While they certainly have some neat technology that can solve a lot of process problems, the number one question (sometimes more of a statement than a question) I hear is:

“Why can’t I just do this myself?”

The quick answer is that you can, but there are a lot of things to think about. Sciemetric recently released a neat little article to address that very question:

3 Reasons Why Engineers Regret DIY Quality Monitoring

Every year, I meet with hundreds of smart manufacturing, process and quality engineers across the U.S. It’s no surprise that one of the most common ways to monitor the manufacturing processes on the line are the do-it-yourself options of rigging a PLC or building a system with a computer, acquisition card and coding software. These are appealing because it appears that one can meet the requirements at a lower cost. However, I have learned from these engineers that there are also downsides to the DIY approach, downsides that do not become apparent until it’s too late.

Read the whole article here.

DataVS2 is available in four different versions according to the installed control tools: Object recognition (OBJ), Advanced Object recognition (AOR), Identification (ID) and Professional (PRO).

Features & Benefits

• The graphical user interface allows to create a new inspection in a simple and intuitive manner.
• The DataVS2 compact dimensions (70 x 52 x 40 mm) allow to install the device even in extremely narrow spaces.
• The VGA resolution guarantees high quality images thus allowing to check even the smallest details.
• The wide range of controls allows to use one model to solve many different kinds of control.
• The Advanced models (AOR) integrate more functionalities thus broadening the applicative scenarios of DataVS2.

 Info 

The DataVS2 vision sensor series presents all the characteristics able to solve artificial machine vision applications in a flexible and intuitive way. The sensor is configured via PC through Ethernet interface. The configuration software has been developed in order to lead the customer through the inspection creation process step by step.

DataVS2 is available in four different versions according to the installed control tools: Object recognition (OBJ), Advanced Object recognition (AOR), Identification (ID) and Professional (PRO).

Tutorial DataVS2 GUI

Calculate your field of view

Features & Benefits

• The graphical user interface allows to create a new inspection in a simple and intuitive manner.
• The DataVS2 compact dimensions (70 x 52 x 40 mm) allow to install the device even in extremely narrow spaces.
• The VGA resolution guarantees high quality images thus allowing to check even the smallest details.
• The wide range of controls allows to use one model to solve many different kinds of control.
• The Advanced models (AOR) integrate more functionalities thus broadening the applicative scenarios of DataVS2.

Applications

This product series has been developed to satisfy to requirements of the following reference applications:

• Part presence verification
• Object counting
• Measurement
• Part position & orientation
• Quality inspection
• Identification

In the post, he reviewed how the technology looks at the unique signature that the process has and using different algorithms and processes can aid in determining whether the part currently being produced has a signature that matches the criteria set and as such whether it is a good or bad part.

The company that makes this technology has produced a couple of very informative youtube videos.  They can be found here and here.

Light-Duty Vehicle Emissions: CNG vs. Gasoline
Testing has been performed to compare the emissions of light-duty CNG vehicles versus light-duty gasoline vehicles. For detailed results, see the Vehicle Technologies Emissions Testing: http://www1.eere.energy.gov/vehiclesandfuels/avta/light_duty/afv/emissions_description.html

The U.S. Environmental Protection Agency calculated the potential benefits of CNG versus gasoline based on the inherently cleaner-burning characteristics of natural gas, summarized in Clean Alternative Fuels: Compressed Natural Gas. http://www.afdc.energy.gov/afdc/pdfs/epa_cng.pdf 

• Reduces carbon monoxide emissions 90%-97%
• Reduces carbon dioxide emissions 25%
• Reduces nitrogen oxide emissions 35%-60%
• Potentially reduces non-methane hydrocarbon emissions 50%-75%
• Emits fewer toxic and carcinogenic pollutants
• Emits little or no particulate matter
• Eliminates evaporative emissions

As automakers have improved the emissions performance of gasoline vehicles to keep pace with stricter emissions regulations, improvements in CNG vehicles have kept their emissions performance ahead of the pack. The U.S. Environmental Protection Agency has called the natural gas Honda Civic GX http://www.afdc.energy.gov/afdc/vehicles/search/light/autos/985 the cleanest internal-combustion vehicle on Earth. Visit the U.S. Department of Energy/U.S. Environmental Protection Agency Fuel Economy Guide http://www.fueleconomy.gov/ the Civic GX with other cars for environmental impact and petroleum savings.

