Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Kyle Dando (00:10):
Hey everyone.
Welcome back to the EdgeVerse TechCast,your go-to source for all things NXP
software, tools, and edge innovation.
I'm Kyle Dando, and Iam your taller co-host.
Bridgette Stone (00:26):
And I'm Bridgette
Stone, your smarter co-host?
Kyle Dando (00:30):
Possibly.
Bridgette Stone (00:33):
Today we're here
exploring a solution that brings some
serious Analog Awesomeness to the edgeNXP's new Universal Sensing Module.
Kyle Dando (00:41):
So, Bridgette, I know we
talked about this in a prior episode.
You have had a front row seat onthis being developed and you were
part of the project team, correct?
Bridgette Stone (00:51):
I was, and it's been
exciting to see how this solution
combines our Analog and Microcontrollerexpertise in a way that really
simplifies things for Designers.
I think our listeners, whether they'refrom the MCU or Analog world, or going
to walk away with something new today,
Kyle Dando (01:08):
Well, Bridgette, it sounds
like you were paying attention and
you learned quite a bit in thoseplanning meetings, but now for our
listeners, I think it's gonna bereally great to dig into the tech
behind this solution, so you invitedMadhura Tapse, one of the experts
behind the Universal Sensing Modules.
Madhura, thanks for joining us.
Madhura Tapse (01:29):
Absolutely.
Thanks for having me here.
Hey everyone.
I'm Madhura Tapse.
I'm product Marketing manager here at NXP.
I'm based in San Jose, California.
Today I'm excited to take you throughour Universal Sensing Module solution.
We will explore how this flexibleand scalable approach helps simplify
(01:51):
sensor integration and acceleratestime to market for edge intelligence.
Bridgette Stone (01:58):
Thanks
for being here, Madhura.
It's good to talk to you.
So to, set the scene, we've all seenhow much smarter our environments
are becoming from factories to homes.
How exactly does analog sensing tieinto these transformations, especially
with AI and ML pushing the boundaries?
Madhura Tapse (02:18):
This is a great
question to kickstart the conversation.
Bridgette, if you look around us, weare living in an intelligent world.
All the places around us are becomingintelligent, such as Smart Buildings,
Smart Homes, and Smart Factories.
To make these places intelligent, youneed safe, secure, and predictive data.
Bridgette Stone (02:40):
That's such a good point.
Everything is getting smarter, but itstill all starts with the raw data.
Correct?
Madhura Tapse (02:47):
Yes.
Even the smartest AI needs theclean and precise raw data fuel.
No shortcuts there!
And most of the data in ourworld is inherently analog.
What customers require is a solutionthat can convert this analog data
to digital data, and then it canbe processed by these controllers
(03:09):
to make quality decisions and runall these environments efficiently.
However, one thing I have consistentlynoticed is that customers don't prefer
only one type of sensor connection.
Let's say only voltage.
They don't want only voltagetype of sensor connection.
They require something which ishighly versatile, that can measure
(03:32):
not only voltage, but also current,temperature, resistance, pressure,
weight, etc. So having thatmultisensor capability has become a
priority for most of our customers.
When we talk about these measurements,another thing I have noticed is that
they require something which is highlyaccurate and highly reliable, and
(03:55):
this kind of ties into our Industry4.0 and automation trend as well.
Additionally, we are in the middle ofthis AI and ML evolution at the edge.
So customers prefer real timedecision making, which requires
precise high speed data acquisition.
So if you look at all the SmartBuildings, Smart Homes, and Smart
(04:17):
Factories holistically, there are twomain things which customers prefer.
One is definitelyconnectivity and data sharing.
They want robust connectionsfor reliable industrial systems.
Second, they want high performance andlower power with a small footprint so that
they can reduce their carbon footprint.
(04:38):
Now these are the trends that we areseeing in the world of analog sensing.
Kyle Dando (04:43):
Okay.
But now you did mention industrialenvironments, and I know those are
pretty crazy and hard to control.
There's tons of noise, highvoltage, lots of interference with
the data going back and forth.
So how do you ensure for our customersthat these sensing solutions that
the team put together are stillreliable under those conditions?
Madhura Tapse (05:04):
Now that's
a great observation, Kyle.
If you look at any industrialsystem, there are two or three main
pieces of the puzzle that everysystem engineer needs to solve.
First is definitely analog sensing.
Second is processing, andthird is connectivity.
Respect to analog sensing, the analogengineer needs to ensure that the
(05:26):
sensing module meets the accuracyand precision requirements even
in an extremely noisy environment.
At the same time, they need to avoiddamage to the module because there
are many high voltage transience,like ESD, EFT surges and high currents
due to miswiring or short circuits,which can easily damage the hardware.
Kyle Dando (05:49):
All right.
