Talking Smart LED Lighting Control Systems & IoT With Synapse

Introduction

Adam: This is the Wisconsin Lighting
Lab podcast. My name is Adam Rupp, and my guest today is Joe with a company
called Synapse. Synapse is a smart lighting controls company and an industrial
IoT company. Joe, let's start — just give me a few background details on
Synapse and your involvement in the controls business.

Joe: Sure. Synapse is a Huntsville,
Alabama–based company founded in 1997 as a consulting firm. They basically took
the ZigBee application — it took them six months to figure out how to turn a
light off and on. So from there they decided there's got to be a better
mousetrap. They wrote some application software that sits on top of the ZigBee
stack. So the good pieces of ZigBee that form the mesh — we kept those — and
we've supplied the buttons and the hooks to make the network easy. When you
power the device on, it's a native mesh network where the nodes are talking to
each other based on signal strength.

Adam: So it's not necessarily a
competitor to ZigBee — it's working in conjunction with the ZigBee technology,
right?

Joe: Really, our IP is the software
that currently sits on a ZigBee stack but could be ported to any type of other
standard like Thread or LoRa. For us, the use cases with ZigBee are great
because it gives us update, bandwidth, enough distance, and the speeds are
significant enough that we can do more than just lighting applications.

Adam: Why mesh technology? What's mesh
technology 101, and what are the competing communications?

Joe: Native mesh — for our Snap OS —
is something that we do, but you can do a star network or other network
configurations with ZigBee. For us, mesh — especially in large-scale
deployments where you want to save power — a lot of our applications are
battery-powered, so you don't want the nodes transmitting great distances all
the time. Mesh gives you the opportunity to talk node to node, point-to-point,
light to light. There's no bottleneck in the communications network.

Adam: So if you had, for instance, a
baseball stadium and there were 400 fixtures around a circle, every single
fixture would be able to relay the communication, correct?

Joe: Correct. Each device has its own
microprocessor — its own smarts — and it can pick its own route based on signal
strength. That's why we're fond of mesh. For lighting, of course, it's natural,
because most of your fixtures are on the same plane or in the same area, and
you can push out software very easily through the mesh.

Adam: What are the competing standards
with ZigBee? In the lighting industry there's been a lot of debate the last
decade or so about what's going to be the standard people can build their
controls platforms off of. Beyond ZigBee, what are the options?

Joe: You have Bluetooth, of course —
Bluetooth, BLE, Wi-Fi. Thread was a standard as well, and some LoRa-based
applications. They're all standards you can write to. Where everybody gets
nervous is, well, we're talking wireless — so what else is out there that's
running on those standards in my facility? And then you get into security
questions as well. For us, we picked ZigBee — that's the thing we started with.
We have a lot of applications that are non-lighting, non-industrial, where we
sell our radio device with our software loaded and customers can write their
own applications to it. So it's been battle-tested. There are limitations with
ZigBee — you can't push video over it, which a lot of people in lighting and
street controllers want — but for controls and device management, there's
enough bandwidth on ZigBee that it's been a popular choice for us.

Synapse Applications

Adam: Cool. Within the application
space, I know you guys advertise your industrial IoT capabilities and smart
lighting, but it sounds like there are other legs to that stool too — could be
agricultural products. What is the full range of applications you guys work on?

Joe: For Synapse in general, we have
basically three vertical markets. The one we started with — and what I'm
responsible for — is our core IoT development platform, where we sell our Snap
OS, which is our IP, to developers and inventors that want to integrate and
make sensors smart, or any type of device smart. We've sold to people that have
made tags for people moving throughout a hospital to make sure they're washing
their hands when they're entering and leaving a room. Our largest application
is smart agriculture, where they're monitoring for pests in nut orchards and
fruit orchards. Once you have that data highway — we're talking hundreds of
thousands of acres now — you can bring back other information and basically put
a weather station in every acre that you need. So you go from monitoring to
controlling. You can turn pumps on, get advance frost warnings, turn irrigation
pumps on, shut things down if you have an issue in the field, or address pests
— which was the original application: to have pesticide-free pest control.
Really interesting applications.

Adam: That's awesome.

Joe: We also have guys that are doing
motor monitoring — both new motors and retrofits in the market. The
applications are really endless.

Adam: I like the data highway term. In
reality, you guys are the data highway in between the sensors and the machines
and whatever's being output from a building management software or computerized
control. Do you guys do application development downstream of the wireless
standard and the sensors? I know you have an online interface that we've worked
with, but do you get into the software side of it at all?

