a rectangular bump visible in the battery pack. I had thought this was the thermistor but I was wrong. It appears to be a temperature safety fuse. There was also an over-current protection amperage fuse in the pack. NiMH cells are relatively safe but it’s still good to see additional safety measures inside these packs.

 

I needed the battery pack’s wiring harness for my battery experiment, so it was cut free. I kept the temperature monitoring thermistor and wired the positive and ground battery power wires to a JST-RCY connector that will mate with my E-Flite battery pack.

E-flite Pack for Neato Vacuum.jpg

Once soldered, I connected the pair of batteries and plugged them into my Neato XV-21.

Neato XV-21 with adapted E-Flite LiPo batteries

I was greeted by the same boot-up screen I saw before, but this time the screen stayed on instead of flickering off after three seconds. Now it’s time to try talking to the vacuum.

 

Thrift Store Neato XV-21 Batteries Can’t Hold a Charge

The batteries from a Neato XV-21 were found to be flat and possibly damaged. This would certainly explain why the little robot vacuum failed to power on in the thrift store where I found it. Since the thrift store didn’t have its associated charging base, my fallback was to leave its batteries to trickle charge (at 100 milliamps) overnight on a bench power supply.

The battery packs nominal voltage was listed at 7.2 volt per pack, so the voltage limit for charging the pair in series was set to 14.4 volts. Fully charged six-cell NiMH battery packs can be up to 9 volts each, for 18 volts total, but setting to 14.4 volts leave us with a bit of headroom in case these batteries didn’t behave as expected. This turned out to be a good idea.

Neato battery pack on power supply

When I checked in on the batteries the next morning, I saw voltage sitting at 14.4 and amperage had dropped to 40 milliamps. If all goes well, the voltage would be evenly divided across the two packs at 7.2 volts each. This was not the case: one battery pack (I’ll arbitrarily name it pack A from now on) was holding at 7.9 volts and the other (now designated pack B) at 6.5. If it weren’t for the headroom, we might have inadvertently pushed pack A beyond its limits.

At this point I got distracted by other events and left the batteries sitting for a week. Checking in from time to time, I measured their open circuit voltage and wrote them down. It was an unplanned test of the batteries’ self-discharge rate.

Pack A:

  • Overnight trickle charge: 7.9 Volts
  • After one day: 4.67 V
  • After three days: 4.44 V
  • After a week: 2.0 V

Pack B:

  • Overnight trickle charge: 6.5 Volts
  • After one day: 2.77 V
  • After three days: 2.68 V
  • After a week: 2.61 V

While it is normal for NiMH batteries to self-discharge over time, this observed discharge curve was far more severe than is acceptable. Looking at the voltage discharge behavior for pack A, it appears to have two almost-useless cells, three weak cells, and one good cell. Pack B behavior hints at four almost-useless cells and two good cells.

But even if they are unusable to actually run the vacuum, they are still better than dead weight: these batteries can hold a charge for a very brief period. So they’re reconnected to the power supply to trickle charge again. I don’t expect them to actually power any motors, but perhaps there’d be enough to power on the electronics for a short while. Hopefully long enough for me to assess if the rest of the vacuum is healthy.

After another overnight charge, I plugged them into the vacuum and pushed the power button. Good news: they did power on the electronics for a short while. Bad news: It only lasted for three seconds, far too short to assess anything else about the vacuum. But the vacuum computer did boot, and that’s enough motivation to keep going.

Attempt to Charge Battery of Thrift Store Neato XV-21

We’ve pulled the battery packs out of a Neato XV-21 found in a thrift shop. We’re pretty sure the red and black wires are battery positive and ground, but if so we expect to see a few volts across those wires. The fact we measured less than a volt indicates this battery is flat and may even be damaged. Well, at least this would be consistent with the fact it wouldn’t power on in the thrift shop.

Fortunately, rechargeable NiMH battery cells are hardier than lithium batteries and they might possibly revive with a charge. And even if they were permanently damaged, they are less likely to turn themselves into fireworks as damaged lithium batteries sometimes do. Since we read a nonzero voltage, none of the cells have failed as an open circuit. Some might have failed as a short circuit acting as a resistor dissipating energy but not holding the charge. With this unknown, we shouldn’t charge the batteries anywhere near as fast as we could if they were known to be healthy.

Neato battery pack on power supply

A bench top power supply was set to a maximum voltage of 14.4 volts and maximum amperage of 0.1 amp (= 100 milliamps). Using a piece of wire, the two battery packs were connected in series. (Which is where the 14.4 came from: adding up their nominal 7.2 volts per pack.) Once connected in series, they were connected to the bench top power supply and we watched the meters.