• It’s Clean
• It’s Cost attractive
• It’s Abundant
• The Technology is here today, and
• It supports Energy Security

Key features include:

• Available up to 450hp @ 600Vac
• CSA certified, UL listed and IEC 60947-4-2 compliant.
• The User Interface Module (UI) provides an intuitive, easy-to-use human interface with powerful configuration capabilities to maximize system performance.
• System operating parameters can be monitored enterprise-wide through a communications network. Two built-in communications ports enable the soft starter to be connected to a variety of networks including Modbus (resident), DeviceNet, PROFIBUS, and Ethernet.
• Run bypass mode greatly reduces internal heating created by the greater power dissipation in the SCRs. Bypass contactors directly connect the motor to the line and improves system efficiency by reducing internal power losses.
• Advanced monitoring and protection capabilities including overload, underload, overcurrent, phase loss, phase reversal, jam, stall, auto-manual reset, shorted SCR and over-temperature. It also monitors power consumption, voltage, current, and temperature among other parameters.
• Internal solid-state overload protection with a wide range of overload FLA settings (50–100% of rated frame current) and a selectable trip class (5–30).
• Variable ramp times and torque control settings provide unlimited starting configurations, allowing for maximum application flexibility.
• Kick-start feature enables soft starting of high friction loads.
• Pump control option with sophisticated pump algorithms on both starting and stopping that minimize the pressure surges that cause water hammer.

http://www.eatoncanada.ca/MotorControl/ReducedVoltageSolid-StateStarters/starters.shtml

Strong currency, growth opportunities

O’Leary believes that Canadian businesses have a real opportunity to make an impact on productivity while we have a strong currency. With rising input costs, he suggested that Canadian companies buy as much technology to enhance their productivity as they can while it is affordable.

“You should be buying as much technology with the strong Canadian dollar as you possibly can in this cycle to enhance the productivity of your manufacturing by as much as 15 percent. That’s how you fight input costs. Any firm that doesn’t invest now when rates are low and the Canadian dollar is high is making a big mistake,” he said. “Even though there’s uncertainly in the markets, it’s the time to be doing that.”

As part of that investment, however, he also believes that manufacturers have to sell the value of automation to the general public.

“People think that automation is killing jobs in Canada, when in fact you’re enhancing the value of the jobs you do provide. I don’t think you’re getting the message out. Automation is good. It enhances the job creation and the value creation in a big way. It’s your responsibility to communicate that because the public doesn’t get it. When they see a robot go into a plant, and a big button that you push to start it up, they think that somebody lost their job, when really somebody got a much higher paying job that came out of a great education that provides for a much better standard of living for Canadians. That’s what the message should be.”

He also encouraged manufacturers as a coherent group to lobby government to relax environmental regulations so that manufacturers here are playing on a level field with other countries. At the same time, he encourages forging closer ties with government to keep the line of communication with them clear.

“Make government your friend,” he stressed – to protect manufacturing interests and to get the industry a proper hearing.

Meet Sciemetric, a company that has pioneered in-process review and testing of your production process.  Sciemetric has developed a complete hardware and software offering that will allow you to systematically get a “process signature”, effectively a picture of the electrical signals within your process, and then compare each and every component you produce to that original signature.

Using advanced analytics, we can setup thresholds, values, slopes, and many other analytics on your process to ensure that your process is repeating itself correctly over and over, and in the event it isn’t, give you an immediate failure on the part, allowing you to quarantine the part, and more importantly, look at WHY the part failed.

I could go over this in very fine detail, but as they say:  A Picture is worth 1000 words, so check out the pictures below and the explanation that follows. (Click on screenshot to enlarge)

Here, we have several quick screenshots of a peak force check.  This could be a press exerting force on a part, or the peak force of fastening something to another using a bolt.  You will see that this part has failed due to the peak force being too much, this could cause cracking or damage to the component (like my tires bolts when the nuts were tightened too tight and cost me $400 in new studs and bolts).  Below you will see the same process on a part that has passed all of our QC checks:

Notice how quick and easy it is to determine what failed?  We could be doing dozens of checks on a process and can easily drill down from the main screen (pass/fail), to the specific checks and see what feature (check) failed.  From there we can drill down into the waveform itself to see where the failure occurred, all within seconds of the product being manufactured.

A picture truly is worth 1000 words, and possibly several 1000s of dollars.  There is a lot more to this technology that I will be drilling into in future articles, check back often to see what else Sciemetric can do for you.

