Old McDonald Had a Factory…

OldMcdonaldOne of my favorite songs from childhood is:

Old McDonald had a factory, E-I-E-I-O.
And in his factory he had some sensors, E-I-E-I-O.
With a flow meter here, a level sensor there,
And a photo-eye here and a temperature sensor there,
Here a sensor, there a sensor,
Everywhere a sensor, sensor,
Old McDonald had a factory, E-I-E-I-O.

OK, so that’s not a childhood rhyme, but it does have an element of truth. There are sensors everywhere and more every day. It used to be that we only put sensors in when they were part of a control process. A curing oven, for example, would have temperatures sensors in each section of the curing oven. They were part of the feedback loop that kept the ovens at temperature.

Now we are in a world where we are putting in sensors for monitoring. We want to monitor everything, record everything, and put everything in a big database in the cloud. Instead of just having temperature sensors in that curing oven, now we have vibration sensors, energy sensors, humidity sensors, speed sensors, current sensors, and more to monitor everything about the machine. If the end product isn’t cured right, we’ll be able to look at not just the temperatures, but a whole host of factors that might contribute to a quality or operational issue. The need for sensors is exploding. It is a very good time to be selling sensors.

The problem is not, of course, collecting data. It’s moving it to someplace where we need it to be. You can connect those sensors to data collectors that can record the data – that’s what we did for a long time. Put it on paper. But now we want it digitized and somehow moved from where it gets digitized to someplace where it’s useful.

Sometimes that useful place is just the local controller. Sometimes it’s the local historian or a plant historian. Sometimes it’s an application on the Enterprise network and sometimes it’s an SQL or Oracle database in the cloud. There are lots of destinations for this data, which makes it difficult for the sensor company. They know that this data needs to be moved. They just don’t know where or how.

If they start looking at what networks the customers are using, they quickly throw their hands up in the air. Some are using wireless – 802.11 or a mesh network like Zigbee. Others are using low level networks for this kind of data like IO-Link, DeviceNet, Profibus DP, or Modbus. Others want the data on their Ethernet network. There is an alphabet soup of physical layers, transport layers, protocols, and now, IIoT networks to look into. It’s a real mess for the sensor company.

It’s pretty clear to me that industrial Ethernet is the network of the future. Even for sensors. Manufacturers are weary of all the different networks and the special platforms, protocols, and wiring that are needed. It’s inefficient, costly, and complex. What they really desire is one single network, Ethernet, that connects everything, from the factory floor to the accounting systems. Any data that needs to move is on the network. It’s simple and efficient (and maybe even cost effective, but we’ll see).

But if you’re a sensor manufacturer, how do you equip a temperature sensor with Ethernet network connectivity. Isn’t that going to cost an arm and a leg? Won’t the power requirements, the size of the chip, and the complexity be prohibitively expensive?

The answer is no! There is something brand new that’s being tested right now called low complexity Ethernet (LCE). LCE is a way for small sensors to get connected to the Ethernet world without that prohibitive cost.

You’ll be hearing more about this technology in the coming days. I’ll be writing about it. Stay tuned, or contact me here to get the details now.