MEMSIC MMC5983MA 3 Axis Digital Compass & Arduino MCU – The Basics

Summary

In this video, the presenter reviews the MEMSIC MMC5983MA 3-axis magnetic sensor, a high-precision digital compass chip. They compare it with the QST QMC5883L, noting the MMC5983MA's superior accuracy of ±0.5 degrees and lower noise levels. The sensor is more expensive but offers impressive performance right out of the box. The presenter explores the sensor's documentation, evaluates its features, and demonstrates how to interface it with an Arduino using the SparkFun library. They also discuss the need for calibration to achieve accurate compass readings.

Takeaways

• 😀 The presenter discussed a 3-axis magnetic sensor, the QST QMC5883L, which is a digital compass chip costing only five dollars.
• 🔍 The QMC5883L promised 1-2 degree compass heading accuracy, but with a significant reaction time of over 12 seconds.
• 🆕 The presenter introduced a new sensor, the MEMSIC MMC5983MA, which is more expensive but offers higher accuracy of ±0.5 degrees.
• 📈 The MMC5983MA has very low noise levels and is specified in the data sheet, making it superior to the QMC5883L in terms of precision.
• 🔌 The evaluation board for the MMC5983MA has eight pins, including an interrupt out, I2C interface pins (SCL and SDA), and multi-purpose pins for SPI operation.
• 📚 The data sheet for the MMC5983MA highlights an 8G full scale range, 18-bit operation, and 0.4 milligaus total RMS noise.
• 💾 The sensor can operate with an output data rate up to 1000 Hz and has a built-in degaussing function to eliminate residual magnetic fields.
• 🛠️ The sensor is compatible with 3.3 volts, which is suitable for use with an Arduino, and has low power consumption.
• 📱 An Arduino library is available for the MMC5983MA, including examples for basic measurements and continuous measurements.
• 🔧 The presenter plans to create their own library for the sensor to include important functions like calibration, as the sensor requires calibration for accurate compass readings.

Q & A

What is the QST QMC5883L?

The QST QMC5883L is a three-axis magnetic sensor, essentially a digital compass chip, which is known for its affordability and the promise of one to two-degree compass heading accuracy.

What was the issue with the QST QMC5883L in the fifth detail video?

In the fifth detail video, the QST QMC5883L was found to have a reaction time of over 12 seconds with a 250 sample median filter, making it unusable as a real digital compass due to its slow response time.

What is the Memsic MMC5983MA and how does it compare to the QMC5883L?

The Memsic MMC5983MA is a 3-axis magnetic sensor that is more expensive than the QMC5883L but promises a higher heading accuracy of plus or minus 0.5 degrees. It also has very low noise levels as specified in its data sheet.

What are the key features of the Memsic MMC5983MA according to its data sheet?

Key features of the Memsic MMC5983MA include a full-scale range of plus minus 8G, 18 bits of operation, 0.4 milligaus total RMS noise, an output data rate up to 1000 Hertz, and a built-in degaussing function.

How is the Memsic MMC5983MA connected to an Arduino?

The Memsic MMC5983MA can be connected to an Arduino using either an I2C or SPI interface. For I2C, the SDA and SCL lines are used along with a ground and power supply. For SPI, the slave data out, slave data in, slave clock, and chip select pins are used.

What is the maximum supply voltage for the Memsic MMC5983MA?

The maximum supply voltage for the Memsic MMC5983MA is 3.6 volts, but it can operate at 3.3 volts, which is compatible with an Arduino.

Is there an Arduino library available for the Memsic MMC5983MA?

Yes, there is an Arduino library available for the Memsic MMC5983MA, specifically the SparkFun MMC5983MA magnetometer Arduino Library.

What are the different operation modes supported by the Memsic MMC5983MA?

The Memsic MMC5983MA supports I2C single supply operation and SPI single supply operation. It does not support I2C or SPI dual supply operation.

What is the significance of the 0.4 milligaus RMS noise in the Memsic MMC5983MA?

The 0.4 milligaus RMS noise is the best value found in data sheets for Hall element-based magnetic compasses. It indicates a very low level of noise in the sensor's measurements, which is crucial for accurate compass readings.