Joe: In our industrial IoT verticals,
we do. In lighting, we have our Simply Snap software, which was originally
built to be an on-prem standalone system because of concerns with internet
security and devices being plugged into the internet. But we do have a cloud
platform version of that too, which ties into our industrial IoT where we're
trying to use that superhighway of lighting controls. You know, we have
thousands of lights inside a warehouse, and we want to get information from
devices on the floor — motors, pumps, submetering — and bring all that data
back. Where the software comes in at the backend is, you can't do everything on
a gateway. You have to be connected to the cloud in order to do the real
analytics and bring value to the data you're pulling in. So that's where we do
a lot of cloud software work.

Adam: In industrial IoT specifically,
it seems like a lot of the stuff is around production and inventory management.
Do you guys use computer vision for inventory monitoring, or what does that
sensor suite look like?

Joe: A lot of the sensors are
temperature, pressure, weight, vibration. So it's taking basic off-the-shelf
sensors and making them smart just by adding a battery-powered node and tying,
you know, a 4 to 20 milliamp output to our system and hauling it back — just
like we would interface to a lighting driver. That data can be pulled every 15
minutes, or it can be event-driven — if there's an issue, you can get an email
or text alert, or the system can automatically override it and turn something
off.

Adam: What's the ultimate goal of the
building manager, the facility manager? Is it controlling inventory levels?
What's the end result of all this — is it production throughput?

Joe: That's a great question. I guess
we're still trying to figure that out. What's the data worth? Who's willing to
pay for the data? I can speak for our parent company, which is in industrial
manufacturing and makes everything in the municipal water supply, basically.
Their concerns are inventory waste, high carrying costs of powders that are
used in fire extinguishers, airflow loss — that's a big one — so extra
compressors that they have to pay for. Inventory management. We have weights on
carts for material that's going through a process — all that adds up to extra
inventory costs. Some of it might just be asset tracking — where is the fork
truck we've been renting for two years, the lease is up and we can't even find
it?

Adam: Yes, yes. So the sky's the
limit, really. I was on a vendor visit trip recently — huge manufacturing
company, hundreds of thousands of acres — and they actually manage a lot of
inventory outdoors. They were trying to come up with a system to track inventory
because in some cases you'd have to go what's the equivalent of 20 city blocks
to get to the inventory, and they couldn't find it. So they actually found it'd
be easier to produce new product than it would be to find the inventory. So I
can certainly see the benefits of those types of capabilities.

Joe: Yes, speaking of the core
product, we have a guy that actually developed his own asset tracking system
with our Snap products. And that was the reason — he was at a very large
manufacturing facility. The story about the missing fork truck, that's a real story,
it's a real thing. So it really depends on what they're looking for. A lot of
it is, with the incentives from the utility companies, power monitoring and
getting reporting so you can get energy rebates, and smart use of your energy
of course.

Adam: In the fork truck situation, was
it a GPS tracker of some kind, or was it more or less just something to connect
with your data highway to know exactly where it was in the facility?

Joe: He wrote his own application
software and his own user interface, based on the equipment. They had a barcode
scan when you sit in the fork truck that tells you all your OSHA security
things, the kind of clearance to run it. He set up repeater pings throughout
the facility — which is one of the largest facilities undercover under one
roof, next to the Pentagon. His accuracy was within six feet, today, based on
signal strength.

Adam: That's really good. And on the
compressed air side, would that be like a pressure sensor that then connects to
your data highway?

Joe: It's an actual airflow sensor. So
basically the plant would put those in throughout their process and check for
air leaks using those airflow sensors. Having a baseline of knowing what the
pressure should be at those areas, they could detect leaks and get those fixed.

Adam: You said you can't send video
data through your network. We're in the early stages of looking into using some
computer vision for data analytics on various applications. The basic idea is
you get a camera 10, 20, 30, 50 feet in the air on a light pole — you can do a
lot of stuff with that. What type of network would be required in that case for
video?

Joe: For video, you would definitely
want power-over-ethernet, or Wi-Fi could certainly handle that video
requirement. For that, we have customers that integrate lighting controls —
they have their own system that's doing video and sound and architectural lighting
for city streetscape lighting — and they have all that technology. What they'll
do is they'll use their technology for four or five different RF devices on
those poles, and then in the surrounding areas they'll use our network and tie
the two together via API, so the customer only sees one interface. So it's a
lower-cost control function for the — I wouldn't call them dumb poles, but the
ones that aren't lighting up the city street, which might be a couple hundred
poles. There may be a thousand poles around the area that still want controls
and data management, and we'll use our network and tie them together into one
dashboard.