You mentioned miswiring, but Ipromise you, nobody would ever put
me in charge of wiring up some ofthose complex industrial connections.
But like you said, it's not just aboutthe clean data, it's about surviving.
There's just so much they have to dealwith when they're building a product
that's going into these factories.
Madhura Tapse (06:06):
Yes, absolutely.
So for every analog systems engineer,it is essential to ensure that
signal integrity and noise rejectionrequirements are met while also making
the module highly robust with somewhatinherent diagnostic capabilities.
Now, if you move to processing andconnectivity side, you will hear from
(06:29):
many AI/ML engineers or even industrialsystem software developers that it is
very hard to meet throughput requirementsand ultra-low latency requirement,
Kyle Dando (06:41):
And no pun intended,
but I think that's where the
pressure really starts to build.
Trying to do all that you justmentioned in real time with, I
think you said minimal power draw.
It's just getting socomplicated for the customers.
Madhura Tapse (06:56):
Yes, that is correct.
And I really empathizewith these engineers.
So if the engineers want to performreal time decision making, it
becomes challenging to do so at lowerpower or with better connectivity.
This has become difficult for themwith the available technology.
Building such an inherently robust,highly precise, and fast solution has
(07:18):
become a key challenge for many customersdecelerating their time to market.
This is crucial when themarket is evolving so quickly.
You definitely want to developsomething that is highly valuable,
but in a shorter time span.
Bridgette Stone (07:34):
Thanks Madhura.
So let's get into thenuts and bolts of it.
Tell us about the technical magicbehind the universal sensing module.
What's the real power behind the solution?
Madhura Tapse (07:46):
To answer your question
in simple words, Bridgette, what we
are trying to do with this solution isto reduce customers cost of ownership.
So on the hardware side, if you lookat the analog side, typically we have
NXP's Analog Front End family, and wehave used the superset product here.
NAFE13388 is a highly integrated solution,and you can achieve 24-bit precision
(08:13):
with accuracies of 0.01% in very noisyenvironments across high temperatures.
So this makes it highly suitable forall industrial system applications where
software configurability is needed.
And at the same time, you requirehigher performance and robustness.
(08:33):
As we discussed earlier,everything can be provided by NAFE.
This product is equipped with ESDand search protection, as well as
temperature and voltage conditionmonitoring with anomaly detection
circuitry for system and self diagnostics.
So from an analog point of view, thissolution provides a lot of leverage that
(08:55):
you can quickly reuse in your design.
Bridgette Stone (08:58):
Wow.
So not just accurate,but very tough as well.
Very solid combo.
Madhura Tapse (09:04):
Yes, call it
Chuck Norris of signal chains.
Accurate and indestructible.
With respect to processingand connectivity.
Now we have the MCXN947 dual ArmCortex-M33 core, which runs at 150
megahertz and has 2 megabytes of flash.
It also features the eIQ NeutronNPU for AI acceleration, and it
(09:29):
also has DSP Accelerator inbuiltalong with Edge Lock Secure enclave.
Now this edge processing will giveyou leverage in achieving lower
power consumption, and at thesame time ultra-low latency for
your real time decision making.
And now we just don'tstop here on the hardware.
(09:49):
We actually have developed a softwarestack so that this analog sensing data can
be transferred over any of the interfaces,like CAN FD, 10 /100 Mbps Ethernet or USB.
The entire software stackis available on ACH.
Having this complete hardware stackas a hardware solution as well as
(10:11):
a software stack with SDK, driver,middleware, and a web server GUI will
actually accelerate your time to market.
Kyle Dando (10:19):
And that's all available
now on NXP's Application Code Hub.
What a great commercial!
You're telling me that anydeveloper listening right now
can basically jump in, go to theApplication Code Hub and get rolling?
Madhura Tapse (10:33):
Yes, everything
is available on ACH.
So developer can just go toACH and quickly run the example
code with his or her own sensor.
Since all the source code is available,further customization is also
possible based on system requirements.
This ease of use helps toaccelerate the time to market.
Kyle Dando (10:56):
That's great.
We'll put the link Madhura in theepisode notes to get people quickly
to that Application Code Hub example.
So that they can get started evaluating.
But before they do that, maybe we can talkabout some other opportunities that they
might consider using this product for.
You said it's versatile.
What are some other unexpectedindustries or applications where
(11:18):
everything you've outlined couldbe a game changer for someone else?
Madhura Tapse (11:22):
You are absolutely
right to note that this is a
highly versatile solution, Kyle.
So our use cases actuallyspan multiple areas.
The first and foremost is industrial.
Smart Factory is a very goodtarget for this solution.
The analog front end has a lot ofparameters and features suitable
(11:43):
for applications like ProgrammableLogic Controller or Analog IO Module.