What is the purpose of the degaussing function in the Memsic MMC5983MA?

The degaussing function in the Memsic MMC5983MA is used to eliminate any residual magnetic fields that might have built up when the device is powered on, ensuring accurate compass readings.

Outlines

00:00

🧭 Introduction to the MEMSIC MMC5983MA 3-Axis Magnetic Sensor

The script starts with the host welcoming viewers back and mentioning a previous series of videos about the QST QMC5883L magnetic sensor. They compare the older, cheaper sensor to the new MEMSIC MMC5983MA, which is more expensive but promises higher accuracy of ±0.5 degrees. The host unboxes the evaluation kit for the MMC5983MA and describes the board's pins, including I2C and SPI interfaces, and power supply connections. They also note the absence of a connection for VDDIO and the presence of capacitors and resistors on the board.

05:02

📄 Exploring the Datasheet and Breakout Board

The host delves into the datasheet for the MMC5983MA, highlighting its features such as an 8G full-scale range, 18-bit operation, low noise levels, and the ability to set heading accuracy to ±0.5 degrees. They discuss the output data rate, degaussing capabilities, and the I2C interface's fast mode. The script mentions the block diagram, showing the sensor's components and how data is stored in registers. The host also covers the supply voltage, current, and temperature ranges, as well as the noise levels achievable with maximum filtering.

10:02

🔌 Wiring and Testing the Sensor with Arduino

The script describes how to wire the MMC5983MA to an Arduino, covering the power connections, I2C pull-up resistors, and SPI configuration. The host confirms that the sensor operates at 3.3 volts and is compatible with Arduino. They mention finding an Arduino library for the sensor and briefly review its features, including error handling and support for both I2C and SPI interfaces. The script ends with the host sharing their initial positive experience with the sensor's performance and accuracy right out of the box.

15:05

💻 Reviewing the SparkFun Library and Testing Examples

The host reviews the SparkFun MMC5983MA magnetometer Arduino Library, discussing its features and functions. They mention the library's support for I2C and SPI, and note some private functions and the ability to set an error callback. The script describes the process of using the library with Arduino, including initialization, checking for connection, performing a soft reset, and obtaining sensor measurements. The host demonstrates the library's functionality by running an example that outputs compass headings with minimal noise and high accuracy.

20:07

🔍 Analyzing Raw Data and Considering Calibration

The script continues with the host analyzing raw sensor data and normalized values, noting the low levels of noise in the measurements. They compare the sensor's output to a physical compass to demonstrate its accuracy. The host reviews another example from the library that outputs raw values and magnetic field measurements in Gauss. They express their intention to write their own library for the sensor, focusing on I2C and adding calibration functions for improved accuracy.

25:10

🔚 Conclusion and Future Plans

In the final paragraph, the host summarizes the video's content, which includes an overview of the datasheet, wiring instructions for the Arduino, and a review of the SparkFun library. They express satisfaction with the sensor's performance and outline plans for future videos, which will involve writing a custom library and implementing calibration to improve the sensor's accuracy.

Keywords

💡QST QMC5883L

The QST QMC5883L is a three-axis magnetic sensor, essentially functioning as a digital compass chip. It is highlighted in the video for its affordability and capability to provide compass heading accuracy within one to two degrees. The video creator previously made a series of videos discussing this sensor, indicating its importance to the discussion.

💡MEMSIC MMC5983MA

The MEMSIC MMC5983MA is another 3-axis magnetic sensor introduced in the video as a more advanced and expensive alternative to the QST QMC5883L. It promises higher heading accuracy of plus or minus 0.5 degrees. The video focuses on this sensor's superior specifications and its potential applications.

💡Magnetic Sensor

A magnetic sensor is a device that detects and measures magnetic fields. In the video, the discussion revolves around two types of magnetic sensors, the QST QMC5883L and the MEMSIC MMC5983MA, which are used for compass applications. Their accuracy and performance are central to the video's theme.

💡Heading Accuracy

Heading accuracy refers to how precisely a compass device can determine the direction it is pointing. The video compares the heading accuracy of two magnetic sensors, with the MMC5983MA promising superior accuracy of plus or minus 0.5 degrees compared to the QMC5883L.

