Efficient Factory

M2M stands for Machine to Machine. Coined about 5 years ago, it is the emerging market for enabling machines to communicate directly to other machines. To put this in context, there are other types of communication that can be visualized in a matrix. I’m taking license here as I’ve found no good resources explaining this idea in a simple way

Communication in its simplest form is person to person; me talking to you — no machines involved. There are two types of communication where humans interface to machines. Think of a telephone. When you speak into it you are actually speaking into a receiver that translates your voice into electrical signals. When you listen you are listening to a speaker that sends sound waves translated from electrical signals to your ear – and you thought you were talking to mom.

Machine to machine is the promising emerging market of enabling machine to operate and trade information with other machines without human intervention. In order for this to happen a certain level of intelligence has to be embedded into a machine. Now this just didn’t occur 5 years ago this trend has been happening for decades. It’s just in the last 5 that the market has become important enough to be classified as its own segment. The sophistication of the intelligence has been increasing as well as the scope of devices and types of devices. The exciting thing is that M2M communication is getting more and more sophisticated with less and less human intervention.

The future growth is in getting disparate or non-similar machines to communicate. In the example above, a computer had to be using the same operating system and communication protocol before the days of the internet.

If you think about it it’s emerging all around us

Your car to your cell phone

Your cell phone to a satellite for GPS

Your home alarm system to your computer

Your cell phone to internet computers

Radio Frequency Identification Chips (RFID)

Expect more and more “things” to be connected in the future.

There is still an untapped market in every factory. Integration of device sensors and machines still takes a great deal of engineering on a case-by-case basis. Lowering the engineering expense will create a groundswell in the M2M marketplace.

To learn more check out M2M Magazine. This month they have a review of where the technology is going.

www.m2mmag.com           

Hurdles, Inefficiency, and Mistakes

If someone is trying to sell you a wireless system to “cut costs” run like hell, or at least do some homework. Our company has seen many of them and you had better be prepared for the headaches. Lost packets, wireless reboots, extra layers of encryption, increased latency… The simplicity of setting up a wireless access point then communicating with everything you want is easier said then done

The way to properly use wireless is NOT for proprietary data transfer. Not for large files and not for files being executed as they are transfered. It’s the nature of manufacturing. When production, Numerical Control (NC) Files are executed they are the edge of turning the simulation of the event into the event. It is the interface between the unreal (the program) and the real (the part). These files have to be executed flawlessly or you risk scrapping a part. The more complex the part; the more complex the file; the more money in scrapped parts and lost time at risk.

That’s not to say that wireless has no place on the factory floor. It just shouldn’t be your backbone. It shouldn’t be your LAN. It shouldn’t be the way to get production files to and from you CNC machine. There is just too much to go wrong on a day to day basis that you have you yearning for your manual “sneaker-net”. (Sneaker net is the use of your sneakers when you transfer a file to your machine from the office by walking)

Don’t get rid of the wires just yet, but that’s not reason to avoid wireless altogether. We’ll talk about that in the next installment.

There is a way to effectively use wireless and that is for event driven workflows. Instead of having every device on the network communicate wirelessly to the server wireless is best deployed in a manner that is a “very” local network. The immediate area around a work center is ideal for a small Bluetooth type small radius wireless zone. Think of an access point that can double as a processing center for simple information. A good analogy may be the human body. The spinal chord is makes the decision in the field by taking control of motor functions. The brain knows a stimulus has been sensed but by the time it does the spinal chord has reacted to solve the problem by taking your hand off that hot stove.

Get your ducks in row before you try to shoot them. Most software installations go awry due to poor planning. The underlying processes are culprit in the garbage in garbage out paradigm. Going wireless won’t help am unstable process. Using wireless is just a bad idea that results in more cost and a white elephant installation.

Look at the cost to deploy carefully; pulling wires may be more expensive as a hard cost but implementation of getting a good wireless signal can creep those installation costs too.

Weigh the maintenance costs of your wires vs. maintenance of your signal. Wires are pretty much fixed unless a forklift runs over them. Future changes in your shop can affect the signal as well as the hard-wired layout.

Are you process or discrete? Process industries are closer to commodities type of products, petroleum chemical the product “recipe” doesn’t vary that much over a production run. Whereas in discrete the product run can be very different form the last run or even a batch size is one unit. Process has a definite lead on discrete when it comes to computerization of processes. Standards have been developed over many years and have been accepted recently. For discrete the party is getting going. ISA is forming groups now to develop the standards.

Security is the key. Having a broadcasted signal with all of your manufacturing secrets is risky. There are more levels of security needed in wireless ands hard wired networks. As these encryption issues get settled your packet waits longer. External attack is prevalent in both, but the drive-by acquisition of your signal can be troublesome too.

Ultimately a hybrid of wired and wireless make the most sense for CNC and discrete processes. Wireless has a good future but it just may not be ready for the tremendous amount of data and the very fast processing times required on the factory floor. A loss of just one packet may make your manufactured part scrap.

Stay tuned for some more ideas on Wireless

Is it really?

In an effort to further develop their product many developers of factory floor software have jumped “on the bandwagon” by announcing wireless connectivity. Notable are the Distributed Numeric Control (DNC) purveyors. This software used to send file to machines from a central server. Here are some issues to watch out for.

·        The factory floor is a complicated place full of electromagnetic interference.

·        Wireless access points cannot always be close enough to a number of machines in order to be cost effective

·        Certain applications are mission critical and cannot be subjected to the instability of wireless data transfer. You’re an aerospace parts manufacturer and deal with very complex programs and very expensive parts. Upon failure of drip feeding DNC programs you’ll not just waste time of doing it again you also run a very high risk of  scrapping a very expensive part.