Sports Lighting Applications

Adam: On the lighting topic, which is
our business — sports lighting — I know you guys do a lot of sports lighting
controls. What are our customers, what are end-users trying to accomplish with
sports lighting? I know it's one of those situations where the sky is the limit
in terms of what can be done, but if you had the 80/20 — narrow it down to the
things that provide the most value for the best cost architecture — what do you
think people are after?

Joe: It really depends on the end user
and the application. Obviously, the small municipalities care about not having
power on when it shouldn't be on, or having a guy there, or having coaches and
players with access to the lighting controls all the time. For them — say a
City Parks Department — it may be: we want to set schedules, turn lights on and
off remotely, so we don't have to send a guy out every time someone wants to
use the field. We want practice settings, we want gametime settings. There's a
lot of that in the municipalities. When you get up into the larger scale
Division 1 and the pros, then it's more flash — they want to see light shows.
So it really depends on the application. With our system, we try to make it
flexible enough to handle both. We have a very simple user interface, and you
can set different administration levels on it so you can have the coach be the
user and the city plant manager or city manager being the administrator. But I
would say most of the time, it's mainly switching and schedules.

Adam: We kind of see the same thing.
We've got some larger scale applications where they want the fancy light shows,
but like you said, a lot of it comes down to: if you're going to be paying and
investing in a solution that limits energy, let's make sure it's not turned on.

Adam: You mentioned power monitoring —
this is something we're taking a close look at right now. Solid-state
electronics are more susceptible to power issues than some of the legacy
lighting technology. When you say power monitoring, is it energy usage, or are
you guys actually looking at voltage transients and other types of dirty power
issues?

Joe: For us it's more energy reporting
— to meet DLC specifications or Title 24. DLC 5.0 is coming out next year, and
I think power monitoring will probably be a requirement, the way it looks. So
for us it's just energy reporting.

Installation, Support, & Troubleshooting

Adam: On the sports side, I know you
guys have commissioning services, ongoing service contracts or service models.
What does that sales process typically look like for a sports lighting
application?

Joe: Honestly, it really depends on
the lighting OEM. They have their choice. Some people just want to sell lights
and have a controls company — they'll say, call this guy, give you a phone
number.

Adam: That's exactly...

Joe: Some people actually want to run
it as part of their value-added services, and that's certainly an option. We
have a remote access where we host the cloud — for an annual contract fee you
could have up to 50 of your sites being monitored, and then you can parse those
off individually as you see fit. We can do it that way. Or, if you hand us over
the end-user and give them our 800 number, they can contract that service with
us. Or they can run it behind their own VPN on their own network.

Adam: So when issues or
troubleshooting situations come up, what are often the reasons behind it? Is it
an educational component where the end-user was potentially sold a really cool
product and they weren't educated on it? What do you guys see come up as the main
pain points?

Joe: A lot of it stems from who's
installing it, who is putting the map together, who's getting the floor plan
together. We have a thing called a census that goes out and finds all the
radios that are in that area for that application and pulls them onto the map.
But if nobody's really laid out the map with the proper MAC IDs — where these
things are located — then it becomes kind of hard to build out the map. It's
easy to control, but the end user typically wants to make sure the lights are
in the location you say they are on the map. Then you can set schedules and
zones based on that. After initial startup, we really don't see a lot of issues
— other than a guy forgot his login password.

Adam: It's interesting — we work with
a lot of contractors, electrical professionals, and I think there's a bit of a
lag in knowledge and training where your traditional journeyman or contractor
is very, very good at traditional products, hooking up basic products into
their electrical systems — they do a phenomenal job at that. But it's almost
like there needs to be a separate skill set where it's Joe the contractor meets
Geek Squad.

Joe: Exactly.

Adam: I've always wondered, with
larger industrial facilities, do you guys work with their IT departments more
so than their building and maintenance departments?

Joe: I tell the story all the time —
I've walked into projects to commission them and the IT guy had no idea this
was even going on, and that there was going to be another network running.
Everything just comes to a screeching halt. You have to make sure you keep
those guys involved from the front end. And to your point on contractors and
electrical, the other biggest thing we see is missing antennas — antennas
mounted in the wrong spot or not connected. Those little things, we need to do
a better job educating before things get installed in the field. We go out for
commissioning — there are a little bit of hiccups, but yeah, IT is the critical
one. They can stop the presses real quick.

Adam: And in IT, a lot of this is — I
won't say it's old news, but it's already part of that skill set. So I think a
fusion of both of those skill sets is needed. From a compliance standpoint or a
National Electrical Code standpoint, who are the folks that have to install
these? Do they have to be a licensed electrical contractor, or can the IT
department in the long run manage the install?

Joe: Typically it's going to be a
licensed electrician — either a plant electrician, or if you're selling through
the energy services market, usually they have local electricians that are doing
the install. From a commissioning standpoint, certainly it can be the IT guy.
But for the physical installation, it's typically electrician.