💡I2C Interface

The I2C (Inter-Integrated Circuit) interface is a multi-master, serial computer bus that the video mentions as the communication protocol used by the MMC5983MA sensor. The video explains how to wire this interface to an Arduino, emphasizing its importance for sensor operation.

💡SPI Interface

SPI (Serial Peripheral Interface) is another communication protocol mentioned in the video. Although the focus is on the I2C interface, the video also discusses the SPI as an alternative method for interfacing the sensor with a microcontroller.

💡Arduino

Arduino is an open-source electronics platform used for building digital devices. The video describes how to interface the MMC5983MA sensor with an Arduino, demonstrating the practical application of the sensor in a real-world project.

💡Data Sheet

A data sheet is a document provided by the manufacturer that details the specifications and capabilities of a component. The video refers to the data sheet of the MMC5983MA, discussing its specifications like noise levels, supply voltage, and output data rate.

💡Calibration

Calibration in the context of the video refers to the process of adjusting the sensor to ensure accurate readings. The video mentions that despite the MMC5983MA's low noise, it still requires calibration to obtain precise compass readings.

💡Noise Levels

Noise levels in sensors refer to the random variations or inaccuracies in the sensor's output. The video emphasizes the low noise levels of the MMC5983MA, which is a key feature for achieving high accuracy in compass heading measurements.

💡Library

In the context of the video, a library refers to a collection of code that facilitates the use of a sensor with a microcontroller like Arduino. The video discusses a specific library for the MMC5983MA and the process of using it to interface with the sensor.

Highlights

Introduction of the QST QMC5883L three-axis magnetic sensor as a digital compass chip.

The QST QMC5883L's affordability at five dollars for the whole breakout board.

Claimed one to two-degree compass heading accuracy for the QST QMC5883L.

Achieving one degree to one and a half degree accuracy with a 250 sample median filter on the QST QMC5883L.

Introduction of the MEMSIC MMC5983MA 3-axis magnetic sensor as a more advanced alternative.

Higher cost of the MEMSIC MMC5983MA chip and evaluation kit.

Promised heading accuracy of plus minus 0.5 degrees by the MEMSIC MMC5983MA.

Very low noise levels specified in the MEMSIC MMC5983MA data sheet.

Overview of the MEMSIC MMC5983MA evaluation board and its pins.

Documentation indicating that the breakout board pins should refer to the actual data sheet for more information.

Description of the multi-purpose pins on the breakout board for SPI and I2C interfaces.

Mention of unpopulated positions for resistors and capacitors on the breakout board.

Data sheet highlights including plus minus 8G full scale range and 18 bits operation.

Low noise levels of 0.4 milligaus total RMS noise as per the data sheet.

Output data rate up to 1000 Hertz for the MEMSIC MMC5983MA.

Built-in degaussing function to eliminate residual magnetic fields.

Block diagram explanation showing the three Hall elements for X, Y, and Z-axis.

Information on how to wire the MEMSIC MMC5983MA to an Arduino using 3.3 volts.

Availability of an Arduino Library for the MEMSIC MMC5983MA.

Review of the SparkFun MMC5983MA magnetometer Arduino Library.

Demonstration of the library's functionality with an Arduino and immediate accuracy results.

Comparison of the noise levels between the QST QMC5883L and the MEMSIC MMC5983MA.

Intention to create a custom library for the MEMSIC MMC5983MA focusing on I2C and adding calibration functions.