·        The only true measure of the cost-benefit analysis of going wireless is installation costs and relocation costs. Many would-be providers would have to you compare apples and oranges.

·        Don’t be fooled by low installation costs. Data access points need to be installed so they are efficiently placed and wire needs to be run to them. The wire is a small expense; it’s the labor to run the wires that’s expensive. Repeater hubs you say? Think about the cost of ownership maintenance and unnecessary added complexity.

·        Wireless technology seems to be a logical next step for data transfer models. This logical step is based on convenience and flexibility of the installation. The convenience and flexibility you’ll get later just may be sales talk.

·        Ask yourself do you often move your machines and reconfigure your shop? For most of us this is a “vapor” benefit. The cost of moving a machine can be extremely expensive so much so that the cost of running a little more wire shouldn’t be cost prohibitive.

·        You have a great investment in your machines; the processes that the data drives are key to whether you’re making money or not.  Why skimp on their reliability and functionality?

Ultimately the decision to go wireless is akin to the classic “pay me now or pay me later.” A wireless installation SHOULD cost less but when you are considering this route be aware that running wires will be more reliable in the long run. For now until there is perfection. How many files that didn’t transfer properly are going to cause you headaches? What is that worth to you?

I’ve thought about this for a while due to the issues I have with a simple wireless network. But then there was the instance that iced it for me. I was looking to update the driver for a printer and come across a new firmware version for the printer that had been recently release on the manufacturer’s site. What I found in the instructions was the warning not to upgrade the firmware over a wireless connection and to choose a wired connection. Why? Inherent issues with transmission; “the update may fail if the connection becomes unstable” Most tasks in a wireless computer network are often printing or file transfer that can be repeated if necessary. Production is critical to you business making money. If your machines aren’t running you’re not making money. It’s up to you to decide if your task is critical.

Many of the DNC packages out there are now touting wireless as the next evolution in DNC. It seems as though the DNC developers are groping for a new way to improve their product. You have to ask is this what I need or is this what I’m being sold?

If the PC world deployed in your home cannot offer a stable environment how can they do it in a hostile environment like the factory floor?

Stay tuned for the right way to do wireless.

David Strom Posted a link to his article on real-time analytics here

http://www.cioupdate.com/features/article.php/3800861/Taking-Control-with-Real-Time-Process-Monitoring.htm

It’s also available on his blog Strominator.

Going a little further than he did, real time monitoring in a manufacturing process has two advantages.

First; knowing what’s going on. The days of business report being available the next day or the next week are going by the wayside. In order for a person to act alert monitoring using machine alerts. Assets in a factory are largely production assets. If they are not doing anything, the right thing, or the right thing properly, they aren’t productive; and not making you money.

Second; creating a closed loop feedback system. This is adaptive control. By enabling a small level of smarts in a machine, that machine can capture sensor information and adapt to what it is doing. This takes the first advantage to a new level. It automates it.

A lot of what David quotes in his article is a good example of the application of technology creating a definite business benefit

I’ll admit. My timing’s WAY off. In my defense – everything’s off

We are I believe going through a fundamental shift in our economy and it’s going to take a while to sort things out. I’ve heard that Monday January 27^th will be known as “Black Monday” for the over 70,000 layoffs made by large companies; Sprint, Caterpillar, Sony… and the rest of the week’s news was congruent with its beginning.

How things are going to look on the other side? Well we’re still going to need things and want to hire people for services and need financing.

Down but not out.

Manufacturing things we need and buy has been an easy target for long time. The current situation notwithstanding, layoff announcements have been a fairly common occurrence for years. The makers of things have had their share of pain but remain fairly resilient.

There are some bright spots to talk about. American Manufacturing as compared with GNP of world countries has it ranking as the 8^th largest economy in the world.

What about growth? China (again my timing is off) has had growth rates hovering around 10% for over a decade, won’t they overtake us? In the effort of full disclosure China has  been hit hard by the recession.. On February 2^nd a Chinese Government Official reported that 20 million had lost their jobs – more than the population of Australia – and that may double as the recession continues. Growth rates all well and good measurements but they rely on a base from which you start. Per capita Income in China is around $4000/yr whereas in the States it is closer to $40,000. Even at a higher growth rate it will take a long time to become equal in consumption. Estimates are that this will happen in about 50 years.

China’s economic policy is heavily centered on exports to developed nations.

The reality is that the trend is to better jobs in manufacturing through the application of technology. Politicians like to talk about “job creation” as if any old job will do as long as it adds to the total. Manufacturing’s future is bright in the creation of more highly skilled jobs. Even now, (talking about overall trends — not the current downturn) the number of automotive workers has remained relatively constant for that last 15 years at about 1 million. American car makers’ downsizing has been offset by foreign producers opening plants.

China’s recent losses are the result of the recession but there is a supporting trend that has its cause in productivity. China’s economy is learning that it can’t compete by throwing labor at production – they must become more efficient and more productive. They are also learning that their export strategy while it’s been great in the short term they need to develop their own consuming middle class for the long term.

Check out Gary Mitchell’s (from Automation World) take on the issue

http://radio.weblogs.com/0133292/2009/01/26.html#a1916

Industry Week has two good articles out this week

First is the Cover story “OEE: Heart of the matter”. It’s really a simple calculaton but getting the data can be difficult; Here’s the link

http://www.industryweek.com/articles/oee_the_heart_of_the_matter_18211.aspx

Also see “Where Procees and Discrete Manufacturing Converge” Process industies have had the jump on discrete in terms of automation for a long time. There is a lot to share including LEAN priniciples. Here’s the link.

http://www.industryweek.com/articles/meeting_of_the_minds_where_process_and_discrete_manufacturing_converge_18210.aspx

Both of these articles are good primary reading about the manufacturing concepts involved.