It appears that for many, this is a very new protocol and there are a lot of myths and mis-information floating around, surrounding this protocol.

Myth 1: Profinet Isn’t Ethernet

I have now had several people mention this to me while discussing ProfiNet.  I believe this myth to be propagated due to the protocol name not having “Ethernet” or “TCP” in it’s name.  This myth is completely false.

To counter this myth, we look to the OSI Model.  Ethernet (also referred to as IEEE 802.3), is part of the Physical, and Data Link Layers of the OSI Model.  What this means is that Ethernet is nothing more than a means of physically hooking two devices together and allowing them to talk to each other.  Profinet is a protocol, much like Modbus TCP, TCP/IP or EthernetIP, that runs on top of this physical/data link layer.

As Profinet was designed to run on an Ethernet based media, we have the benefit of using standard cabling, in the case of the GE Intelligent Platforms products, this can be standard copper (100/1000), or fiber.

Myth 2: Profinet is Proprietary

This particular myth typically involves Siemens.  Although Siemens is a big proponent of Profinet, they do not own it, nor is it proprietary.

Profinet is in fact, an open protocol handled by PI.  PI is an open standard working body who’s sole responsibility is to develop and test against the Profinet specification.  Theirboard of directors and advisory boards contain individuals from many different companies, many of whom are considered to be competitors to each other.

Myth 3: Profinet is new Technology

New is a rather subjective term.  That being said, a technology that made it’s debut in 2003 is hardly new.  Although newer to some manufacturers, the specification and protocol have been around for a very long time.

The benefit of this, is we have an established mature protocol, that has rigorous standards that each and every piece of equipment has to be tested against in order to be Profinet certified.  Unlike most other protocols where a manufacturer only implements a small subset of the functionality and you as a user are left wondering if this will talk to widget “A”,  Profinet requires a manufacturer to put their equipment through a testing facility in order to call it Profinet capable, meaning you should know when buying this equipment that it will in fact work with widget “A”.

Myth 4: Profinet isn’t Deterministic

In order to debunk this particular myth, we first need to look at the word “deterministic”.  While many people like to look at a network in terms of determinism, a better metric to use is that of jitter.

As mentioned in the Wikipedia article, jitter is:

the undesired deviation from true periodicity of an assumed periodic signal in electronics and telecommunications

A deterministic network is not necessarily the best metric.  We can guarantee that every device is given the opportunity to talk every 10ms, however if the jitter on that network is so high that we deviate significantly, we could be recieving information in 10ms, then 14ms, then 11ms, then 22ms; not very reliable at all.

Having a network with low jitter, means we can effectively guarantee when we receive the information, and when talking about I/O this is a much more effective metric.  Taking a look at the graph below you will see a sample application measuring the jitter of a Profinet network, the overall standard deviation was a miniscule 17 nanoseconds!  Compare that with EthernetIP (considered to be solid and suitable for I/O level communications by most) that has a standard deviation of 1.89 milliseconds, and Profinet is looking to be quite reliable.

Profinet Jitter Analysis

EthernetIP Jitter Analysis

So how exactly do they get this jitter out?  The secret lies in the Profinet protocol itself.  Most “industrial ethernet” protocols utilize the TCP and IP layers of the OSI model to handle the communications between devices.  The TCP and IP layers were designed to be as robust as possible, this is why so many devices and protocols can easily use these layers.  The problem with this, is in order to be this robust they have to have a lot of elements in them, that slow these layers down.  Normally this isn’t a problem (surfing the internet for example, you don’t really care if there is a 50-100ms lag on a connection), however when we start talking about industrial I/O that type of delay could be catastrophic.  This is why Profinet does not use the TCP and IP stacks.  They have been replaced by a custom stack that allows the protocol to effectively eliminate jitter and become a rock solid protocol.

Myth 5: Profinet is Profibus on Ethernet

I hear this from people a lot.  Because everyone is familiar with Profibus, they assume that Profinet is Profibus on Ethernet (much like Modbus TCP is Modbus RTU encapsulated in a TCP frame).  This is most definitely not true.  Earlier we talked about the custom stack that replaced the TCP and IP layer, this effectively debunks the myth that Profinet is Profibus on Ethernet, but this story goes a lot farther than just that.

Profinet is so much more than Profibus ever was.  Taking a look through the specification, you will see that the PI group has thought of all of the different uses for an industrial Ethernet network, and has designed the specification around all of these thing.  You need a network to control synchronized motion?  Profinet can do that.  Peer-to-peer communications?  Profinet can do that too.  Safety?  Check out ProfiSAFE.