If we instantly hit maximum voltage and maximum amperage, it indicates a short circuit somewhere, possibly in a dead battery cell acting as a resistor turning that 1.44 watts into heat. Fortunately this did not happen. We hit the 100 milliamp limit immediately as expected and the voltage started rising, which is consistent with a battery pack trying to charge.

After a few minutes, we disconnected the power supply and measured battery pack voltage: they now measure a few volts, higher than before. This increased confidence that our positive/negative wiring guesses were correct and also proof at least a few cells can hold some amount of charge.

The batteries were reconnected and, for the first half hour, we periodically touched each of the cells to see if they were hot to the touch. Some of the cells were warmer than others, but none were alarmingly so. We’ll let the power supply continue slowly feeding the battery pack overnight and see how they behave.

Examining Battery of Thrift Store Neato XV-21

This Neato XV-21 robot vacuum did not power on when I found it in the thrift store. Now that I’ve verified there isn’t anything fundamentally broken or missing on the robot, the next thing is to try to charge its batteries. But since the thrift store didn’t have its charger, we have to go to plan B: Pull out the batteries and see if we can work with it.

The interior component packaging of this vacuum divides its battery into two packs, each in their own battery bay on either side of its dust collection chamber. These two bays are accessible from the bottom of the vacuum, each with a door that opened by unscrewing a pair of Philips screws. Each battery pack has a 4-wire pigtail terminating in a connector that plugs into the vacuum.

It takes quite a bit of force to remove the connector. One battery was removed without incident, but the other battery required so much force it actually broke something holding the vacuum side connectors to the vacuum body. Fortunately this appears to only be a minor mechanical annoyance and will not derail the project. At worse, we now have a connector that will rattle as the robot moves around.

Neato XV-21 battery pack

Once the battery pack was removed we could read its printed information, identifying itself as a nickel-metal hydride (NiMH) battery pack. The general shape implies six cells, and the stated nominal voltage of 7.2 is consistent with six cells in series. Each cell is rated with a capacity of 3200 milliamp-hours. I didn’t know what “4/3A” meant but a quick web search indicated it was the physical form factor (size) of the battery. “McNair” appears to be the name of a battery manufacturing company, and finally “MC20120924” probably meant this battery was manufactured on September 24th, 2012.

Attention then turned to the four-wire connector. Typical convention hints black is probably negative and red is probably positive. A voltmeter said there was less than a volt across those wires, which is not just flat but might also indicate damaged battery cells. The next question: What are the two yellow wires? It is something the vacuum control board would want to know about, and our hypothesis is that it is a thermistor to measure battery’s temperature. As an experiment to test this hypothesis, we put a ohm meter across the two yellow wires and warmed up the battery pack with our hands. We saw resistance change in response to temperature.

With some confidence the yellow wires are not directly involved in charging, it’s time to put a little bit of power across those red and black wires to see what happens.

Digging Into Thrift Store Neato XV-21

Of course, I didn’t buy a nonfunctional robot vacuum from a thrift store just to admire how clean it is. I bought it to see if it can be brought back to life. So out comes the screwdriver set and it’s time to dig into those non user serviceable parts inside. Fortunately Neato didn’t feel the need to complicate this effort with “security” bits that would only hamper resourceful hackers by a few seconds – they were straightforward Philips head screws. I had expected fine pitched machine screws, but they were actually coarse pitched screws that self-tap into plastic. This limits the number of times we can assemble and disassemble this vacuum before the screws become too loose to hold themselves in place, but that’s a concern for the future.

Neato XV-21 with bottom panels removed

There are three easily accessible panels. Two rectangular battery compartment doors each held by two screws, and a large semicircular panel held in place by four screws.

Underneath the semicircular panel is a large centrifugal (squirrel-cage) fan responsible for creating the suction that gives vacuum cleaner their name and purpose. The filter sitting in the back end of the dust bin is visible. Air would move past the filter, through the fan, and exhaust out the grating at the back. It would not have been surprising to find a coat of fine dust that made their way past the filter, but the fan compartment is pristine. Either the filter is far more capable than I give it credit for, or this vacuum was indeed barely used.

Neato XV-21 wheel encoder

On either side of the fan are motor gearbox assemblies for driving the vacuum around. The motor appears to be a commodity DC brushed motor, but it does have some sort of encoder mounted to its back. Looking at the circuit board we saw a single sensor, indicating this encoder can sense motor movement but not its direction: a full quadrature encoder would have had two sensors on the board.