Transcripts

welcome back about one year ago I made a

total of six videos uh one basic and

five detailed videos about this qst qmc

5883l three axis magnetic sensor

basically a digital compass chip and

that thing was dirt cheap I spent five

bucks for it for the whole breakout bot

that is and it promised to enable you to

have one to two degree Compass heading

accuracy and at the very end in the

fifth detail video I was able to reach a

one degree to one and a half degree

accuracy

with a 250 sample medium filter

which had a reaction time of over 12

seconds so basically as a real digital

compass that thing was completely

unusable

uh car to the last video here and Link

in the description in the description

for the last video you find links to all

the videos about that thing

enter the stage of the memsic MMC

5983 m a 3-axis magnetic sensor that's

another Beast first of all it's more

expensive the chip alone will set you

back a little less than five bucks and

this evaluation kit I have here costs uh

almost 24 Euros however it promises you

a heading accuracy of plus minus 0.5

degrees and it comes actually with a

specified noise levels very low noise

levels in the data sheet

in this video we will have a look at

that set data sheet at the evaluation

board that should be in here and we will

try to get the whole thing up and

running with an Arduino

enjoy

I really should have made a unboxing

video for that thing so let's open the

box and oh

that looks very nice and

there's the board and it comes in

of course a little ESD bag which is open

which is open uh don't stop the video

right now

and there we have

the evaluation board

we have eight pins in total on this side

we have an uh interrupt out I assume pin

scl and SDA that's your uh I squared C

interface and then we have an sdo uh

these pins are probably multi-purpose

either as the ascl for i squared C or a

slave data out slave data in enslave

clock for SPI operation and then we have

the chip select pin also for SPI

operation and then we have an vddio so I

O Supply voltage and vdd General Supply

voltage and ground pin but let's have a

look on the documentation for that thing

here's the documentation for the

breakout board they tell us right away

that we should refer to the actual data

sheet for more information

and they give us a table with the pins

we already talked about that there are

these two multi-purpose pins so I have a

green for SPI that would be a SPI slave

clock and slave data in or in blue I

squared C slave data and slave clock and

then we have two more SPI pins for slave

data out and the chip select and we have

an input output interrupt output and we

have the two power supply pins vdd on

ground what's marked on the little

breakout bot here as vdd I O is not

connected and I just verified that

visually there's really no Trace going

off from that vdd I O pin

there's even a little schematic of that

breakout board here and we see here that

pin 2 which is marked vddio not being

connected but instead the vdd io and the

vdd pin of the chip being connected to

the same output pin here at the board we

also saw that these two C1 and C2

capacitors are populated at that board

and their wall there were unpopulated

positions for that are one and R2

resistor here which are obviously the I

squared C pull ups for the SDA and scl

line

however there was a populated Arrow 3 on

there with the value 0 1 0 and an

unpopulated of four and R5 position two

so uh I guess that 010 is just a bridge

and uh yeah they're using that breakout

board for different chips

again on the board there's the

population C1 and C2 capacitor there's

that uh

010r3 resistor the unpopulated ll4 and

R5 which are not even mentioned in the

circuit diagram and the unpopulated

positions for an R1 and R2 2.7 K pull up

resistors for the I squared C lines

let's have a look into the data sheet

then

the highlights plus minus 8G full scale

range that's the usual for a compass

chip 18 bits operation not only 16 but

18 bits 0.4 milligaus total RMS noise

and that's the best value in a data

sheet I found for these Hall element

based magnetic compasses you can get

ships that are better but these are no

longer based on Hull elements but they

are really using three different coils

to measure the magnetic field of the

earth and they also in another price

range

enable setting accuracy of plus minus

0.5 degrees we talked up already about

that uh output data rate up to 1000

Hertz now I don't need that but that's a

lot and degaussing with built-in set

reset function so that thing can

eliminate when you boot it up any

residual magnetic fields that might have

built up a lot of blah blah blah yeah

output data ready interrupt output of

course uh then the I squared C interface

is fast mode 400 kilohertz that's good

3.0 volt single low power supply ER I

hope it will run with 3.3 volts too we

will see that in the details and of

course it has an alternative SPI

interface available

here's the block diagram very tiny so

you have here your three Hull elements

for the X Y and z-axis and then you have

some signal processing and then the

results are stored in three registers X

Y and Z which you can read while I

squared C and SPI I'm sure there are

also some registers that you can write

to set different options uh clock

generator blah blah blah set reset