While Profibus and Profinet share a governing body, Profinet is a next-generation protocol that was designed from the ground up to be open, extensible, and fast.  Regardless of your application, Profinet can help you.

Summary

Having had a chance to work with Profinet, I’m quite excited with what this network will bring to customers in terms of new and exciting architectures, functionality, speed and reliability.

If you are interested, GE will be running a series of workshops in late September on Profinet.  Should you miss these, please give us a call and we would be happy to stop by and discuss Profinet and how it can benefit you.

Before getting started on the conversion, it is imperative you have the following materials at your disposal:

• Users Manual for the HE693RTM705
• IC695CMM002 Manual (Login Required)
• Lots of Coffee, or other similarly caffeinated beverage
• Good ol’ Microsoft Excel

The first step in our conversion process is to determine what our RTM705 is actually doing.  This is probably the most time consuming process and will require a lot of cross-referencing.  Make sure your spreadsheet is ready and your coffee topped up.

The RTM705 uses a communication request to handle all of it’s communications.  If you are not familiar with this instruction I highly recommend a quick read of the Machine Edition help to brush up on how it works.  To find where this instruction resides in the code, do a quick search and in the search dialog box type “COMM_REQ”, this will search through your entire program and give you a list of all of the communication request instructions in your feedback zone.  Double click on a search result to jump to that portion of logic in the editor.

Step 1: The COMM_REQ

The COMM_REQ instruction has 4 inputs to it, the enable, IN, SYS and the TAS.  The enable input is what triggers the COMM_REQ to execute, this is typically delayed by a small timer as the RTM module takes a little bit of time to come online when the PLC first starts up.  The IN input will have an address assigned to it, this is where the brunt of our configuration resides and where our spreadsheet will come in handy.  Jot down the address for future use.  The SYSID input lets us know the rack/slot location of our RTM module, this will be in the format 16#0R0S (hexadecimal, rack/slot).  This lets us know where on the RX3i this module must reside, in our example we have a SYSID of 3, meaning the module is located in Rack 0 Slot 3.  Go ahead and add the CMM002 to the hardware config in this location.  Lastly we have the TASK, this will be one of either 16#65 (port 1) or 16#66 (port 2), this may also be expressed as a decimal 101 (port 1), or 102 (port 2).

In my customers scenario, there was only 1 COMM_REQ function as they were only using port 1 (see example, TASK is 101), other situations may have multiple COMM_REQ functions, 1 for each port.  In that case these instructions would have to be followed for each COMM_REQ.

Now comes the hardest part.  Determining exactly what the COMM_REQ actually does.  In order to do this, we need to dissect the address of our IN (%R471 in our example).  These series of registers actually sets up our COMM_REQ and tells it how to behave (baud rate, number of slaves, etc), so our next step is to search through our Machine Edition software for instances where %R471 is being used.  If we are lucky it should be used in a block move statement (BLKMOV_WORD).  Double click on this in the feedback zone to jump to that location in logic, and hopefully you will see something like this:

According to the Horner manual (pg 27), the addresses equate to the following:

Several important pieces of information need to be extracted from these registers.  The first is our port configuration, we see from Address + 6, Address + 13 and Address + 14 that our port is configured as RS232, 19200, 8/Even/1.  This information can be added to our Port 1 configuration of the CMM002.

Step 2: Slave Control Blocks (SCB)

Next, we need to determine how many devices we are talking to, as well as their configuration setup (node ID etc).  From our COMM_REQ, we see that we have 1 SCB (slave configuration block), which makes sense as we are using RS232 communication, so we would only expect to see 1 slave.  In the event of RS422 or RS485 communication, there may be multiple SCB’s, 1 per slave.

To determine the setup of our slave, we look to %R295.  Notice that our SCB pointer offset is 294.  As this is an offset from 1, our SCB starts at %R295.  Examining the contents of our SCB yields the following:

The configuration of Dial Strings using the CMM002 is out of the scope of this HOWTO, so the primary things to take note of is Address + 3. This is our slave ID, and as mentioned before, there may be several of these in sequence, be sure to jot down each slave ID we will be working with. Our example only has a single slave using ID 1.

We now have our port configuration, as well as a list of all of our slaves, it’s time to take a nice break and enjoy some of our coffee before we continue on to the last part of our HOWTO, the MCBs.