The Future of IoT

Adam: So the future of IoT — IoT, AI,
computer vision, 5G — there are a lot of buzzwords involved here. Every time I
read about things or watch YouTube videos on how this big IoT wave is coming,
then I think, well, on the commercial industrial side it's been around for a
long time — automated manufacturing, robotics, the things you guys are doing.
It's interesting how there's a disconnect between the past and the future. I'm
sure on the industrial side things will continue to get better, but do you see
it as a lag in the retail space or the residential space, with the technologies
you guys are already deploying? Or is it indeed a new technology on the retail
and market side of things — home building, everything from digital thermostats
to refrigerators that can send you a notification that you need to replenish
your tomatoes? Is it the same stuff in a new application, or are there new
standards and technologies being developed?

Joe: That's a great point — it's a
combination of both. I would say the home consumer market is way ahead of the
industrial side, really. You think about it — your phone has how many different
radios; you can connect to anything from anywhere already with it. Home
thermostats, the Philips Hue lighting, multicolored lighting controls, Ring
doorbell guys that can alert you. That space is already using IoT. But think
about who — there's data, and the big guys love data. The Googles, the Amazons,
the Microsofts of data. So they can push that in. When you flip that over to
the industrial commercial space, then you start — somebody's got to pay for it,
somebody's got to understand it, somebody's got to maintain it. It's still in
its infancy, I think, from the industrial standpoint. Your point on AI — one of
our CEO's visions is all this machine learning, the artificial intelligence
machine learning with this data that we are trying to help pull into the cloud.
That is the real future, that's the real endgame, and it's really going to
transform manufacturing. We talk about Industry 4.0 — I'm not a big buzzword
guy — but you talk about Industry 4.0 and all these things when you go to an
IoT show, and everybody's talking about them. We're trying to go from the node
to the gateway to the cloud. Those are the three big pieces. You have to get
the data, you have to get it somewhere, and then you get it to the cloud, and
that's where all the analytics and the real power come.

Adam: Getting it to the cloud is the
data highway that you guys have created.

Joe: Yeah, and we're just one flavor
of it. But that's the other issue too — trying to educate. There's no standard
in gathering that data. So that's the hard part right now. People are like,
alright, we get it, what do we do with it? Is that the right mousetrap for us?
So it's a definite education process. But if you believe everything you read —
or, at least if you believe you're not supposed to believe a small fraction of
what you read — it's still in its infancy. It still will transform the way
manufacturing is.

Adam: When I said it seems like it's
been around for longer on the industrial side, I think what I was referring to
is at the machine level. You look at, whether it's using optics to track
throughput, or other types of automated manufacturing — that stuff's been
around for a long time. I think the difference is, it's not connected. So it's
at a local machine level, but the data isn't being accumulated in a way that
can be understood and potentially made sense of.

Joe: Right. We came out with the term
“illuminating dark data” — I think it was a marketing buzz a few years ago —
when we were talking about lighting as the superhighway of data. But if you
think about the dark data on a factory floor, or even in a sports facility —
what are you missing that isn't connected to a PLC, or a robotic system, or
your BACnet system? There are pieces of equipment that aren't being monitored,
and it's expensive to try to get those connected and hardwired to your PLC
backhaul. So the idea is to have these plug-and-play, off-the-shelf sensors
that become smart and become part of that information system. You take that
data — I joke about all the cellphone photos that are taken now, all that data.
You use the ones that are good — back in my day, you had to care about what you
were getting developed, and now you just take 10,000 pictures and see what
works. It's similar with data — you're bringing all this data back, you're
taking those analytics, and you're applying what you've learned from those
analytics and changing your production.

Adam: So you have customers or markets
that are grabbing the data and storing it for future use? Or is it just a case
where the sensor infrastructure is there, the data highway is there, and from
there you need some type of a server either on-site or off-site? Do you have
people collecting it knowing that down the road they're going to be able to
make sense of it? Or are they using it immediately?

Joe: I think we've had a few proofs of
concept at some of our McWane factories where they've learned things they
didn't even know were happening, so they've made immediate changes. But
certainly they can store the data over time and year-over-year take a look at
this piece of operation versus that piece of operation, and make executive
changes to the way they're operating those areas of the plant.

Adam: Cool. Anything else you'd want
to touch on?

Joe: No, I don't. Yes, very exciting
time. I think we're still in the infancy of it, and we'll see what the next
three to five years bring.

Adam: It's all about making dumb
lighting products into smart computers, right?

Joe: Exactly. There will be a
computer. Thank you very much.

Adam: Thanks a lot.