control that was that degaussing staff

voltage reference test and buffer

switches temperature sensor I think I

skipped that it has a built-in

temperature sensor for temperature

compensation too

maximum Supply voltage is 3.6 volts so I

want to run that thing on 3.3 volts on

an Arduino so that's fine uh

Supply current okay uh yeah not even

half a milliamp and yeah the uh there

are these BW zero zero two one one these

are bits in some of the control

registers and they control how long a

measurement can take or will take

meaning how much filtering the chip does

internally that's important that ranges

from eight milliseconds to 0.5

milliseconds that's important because if

we go further down here our operating

temperature range minus 40 to 105

degrees with golden there linearity

Arrow 0.1 percent of full scale that's

all very very nice repeatability error

0.1 of false scale alignment error and

so on and so on until we come here to

the RMS noise

and

the promised

0.4 milligaus RMS noise you only get if

you take the filtering to the Max and

then each measurement takes eight

milliseconds and uh yeah you don't get a

1000 measurements per second out of that

thing but that's okay for my application

uh new value every eight milliseconds is

more than I need

there's a whole lot more information

here in the data sheet but uh yeah we

will come back to that when we actually

need it

absolute maximum ratings so Supply

voltage from minus 0.42 plus 3.6 volts

again we're golden here with 3.3 volts

from our Arduino

on page 10 respectively 11 of the data

sheet we find for typical application

circuits for or four different modes

operation modes I squared C dual Supply

operation we can forget about that

that's not supported by our little

breakout board SPI dual Supply operation

is also not supported so we can

concentrate on I squat C single Supply

connection and SPI single Supply

connection

I squat C single Supply connection no

surprises here but we have to bind the

SPI chip select pin to vdd to indicate

that we are actually want to use I

squared C and the SPI slave data out

chip is just left dangling here

please note that the pull up resistors

here for our I squared csdi and as clock

lines are not on the breakout board at

least they are not populated for SPI

single Supply operation you just feed

your SPI signals in the four pins that

are for SPI so slave data out slave data

in slave clock and Chip select

now that we know that the thing will run

with 3.3 volts and we have an idea how

to connect it to our Arduino the

question is is there an Arduino Library

available and yes there is but only one

at least I only found one library and

that's the spark fund SparkFun MMC

5983ma magnetometer Arduino Library

what else

um

we will probably go into the details a

little bit but I want to point out here

this little remark added last year added

SPI support Dash partial okay not that

this matters to us because I want to use

I squared C but this Library does

somehow also support SPI but maybe not

completely anyway let's have a look at

the header file if we like that Library

sometimes uh you don't like it or

you know me so yeah here

all the includes and we need the

SparkFun IO library on Arduino Library

constants and different files yeah or

yeah I can live with that lots of

private functions here nothing too bad

than a default Constructor and the

default Destructor okay or you can set

an error callback function that's not

too bad I like that uh error code string

also not too bad uh begin that is there

are two different begins here one

obviously for I squared C with a two

wire object

and then here one with uh for SPI

and then you get the usual is connected

get temperature yeah we know this thing

has a temperature sensor software Ascent

enable interrupt disable and drop that

obviously for the intro pin is interrupt

enable yeah query

uh what else we have here our enable

three wire SPI and disable free via SPI

we don't care about that is three wires

yeah no SPI uh perform set operation

perform reset operation okay enable

automatic set reset okay whatever that

is we will dive into those details in

the first the details video I guess

is automatic said okay enable X Channel

disable

okay Channel disable

enable for X Y set set build a filter

bandwidth we talked about that okay

anyway uh

yeah uh complete Library

using I don't know if it's really using

all of the functions but uh

it doesn't look that bad so let's use it

the library is available within the

Arduino Library manager in the version

one one one that was also the one we

just saw on GitHub great

and that Library comes with a bunch of

examples seven to be exact starting from

a simple measurement digital compass or

we will try that a continuous

measurement simple measurement digital

compass two

fast continuous measurement and sensor

offset

let's start with a digital compass and

um

yeah

load that to our Arduino and connect

everything up

everything is wired up now nicely so we

have our ground connection to ground and

we have a 3.3 volt going to our vdd VD

Dio has no connection we already talked

about that and the SPI chip select is

bridged to our vdd so that should put

the whole thing into isolate C mode

we have a two 2.7 K as was mentioned in

the datasheet pull up resistors here on

our chip clock uh slave clock and slave

