Posts from — November 2012
The Motion System Design magazine recently featured an article titled Trends in Automation Packaging.Â The author is John Kowal, who works at B&R Industrial Automation, and is also a OMAC board member. The article relies heavily on the views of another OMAC board member, Dr. Byan Griffen of Nestle.
Even though I am not involved in packaging machinery, I found the article interesting and want to highlight some interesting sections:
- Nestle is trying to standardize basic HMI layout, because right now every machine has its own layout.Â I think this makes sense, especially since they are not trying to specify too much detail (I could insert a car analogy here: automotive controls are similar but not identical, which really helps when you’re trying to drive off in a rental car).
- The section of HMI-based maintenance promotes adding extensions (such as interfacing to conveyors or stack lights) by modifying the HMI instead ofmodifying the packaging machine’s source code.
- My favorite quotation: “What’s more, so many processor and software advances have taken place in recent years that keeping code inside a maintenance technician’s comfort zone can seriously impair a machine’s performance potential.”
- I totally agree with that – and that it should be possible to create maintainable, extendable machines without having to modify the machine’s base source code; I have extensive experience doing this in a different market segment.
- The section “Tapping into the network” starts by saying “Sadly, there is no universal device bus or industrial Ethernet” and ends by recommending that “For communications that do not require determinism, TCP/IP and OPC suffice”.
- I think there has to be more than one standard, because different areas (process, motion control, etc) have such different requirements.Â Yes, >20 industrial Ethernet standards is too many, but we do need at least two: a non real-time version that runs on standard Ethernet hardware, and a real time version for motion control and such but require special hardware (e.g. Profinet-IRT or EtherCAT) or a special network segment (e.g. Ethernet PowerLink).Â (OK, a single standard with real time and non-real time versions could work).
- I pretty much agree that TCP/IP is a good start, but it’s even better if everyone (at least in a given industry segment) agrees toÂ higher level standards, so that just like with HMI’s, you don’t have to re-invent the wheel for each new machine you deal with.
- Universal User Requirement Specifications (uURS) documents are being prepared by the OMAC PackSpec committee.Â The idea is that the uURS documents provide a standard set of packaging equipment acceptance criteria, including electrical specifications (labeling, grounding, shielding, wiring, buttons, etc), safety, ergonomics, etc.
- Other industrials already have similar standards, such as the SEMI S2 (safety) and S8 (ergonomics) standards for the semiconductor industry.
- The last section, Simplified motion programming, discusses IEC 61131.Â I do think 61131 is a major advancement in PLC programming, but it’s still far short of what can be done with modern mainstream programming languages (such as the Microsoft .NET languages).Â Â Sometime I may go into this point in more detail.
Post Script 12/3/2012: Another detail I liked: Mr Kowal did not push B&R products (unlike many articles in trade magazines written by people working for outside companies).Â I’m pretty sure the approaches described in the article are compatible with B&R’s vision for the future of packaging machinery, but I think they’re compatible with several other companies’, too.
As far as field buses go, I think they’re like Operating Systems and Programming Languages: one standard to rule them all sounds great, unless your problem domain doesn’t fit (e.g. real time & Windows).Â In other words, I think there are valid technical reasons for more than one bus standad, but we have too many.
November 30, 2012 1 Comment
I think the data acquisition (DAQ) market could use some new players.Â For example, affordable but low jitter USB D/A converters do not exist, and Ethernet DAQ systems seem stuck in the stratosphere.
However, I recently come across a couple interesting products that are quirky, but would make a great fit for the right application:
- The Dataq DI-145 is, at $29,Â probably the cheapest USB A/D converter available.Â It has a lot of limitations (such as maximum 240Hz sample rate, fixed +/- 10V input range, and 10-bit resolution), but the price is hard to beat.
- The Data DI-149 adds substantially more capabilities (such as 10KHz maximum sample rate) for a mere $59.
- The Digilent Analog Discovery Lab is the most affordable high speed DAQ system I’ve ever seen, with dual 14-bit 125MSPS A/D and D/A converters.Â Regular price is $199.Â The documentation is pretty sparse; apparently right now it can only be used with their free WaveForms software, but supposedly they are working on an API so you can use it with your own program.
November 26, 2012 No Comments