Also visible behind the robot’s left wheel (right side of top picture) are wiring for external connectors:

  • Mini-USB for serial data, which would be extremely interesting if I could get the vacuum powered up.
  • Barrel jack for a charging adapter I don’t have.
  • Two thick metal wires to make contact with a Neato charging dock, which I also don’t have.

Given that the official chargers were absent from this thrift store purchase, the next step is to remove these two battery packs and see what we can do with them.

A Curiously Clean Thrift Store Neato XV-21

I went to a thrift store to take pictures, but came home with a Neato XV-21 vacuum that did not power on. At $7.99, it was a cheap gamble to see what might be usable in this robot vacuum cleaner. In my cursory in-store examination, it looked to be in good shape. After I took a little time for a closer look, I realized it was actually in even better shape than I had initially thought.

Super clean Neato XV-21

Of course, there’s no mistaking the vacuum for new. There is dirt in the dust bin and visible black smudges on top of the vacuum. But beyond that I found almost no other signs of wear. The exterior sides of this unit are clean, free of the scuff marks and other marring I associate with robot vacuum veterans. Although my experience are mostly with affordable models of Roomba and similar vacuums that perform a random walk through its environment, bumping into things along the way. The selling point of a Neato is its lidar for object avoidance, so the nearly pristine sides of may just be a result of excellent object avoidance and not lack of use.

Super clean Neato XV-21 brush

Its floor cleaning brushes show some dirt but is actually really clean. My own home vacuum would show more dirt than this after a minute of use. Brush roller drive belt looks to be in good shape.

At this point I’m still undecided whether this was a barely used robot vacuum or maybe the previous owner is just very meticulous about keeping their tools clean. The clean brush and dust bin certainly imply the former, but the latter is also quite possible as indicated by the presence of original protective plastic over the control panel. The previous owner was one of those people who did not remove plastic protective film that were meant to be removed, a mindset very different from mine.

Me? I peel that stuff off before I start taking this vacuum apart to examine its internals.

New Project: Neato Hacking

My ROS learning robot Phoebe was built mainly around a laser distance scanner module salvaged from a Neato robot vacuum cleaner. At the time I knew people were selling them for $50-$75 on Craigslist in varying condition, but I was content to pay $45 for just the laser scanner. It was all I needed for my own robot exploration purposes. I thought a full Neato vacuum might be fun to play with, but I have enough projects on my to-do list that I didn’t feel the need to go out and find one.

Unless when I do. I recently wrote a Hackaday article about bargain shopping in a thrift store, and I needed a picture to go with my article. I went into my local thrift store to take some pictures, but it was also an opportunity to practice what I preached. I spent most of my time in the electronics section and didn’t find anything I wanted to take home with me. On my way out the door, though, I took a glance at the kitchen electronics section and spotted this beauty: a Neato robot vacuum with a price tag of only $7.99.

Savers Neato XV-21 1600

It doesn’t power on, and external accessories were nowhere to be found: neither a wall wart charger nor its charging dock. But it looked to be in pretty good condition with only minor cosmetic blemishes on the exterior. Aside from the missing charger, all other major components appear to be present. But the purchase decision was based on the most interesting part: I looked inside the top bump to verify presence of a familiar looking laser scanner unit. If I all I get out of this $7.99 is a Neato lidar, I’ll be happy. If anything else worked, they would just be icing on the cake.

It’s a big question mark, but it’s one I’m buying to take home for a closer look.

Robot Disorientation Traced To Timing Mismatch

Once the Roboclaw ROS Node‘s wheel parameters were updated to match the new faster motors, Phoebe went on a mapping run. And the results were terrible! This is not a complete surprise. Based on previous experimentation with this LIDAR module, I’ve seen its output data distorted by motion. It doesn’t take a lot of motion – even a normal human walking pace is enough to cause visible distortion. So now that Phoebe’s motors are ten times faster than before, that extra speed also adds extra distortion.

While this is a problem, it might not be the only problem behind the poor map. I decided to dig into an anomaly I noticed while using RViz to calibrate wheel data against LIDAR data: there’s some movement that can’t be entirely explained by a LIDAR spinning at 240 RPM.

The idea behind this odometry vs. LIDAR plot on RViz is to see if wheel odometry data agrees with LIDAR data. My favorite calibration surface is a door – it is a nice flat surface for LIDAR signal return, and I could swing the door at various angles for testing. In theory, when everything lines up, movement in the calculated odometry would match LIDAR observed behavior, and everything that is static in the real world (like the door) would be static in the plot as well.

In order to tune the base_width parameter, I looked at the position of the door before turning, and position after turning. Adjusting base_width until they line up, indicating odometry matches LIDAR. But during the turn, the door moved relative to the robot before finishing at the expected position.