data lines and the slave data line is

going over here to the slave data pin

and the slave clock line is going over

to the slave clock pin let's fire that

thing up

and right out of the bed it's working

and I want you to note how little noise

there is on the signal here so yeah uh

uh it's fluctuating between five

three

I see a six in between can I get a six

nah not really so uh it's fluctuating

about two tenths of a degree

right off the Box no additional

filtering just with that

SparkFun Library no calibration at all

and that's the next big surprise uh I

have placed here a little Compass so you

can see

where North is

and

yeah okay we're 10 degrees off but

without calibration that's absolutely

okay

and if we go to 90 degrees approximately

yeah

seven eight degrees of

270 degrees I mean we will make a

Precision measurements in the details

videos of course yeah four degrees off

and

if we go to 180 degrees approximately

yeah we are at 180 degrees without

without again if you followed the series

about the qmc chip

without any calibration

that's a really really impressive uh

right out of the gate

I like it

let's go over that spark fun example

real quick so we include here wire and

the SparkFun library then we instantiate

that's the class name okay uh our chip

the object for our chip we make a Serial

begin and we make a little serial print

that was already scrolled out of the

screen and if Mac begin is false there's

something wrong then you get a error

message and it goes into an endless loop

here

and then we do a soft reset whatever

that does we will see in the details

videos there will be several of them and

then we say we are connected and then we

have here somehow 32-bit unsigned

integers for the raw values of the X Y

and set sensor then we have doubles for

the normalized values uh please note a

double is only a float in the Arduino

environment so it doesn't really matter

it's only 32-bit whatever you want to do

and then we do I mean that

self-explanatory uh get measurement X Y

set into the raw values and then we

normalize these values into a range

between plus minus

1.0

and we do that for all three values for

x y and set and then we calculate the

heading using the archus tankans two

function and this doesn't deliver uh

yeah a result in degree plus but uh

something between plus pi to minus pi

and we convert that in degrees and then

this is printed out so really straight

forward uh by the way the whole Arcus

tungans two thing to actually calculate

a compass heading from two sensors is

explained in the room in the third

details video about the qmc chip card

here Link in the description let's try

another example then shall we

so the examples are not that interesting

uh this basically gives us the raw

values and the values in Gauss on the

serial military uh let me stop the auto

scroll so the flickering goes back uh

but again I want you to note how little

how little the raw values fluctuate

really this is a a few least significant

bits noise we have on here so that's

really nice and then we have our Gauss

values and I activate the auto scroll

again and of course if we move our

Compass I'm doing that off screen here

uh

forgot to switch on the other camera

then things are changing so yeah really

nice

and the code for that example is also

very trivial uh yeah it starts exactly

the same yeah initialize your serial

interface check if the chip is there

and uh yeah we already scrolled over

that uh get the temperature and

Fahrenheit or Celsius uh then we go into

a loop we have again our six values here

for x y and set count value that is an

unsigned integral 32 and normalized we

get our measurements this time with

three methods each measurement for its

own and then you saw that it's printed

out the intellectual value

and then we convert these 18-bit

unsigned intentional values into

a double which should be

after it's multiplied by eight You

Remember full scale is eight cows for

that chip gives us the field value the

magnetic field value or at least the

component in X Y and Z Direction

in cows

that's it that's it yeah the examples

are not very

extensive uh there's also nothing

absolutely nothing uh with calibration

and stuff so we will have to do all that

by ourselves

that's it for today we had a little

squeeze on the data sheet of that thing

and it's uh about at least according to

the data sheet impressive performance uh

we learned how to wire that thing up to

an Arduino the different modes it can it

supports us with c and SPI and we found

a readily available Library the examples

are a little bit mood but yeah it's

working

very well

next time

the details one uh we will start writing

our own library for that thing because

never trust a library uh from the

internet about we could look into the

code how the library is implemented but

anyway uh I think it's a little bit

heavy with that uh SPI and I squared C

support and then including some uh i o

library from SparkFun I don't like that

we will write our own Library just

purely for I spread C

and then we will add some important

functions over the different details

videos like for example calibration

because while that thing is basically

compared to the qmc

chip uh noiseless

it does need calibration okay otherwise

you won't get a very good values a very

good Compass readings out of that

till then bye

bye