We of PROFIBUS and PROFINET International (PI) are in the middle of our 21st meeting. We are in Krakow and moments ago we had our picture taken by Geoff Hodgkinson:
(click for larger image)
The size of the community is amazing. We have about 80 folks from 22 countries and 6 of the 7 continents here. There are 25 Regional PI Association (RPAs), 42 PI Competence Centers (PICCs), and 16 ~PI Training Centers (PITCs).
More technical details to come…
The question was prompted at the Vancouver PROFINET one-day training class. (Yes, we get PROFIBUS questions at the PROFINET class and vice versa.)
Here’s the summary:
DPV0: Cyclic Data Exchange, Diagnostics (all devices)
DPV1: Acyclic Data Exchange, Process Alarm Handling (process automation)
DPV2: Clock Synchronization & Time Stamp (drives)
For more details, read the “PROFIBUS System Description” (available in seven languages).
It’s important to note that the versions build one on the other and are interoperable and backward compatible. In practical terms, this means that you can have, for example, a V1 controller over V0 and V1 slave devices. The V0 devices will only report cyclic data and diagnostics (but these devices don’t have more to report anyway).
We generated a bit of internal discussion from this question revolving around PROFIsafe. We show a timeline that shows PROFIsafe as being part of V1, but should more accurately indicate that its association with V1 is strictly from a timeline point of view. PROFIsafe is an application profile and can therefore be used with any version of PROFIBUS… and with PROFINET.
After all the discussion, I’m finishing this on the plane for Poland and our international gathering of RPAs (Regional PI Associations) and PICCs (PI Competence Centers). I’m looking forward to the information exchange with other regions of the world. (We have about 80 folks from 22 countries and 6 of the 7 continents – including the Indian subcontinent. Alas, we have no RPA in Antarctica.) I’ll blog and tweet interesting news from Krakow.
Now I’m posting from Chicago since yesterday’s flight across the water was cancelled due to mechanical issues. My first trip to Europe for PTO also had an unplanned 22-hour layover in Chicago.<sigh>
The question came up at the PROFINET one-day training class in Vancouver: “Is PROFINET routable?” In other words, can PROFINET messages be sent through a router?
First, a little background. In an Ethernet network, several types of device can be used: hubs, switches, and routers. Each works at a different layer of the ISO/OSI reference model. Here is a summary of the ISO/OSI reference model:
(click for larger image)
Hubs work at layer 1 of the ISO/OSI seven layer model. Hubs are not recommended for industrial networks because they send an incoming message on any port out of every other port. This creates too much traffic to maintain deterministic behavior on the network. Switches operate at layer 2. In the Ethernet world this uses the physical address of the device – the MAC address. Layers 3 (IP) and 4 (TCP or UDP) are not involved. Routers use layer 3 and communicate using the IP address. Routers are necessary when communicating from one subnet to another. (For more on subnets, see our archived webinar, “Industrial Ethernet, Advanced Ethernet Architecture.”)
PROFINET communicates in two ways: via TCP/IP and via the PROFINET real-time channel. The PROFINET real-time (RT) channel basically skips the TCP/IP portion of the communications stack. In other words, PROFINET RT uses layers 1, 2, and 7 of the ISO stack just like PROFIBUS and other fieldbuses do. We do this to reduce jitter. But since it does not use IP, it can’t pass through a router. Fortunately, PROFINET CBA (Component Based Automation, our peer-to-peer communications) allows us to use TCP/IP for communications, too… and that is routable. So the short answer after the long background: Yes, PROFINET is routable using PROFINET CBA. (Caveat: using TCP/IP and going through a router will introduce some modest delays compared to PROFINET RT.)
More on the Vancouver class: Most attendees indicated on the course evaluations that the course length was “just right” and several though it “too long.” Perhaps best summarized by someone as “Long, but good.” I agree that it was too long, actually. A couple folks suggested we cut the introduction down. We will have that done before our next class in Boston on July 23.
Other comments:
“Effective use of our time. Webinars are a great feature.”
“Worked well with two instructors (also there was consistency with the webinars).” [You can find those webinars here.]
“Well paced. Well led.”
“Instructors have impressive knowledge of PROFINET.” [They probably meant Hunter, but I’ll bask in his reflected glory.]