It is also suitable fortemperature measurement in
process control application.
Now, not just limited to factoryautomation, but also the other areas
of industry such as weighing scale ormotor control, specifically torque and
(12:04):
force measurement, and even AutomatedTest Equipment where parametric
measurement unit capabilities areneeded in wire bonding instrumentation.
This solution could be a great fit.
Additionally, the solution can be appliedto food processing equipment or grain
silos where monitoring, temperature,humidity, and weight is essential.
(12:26):
So in short, this is a great solution.
In case of even lift gate, whereheavy motors require some form
of predictive maintenance, thissolution is excellent due to its
high robustness and performance.
Kyle Dando (12:39):
All right.
I love it.
It does sound like this thing can doalmost anything we'd want it to do.
From factory floors to foodsilos, it's like the Swiss Army
knife of the sensing group.
Madhura Tapse (12:51):
True.
And just like a Swiss knife,Kyle, it is always ready to
deploy with minimal BOM cost.
Now, moving on fromindustrial applications.
In home and building settings, you wouldwant to measure multiple parameters.
Not just temperature, but alsopressure, flow control, and humidity.
Now this solution makes it idealfor building automation controller
(13:15):
applications or thermostat.
Since this solution has the inherentcapability of higher performance
with respect to accuracy andprecision, it is again a great
target for any lab instrumentationsuch as mass spectrometers, LCR
meters, or water quality analyzers.
So the key aspect I have observedis that you can find immediate
(13:38):
use cases for this solutionwherever software configurability,
robustness, accuracy, precision,or higher performance is required.
Whether on the analog sideor on the processing side.
Where ultra-low latency andlower consumption are needed,
the solution is a great fit.
Bridgette Stone (13:59):
I love it.
So wherever precision andconfigurability are crucial or
critical, this solution has a place.
Now I wanna just transition us alittle bit and look at the landscape.
So when you look at thecompetition, what sets N XP apart?
What does the Universal SensingModule bring to the table
(14:19):
that others just can't match?
Madhura Tapse (14:22):
Thanks for asking this
very important question, Bridgette.
In terms of competition, NXP isthe only solution that provides AFE
data over CAN, Ethernet, or USB.
Competitor AFE solutionsconnect through only USB.
Our solution is tested and EMCcompliant so it can readily be
(14:43):
used to achieve higher performanceand robustness in customer system.
Since we provide the entire softwarestack to transfer data over CAN and
Ethernet, the solution can accelerateevaluation and deployment time.
Bridgette Stone (14:57):
That's a pretty big deal.
Especially for those folks whoneed the plug and play flexibility
without hacking together solutions.
Madhura Tapse (15:07):
Exactly.
It's like giving engineers apreassembled IKEA shelf with actual
instructions that makes sense.
There are many data acquisitionmodules in the market as well.
This solution can compete verywell with these modules regarding
functionality and performance.
And the best part is ourentire solution is under $75.
(15:32):
The NAFE13388 UIM board, it just $49.
And the FRDM-MCX947N board, it just $25.
So compared to these DAQ modules, oursolution is much cheaper, but with similar
or higher performance and robustness.
Kyle Dando (15:50):
Wow.
Well that sounds like a great bargain.
I mean, 75 bucks.
They get an excellent NAFEsolution, they get a FRDM evaluation
board and they're ready to go.
That is crazy, and it's awesome whatyou guys have done on that project.
Bridgette Stone (16:06):
Well, Madhura, thank
you for the amazing deep dive into
the world of Analog Sensing and howthe new Universal Sensing Module is
bridging the gap between the Analogworld and the Intelligent Edge.
Kyle Dando (16:20):
Yeah, and we learned
that whether you're building smarter
factories, you're fine tuning your labequipment, or maybe you're trying to
automate a home or a building system.
This new module really deliverson the performance flexibility and
ease of use that the customers need.
Bridgette Stone (16:36):
Again, big thanks to
Madhura for walking us through the tech
and showing us how it all comes together.
And if you're as excited as weare, be sure to check out the
Universal Sensing Module exampleon NXP's Application Code Hub.
Madhura Tapse (16:49):
Thank you
so much for having me here.
This was super fun.
Kyle Dando (16:53):
You did an
awesome job, Madhura.
To our listeners, if you enjoyed thisconversation, make sure to hit that like
button and share it with your friend.
Make sure you subscribe and turnon notifications so that you'll
never miss one of our incrediblyinteresting conversations.
If you got a topic or a questionyou'd love for us to cover,
(17:13):
put it into the comments.
And let us know becauseBridgette is always listening.
Bridgette Stone (17:20):
That I am Kyle!
And until next time, keep innovatingand we'll catch you on the next
episode of the EdgeVerse TechCast