When Phoebe started turning (represented by red arrow) the door jumped out of position. After Phoebe stopped turning, the door snapped back to position. This means non-moving objects appear to the robot as moving objects, which would confuse any mapping algorithm.

Odom LIDAR Mismatch No Hack

I chased down a few dead ends before realizing the problem is a timing issue: the timestamp on the LIDAR data messages don’t line up with the timestamp on odometry messages. At the moment I don’t have enough data to tell who is at fault, the LIDAR node or the Roboclaw node, just that there is a time difference. Experimentation indicated the timing error is roughly on the order of 100 milliseconds.

Continuing the experiment, I hard-coded a 100ms timestamp delta. This is only a hack which does not address the underlying problem. With this modification, the door still moves around but at least it doesn’t jump out of place as much.

Odom LIDAR Mismatch 100ms hack

This timing error went unnoticed when Phoebe was crawling extremely slowly. But at Phoebe’s higher speed this timing error could not longer be ignored. Ideally all of the objects on the RViz plot would stay still, but we clearly see nonuniform distortion during motion. There may be room for further improvement, but I don’t expect I’ll ever get to ideal performance with such an inexpensive LIDAR unit. Phoebe may end up just having to go slowly while mapping.

(Cross-posted to Hackaday.io)

Phoebe vs. Office Chair Round 2

Phoebe was built to roam my house, but the first draft chassis was unable to do so effectively due to a few problems that the second chassis aimed to solve. The first one was ground clearance, which was solved by raising the main chassis and sloping the bottom of the electronics tray. Sloping that leading edge gives Phoebe a better approach angle for smoothly transitioning between floor surfaces.

The second major problem was the LIDAR scanner’s height: it was too high to see the legs of an office chair. Hence the other major goal of the second chassis was to lower the LIDAR mounting point and hopefully bring an office chair’s legs into plane of view.

Placing the newly rebuilt Phoebe next to the chair looks promising at first glance. Unlike the taller first chassis, the LIDAR’s horizontal plane of sight is now low enough it should be able to see the legs.

Phoebe vs Chair Round 2

The proof is in the occupancy grid, and the RViz plot shows that Phoebe can now see the legs of the chair blocking its way.

Phoebe Sees Chair Legs

It’s not a very solid detection, though. Something about the surface texture and/or angle of the plastic results in a weak laser return.  And there’s the risk of a leg going undetected when if approached from the end, as the dark sloped rounded end of the chair leg is nearly invisible to LIDAR.

But it’s a huge improvement from before, where the LIDAR was too high to see any part of the starfish pattern. It’s good enough for us to proceed with the next task: integrate Phoebe’s new faster wheel drive motors into the system.

(Cross-posted to Hackaday.io)

Observations From A Neato LIDAR On The Move

Now that the laser distance scanner has been built into a little standalone unit, it’s easy to take it to different situations and learn how it reacts by watching RViz plot its output data. First I just picked it up and walked around the house with it, which led to the following observations:

  • The sensor dome sweeps in a full circle roughly four times per second. (240 RPM) This sounded pretty good at first, but once I started moving the sensor it doesn’t look nearly as good. Laser distance plot is distorted because it’s moving while it’s sweeping, visibly so even at normal human walking speeds. Clearly a robot using this unit will have to post-process distance data generated by this sensor to compensate for speed. Either that, or just move really slowly like the Neato XV-11 robot vacuum this LIDAR was salvaged from.
  • The distance data is generated from a single narrowly focused beam. This generates detailed sweep data at roughly one reading per vertical degree of separation. However, it also means we’re reading just a very narrow one degree horizontal slice of the environment. It’s no surprise this is limiting, but just how limited wasn’t apparent until we started trying to correlate various distance readings with things we can see with our eyes.

Autonomous vehicles use laser scanners that spin far faster than this one, and they use arrays of lasers to scan multiple angles instead of just a single horizontal beam. First hand experimentation with this inexpensive unit really hammered home why those expensive sensors are necessary.

Neato LIDAR on SGVHAK Rover

After the few handheld tests, the portable test unit was placed on top of SGVHAK Rover and driven around a SGVHAK workshop. There’s no integration at all…. not power, not structure, and certainly not data. This was just a quick smoke test that was very productive because it lead to more observations:

  • Normal household wall paint, especially matte or eggshell, works best. This is not a surprise given that it was designed to work on a home vacuum robot.
  • Thin structural pieces of shelving units are difficult to pick up.
  • Shiny surfaces like glass become invisible – presumably the emitted beam is reflected elsewhere and not back into the detector. Surprisingly, a laptop screen with anti-reflective matte finish behaved identically to shiny glass.
  • There’s a minimum distance of roughly 15-20cm. Any closer and laser beam emitted is reflected too early for detector to pick up.
  • Maximum range is over 4-5 meters (with caveat below). More than far enough for a vacuum robot’s needs.