I was one of four presenters at GlobalSpec’s virtual conference and exhibit. If you need a really brief introduction to PROFINET, the presentation is archived in the “Conference Center” there. I wasn’t able to answer all the questions, but look for a document with all the questions answered on the site as well.
How do you gauge the growth of your favorite technology without an expensive market company study? (PROFINET being the technology; it’s my favorite.) One way is the number of people coming to the free one-day training classes. (An ever-increasing number.) A second is those taking the next step and clamoring for the next class – the full week Certified Network Engineer class. (We’ve had to add a class for July 27-31. Seats still available.) Third, how busy are the test labs in certifying new PROFINET products. (Really, really busy. Device manufacturers need to reserve their test slot early!) The economy may be down, but clearly many of you are using this time to prepare for the inevitable upturn.
One of the other questions that came up in Grand Rapids came up during the PROFINET section and related to safety. I thought it worth expanding on the answer to that. The question: “Can I use PROFINET over wireless in a PROFIBUS PA environment?”
First, let’s be clear that we are talking about two different kinds of safety – intrinsic safety and functional safety. Intrinsic safety is for atmospheres where a spark can cause an explosion. Functional safety is protection against machinery malfunctioning and causing damage or injury.
The answer to the question then is yes and no. No, you can not use PROFINET in an explosive atmosphere; there is no intrinsically safe Ethernet standard. Yes, you can connect a functional-safety-rated PROFIBUS PA device through a PROFINET proxy and from there wirelessly to a safety controller. PROFIsafe is the technology that allows safety over a bus regardless of the transmission media (serial or Ethernet, wired or wireless). More on PROFIsafe. Or click on the PROFIsafe or safety tags in the right-hand column. (I’ve started updating past posts with tags. I’m back about 8 months so far. I really like the new blog software!)
There was a good crowd at the Grand Rapids PROFIBUS one-day training class last week. Manny and Torsten were there while I was in Vancouver with Hunter for a PROFINET one-day training class. One of the questions that came up in Grand Rapids concerned how many PROFIBUS DP devices are allowed on a network. We have a slide that addresses that:
(click for larger image)
The questioner thought that the four devices with no address still counted against the 126 allowable addresses on a network. They don’t. They do count against the number of devices on a segment and that probably accounts for the confusion. The address space is independent of the network segmentation. (Segmentation is provided by the repeaters.) So while the repeaters count against the physical number of devices that are allowed on a segment, they do not have an address. Why the limit of 32 physical devices on a segment? That’s a limit of RS485 that PROFIBUS DP is based on. For more details see the Wikipedia article and especially the B&B Electronics link at the bottom: “Technical library of RS-485 articles and application notes.”
Bottom line: we’re talking about two different domains – physical and virtual. The physical with limits imposed by the RS485 standard and the virtual address space independent of the physical with limits from the PROFIBUS specification. Or, more basically, the devices with no address don’t subtract from the total number of addresses available.
As always there were suggestions on topics to increase and topics to decrease. As always some wanted to increase and others to decrease the same topic! We continue to try to balance that.
Other comments from the Course Evaluations:
“Great course. Learned a lot from ‘the bottom up.’ They (correctly) assumed I knew nothing about PROFIBUS.”
“Was good to have vendors here. Was glad there wasn’t a sales pitch.”
“The Canadian had an awesome sense of humor.” [Yes, Manny, the Canadian, went to Michigan and Carl, the American, went to Canada. I do have to note that I lived in Canada for a couple years and speak nearly-fluent Canadian.]
This month’s IEEE Spectrum magazine features articles about Mars exploration. A related article “My 10 Favorite Mars Novels” by Kim Stanley Robinson reviews a century’s worth of Mars science fiction. In the introduction, he mentions A Princess of Mars by Edgar Rice Burroughs and War of the Worlds by H.G. Wells. [Click on those titles for the free download from Project Gutenberg.] I will admit to reading War of the Worlds and all ten of Edgar Rice Burroughs’ John Carter of Mars series (A Princess of Mars is first in the series) and a few of Robinson’s top ten. Robinson’s own Mars books are here
1. Two Planets
2. Red Star: The First Bolshevik Utopia (Soviet History, Politics, Society, and Thought)
5. Outpost Mars: A science-fiction novel
9. Man Plus