Step 3: Message Control Blocks (MCBs)

This is where excel will come in to play heavily.  Looking at our COMM_REQ setup, Address + 10 tells us how many MCBs we have (how many distinct reads/writes we will be performing).  As you can see from our example, we are using 28 distinct read/writes (this is not uncommon in larger programs, those with multiple slaves may have many many more).  Address + 11 and Address + 12 tell us the series of addresses where our MCBs begin, in our case %R501.  As with the SCB from earlier, note that our value is 500 which is an offset in the register domain from base address 1, thus giving us a final address of %R501.  Using the Machine Edition search and searching for usage of %R500 yields logic that looks a little something like this:

This, like most of the other information, needs to be cross referenced against the Horner manual to determine exactly what it is saying.  Analyzing MCBR1 against the manual on page 24 yields the following:

As with all of our other configurations, notice that the RTM705 uses an offset, so we need to add 1 to get the actual address.  This MCB tells us that we are taking the values in %R300 – %R304 and writing them to slave ID 1, into the address space 40201 – 40205.  As your system probably has many MCBs which could be writing to 1 or more slaves, you can see why excel is a very, very good idea here.

Once we have this information it is quite trivial to transpose this into the CMM002.  When the hardware config portion of the CMM002 was designed, the MCB structure was taken into account, so you will see that the CMM002 needs to know whether we are reading or writing, to what address space, the address, the PLC address to write from/read to, as well as a length.  The figure below shows several entries based on the MCB from above.

Step 4: The Finishing Touches

We’re almost there!  Hopefully you still have some coffee left as we are not quite done.  Once we have setup our Data Exchanges it’s on to the physical changes.  The RTM705 used a rather large 25 pin connector which hung off the unit and routinely got caught, pulled, or ripped out.  The CMM002 uses a much sleeker RJ45 connector, allowing you to plug the cable directly into the unit and easily run it through your existing wireway.  The downside to this is that you will need to make new cables.  The CMM002 manual on page 26 has an excellent pinout diagram that will assist you in creating RS232, or RS422/RS485 cables.

Last but not least is to modify your logic.  If your existing program was executing reads/writes as fast as possible with no on/off control, you merely have to disable this logic (we recommend a normally open contact with %S8 to disable it instead of deleting it, at least until we have tested), however if you have conditional reads and writes, we are on to the bit control.

You will notice while setting up the data exchanges that the operation has a total 0f 6 available options, this allows you to fine tune your communications.  If your communications media is not always available, you can use bits to control when to perform each read and write.  In our example reads and writes were being performed as fast as possible all the time so bit control was not required, however for those of you who do require bit control, the settings tab lists the IO addresses being used for bit control.  You will notice there are 128 bits in the Port 1 Data Control section, this allows us to enable/disable each Data Exchange individually.

In our example our Port 1 Data Control is set to %Q385, so if we wanted to conditionally enable or disable our Data Exchange 1, we would set it’s operation to Bit Control Write Continuous, instead of write continuous.  This would cause Data Exchange 1 to write continuously as long as %Q385 was logic true, if %Q385 was logic false, we would not write at all.

There are a lot of additional status and control bits available in the CMM002 that allow you to tailor your application to suit your needs, however to describe them all is beyond the scope of this HOWTO.  The CMM002 manual has an excellent writeup on all their functions starting on page 89.

Final Thoughts, Cautions and Considerations

Overall, the transition from the RTM705 to the CMM002 was rather smooth for my customer, having spent a couple of hours transposing the information and re-cabling the module (this included several coffee breaks), however each installation is quite different so I highly reccomend you have the aforementioned manuals available when you do the transition.  Also there are a couple of gotcha’s and considerations I will note here:

• The RTM705 supports up to 1024 MCBs.  This is an extremely large amount of data, the CMM002 will only support 64 Data Exchanges per port.  You may need to consider consolidating Data Exchanges, or alternately moving slaves to a different port to balance the Data Exchanges.
• The RTM705 uses an offset addressing structure, the CMM002 does not.  When reading an address in the code for the RTM705, remember that is an offset from address 1.  If you read 299, it actually means 300.  When transposing this to the CMM002 you will need to put in 300, otherwise your code will not work as expected.
• Status words/bits are your friend.  We didn’t really touch on the status words, but they provide a wealth of information and are the first place you should go if your data transfers are not working as expected.
• If at first you don’t succeed, you either need more coffee or perhaps a bigger hammer, however I personally wouldn’t recommend the hammer.