The final observation was unexpected but obvious in hindsight: The detection capability is affected by the strongest returns. When we put a shiny antistatic bag in view of the sensor, there was a huge distortion in data output. The bag reflected laser back to the scanner so brightly that the control electronics reduced receiver sensitivity, similar to how our pupils contract in bright daylight. When this happens, the sensor could no longer see less reflective surfaces even if they were relatively close.

That was fun and very interesting set of experiments! But now it’s time to stick my head back into my ROS education so I can make use of this laser distance sensor.

Making My Neato LIDAR Mobile Again

The laser distance sensor I bought off eBay successfully managed to send data to my desktop computer, and the data looks vaguely reasonable. However, I’m not interested in a static scanner – I’m interested in using this on a robot that moves. Since I don’t have the rest of the robot vacuum, what’s the quickest way I can hack up something to see how this LIDAR unit from a Neato XV-11 works in motion?

Obviously something on the move needs to run off battery, and there’s already a motor voltage regulator working to keep motor speed correct. So that part’s easy, and attention turns to the data connection. I needed something that can talk to a serial device and send that data wirelessly to my computer. There are many ways to do this in the ROS ecosystem, but in the interest of time I thought I’d just do it in the way I already know how. A Raspberry Pi is a ROS-capable battery-powered computer and everything I just did on my computer would work on a Pi. (The one in the picture here has the Adafruit servo control PWM HAT on board, though the HAT is unused in this test.)

Mobile Scanning Module

The Raspberry Pi is powered by its own battery voltage regulator I created for Sawppy, supplying 5 volts and running in parallel with an identical unit tuned for 3 volts supplying power to spin the motor. As always, the tedious part is getting a Pi on the wireless network. But once I could SSH into the Pi wirelessly, I could run all the ROS commands I used on my desktop to turn this into a mobile distance data station. Reading in data via FTDI serial port adapter, sends data out as ROS topic /scan over WiFi.

Using a Raspberry Pi 3 in this capacity is complete overkill – the Pi 3 can easily shuttle 115200 bps serial data over the network. But it was quick to get up and running. Also – the FTDI is technically unnecessary because a Pi has 3.3V serial capability on board that we could use. It’s not worth the time to fuss with right now but something to keep in mind for later.

Now that the laser is mobile, it’s time to explore its behavior on the move…

Shouldn’t Simple LIDAR Be Cheaper By Now?

While waiting on my 3D printer to print a simple base for my laser distance scanner salvaged from a Neato robot vacuum, I went online to read more about this contraption. The more I read about it, the more I’m puzzled by its price. Shouldn’t these simple geometry-based distance scanners be a lot cheaper by now?

The journey started with this Engadget review from 2010 when Neato’s XV-11 was first introduced to fanfare that I apparently missed at the time. The laser scanner was a critical product differentiation for Neato, separating them from market leader iRobot’s Roomba vacuums. It was an advantage that was easy to explain and easy for users to see in action on their product, both of which help to justify their price premium.

Of course the rest of its market responded and now high-end robot vacuums all have mapping capability of some sort or another, pushing Neato to introduce other features like internet connectivity and remote control via a phone app. In 2016 Ars Technica reviewed these new features and found them immature. But more interesting to my technical brain is that Ars linked to a paper on Neato’s laser scanner design. Presented at May 19-23 2008 IEEE International Conference on Robotics and Automation titled A Low-Cost Laser Distance Sensor and listing multiple people from Neato Robotics as authors, it gave an insight into these spinning domes. Including this picture of internals.

Revo LDS

But even more interesting than the fascinating technology outlined in the paper, is the suggested economics advantage. The big claim is right in the abstract:

The build cost of this device, using COTS electronics and custom mechanical tooling, is under $30.

Considering that Neato robot vacuums have been in mass production for almost ten years, and that there’s been ample time for clones and imitators to come on market, it’s quite odd how these devices still cost significantly more than $30. If the claim in the paper is true, we should have these types of sensor for a few bucks by now, not $180 for an entry-level unit. If they were actually $20-$30, it would make ROS far more accessible. So what happened on the path to cheap laser scanner for everyone?

It’s also interesting that some other robot vacuum makers – like iRobot themselves – have implemented mapping via other means. Or at least, there’s no obvious dome of a laser scanner on top of some mapping-capable Neato competitors. What are they using, and are similar techniques available as ROS components? I hope to come across some answers in the near future.