How do OEMs develop communications interfaces?

kdcui

Lifetime Supporting Member
Join Date
Dec 2007
Location
USA
Posts
386
Hi all,
So I am going back to school part time, and am currently working on a student project in which I am consulting for a client in the wastewater industry. I thought this forum would be a nice resource to tap into.

The company in question is small manufacturer of analytical sensors & controllers (pH, ORP, etc...), all of which output 4-20mA.

However, they currently have no offerings that allow any industrial network communication or conenctivity (be it Eth I/P, Modbus, etc...), so these devices basically live on their own island. Although the controllers do have a built in web-page for remote access.

For a small OEM with limited software / hardware resources, how would one go about developing this? I would imagine there are companies in which you can partner with (or outsource) to develop the communications stack and hardware for you, however I am not familiar with this process.

I gut feeling is they could develop multiple communication modules that could be swapped in / out of their controllers to allow network connectivity.

Another issue is that it appears the wastewater industry is slow to move away from analog instruments in general, but that is really topic for another thread.

Any insight would be greatly appreciated.

Thank you!
 
Yes, there are companies that specialize in making communication stacks for OEMs.
If not a communication stack, then a chip that the OEM can build into his hardware.

One thing I am certain will become "the standard" is OPC UA. It is a communication standard that can be implemented both on PCs, as well as "devices". And yes there are already vendors for OPC UA stacks.
 
Hi Jesper,
Thanks for the info.

I've read a bit about OPC-UA, but it was my understanding that it is really not suited or intended to replace Real-Time I/O protocols. An interesting "blog" post at RTA:
http://www.rtaautomation.com/implementation-guide-to-opc-ua/

I know companies like Inductive Automation also are large proponents of OPC-UA and bake it into their product offerings.

Either way, whatever communications the OEM chooses it sounds like they can outsource that development. Would you happen to know of any companies that specialize in this (again, not being familiar with this side of the industry).

Thanks again.
 
I've seen a lot of successful devices that use a standard OEM module or chip. The interface on the device side can be as simple as a shift register, or Modbus RTU over a UART, or a dual-port RAM interface, while the plug handles all of the fieldbus protocol effort.

The most successful ones I've seen have been based on Hilscher and HMS Fieldbus products. When I find one a new third party device that turns out to be Hilscher or HMS under the hood, the odds of success integrating it with a fieldbus double.

I recently worked on a system that rolled their own EtherNet/IP interface, and ended up scrapping it in favor of a GridConnect version of the LanTronix XPort for EtherNet/IP. On the device side, all they had to implement was a super-simple Modbus RTU Slave over a TTL serial port. For our Volume 1 application, the sixty bucks for the XPort was recouped before lunchtime.
 
Generally when it comes to electronics development, especially in something as small as sensors you would use some sort of IC or microcontroller that is developed elsewhere that you can put your own code on to do what you wish. This kind of equipment is very specialized and I'm not that familiar with the process, but often times you can see a specific chip used in multiple different kinds of devices. Try it some time. Just google a chip number on any random board you have laying around. You will likely get a datasheet for that specific chip. This is a pretty generalized response but I hope it helps.

They would want to look for a very small controller meant to handle ethernet communications and all ADC conversions could be done on that chip and it could send out a digital signal to a controller via the ethernet coms.

Also, considering I have a wastewater plant I deal with a lot, 4-20mA signals are convenient and easy to program into controllers, whatever one you may want. Devices in a wastewater plant are exposed to very harsh conditions 24/7 and can fail quite frequently. Its pretty easy to deal with a couple power wires and a 4-20mA signal for anyone (though PoE could mean one cable for all the needs), RJ-45 tends to be quite flimsy when it comes to harsh conditions, and other connections like barrel-type connections aren't a lot better. Terminal blocks are much tougher when it comes to corrosion and wiring connections. Ethernet also uses solid copper wire, where as once again I find braided to be a bit tougher when it comes to corrosion.
 
Last edited:

Similar Topics

For custom stuff at my last company we usually did 12 weeks. My current job is much more standardized and configured stuff, but we also handle...
Replies
3
Views
1,320
Hello All, I've not tried Ignition yet but it looks good. It looks to be aimed at factory level automation so I was wondering if anyone is using...
Replies
8
Views
3,501
F
Could anyone suggest to me a PLC manufacturer that is known to be used with OEM's that has current driven outputs (0-2.5 amps to drive...
Replies
4
Views
2,201
Good Afternoon , This may be a different subject on this forum . I have been brainstorming a project with my hometown. Like many small and...
Replies
3
Views
1,787
hello, I got a job in a factory. I would like to develop my PLC programming skills. I know that every person has his own way of learning but if...
Replies
9
Views
2,465
Back
Top Bottom