Configurable I/O

Not too long ago, a senior engineer from one of the nearby chip companies was holding a Soft-I/O module in his hand and studying it. He looked up and said, “This is a magic box! It can do anything.”

Of course, this made us feel really good. He got it. He understood that—compared to any other I/O module or any PLC in the world today—Soft-I/O is unique.

One of the things that led our engineer friend to proclaim Soft-I/O the magic box is that Soft-I/O is “configurable”. There are a lot of input/output or I/O products in the world, but Soft-I/O is unique among them. Programmable Logic Controllers (PLC’s or PAC’s) are ubiquitous, but every one possesses a rigid, fixed format. Let’s look at how I/O modules work to appreciate the contrast with Soft-I/O. We offer a word of caution here: this is a pretty lengthy discussion, and you may find it a bit dense for casual reading. In fact, this is a 30,000-foot introduction to the fairly complex issue of configurability. To do it justice, we would require far more of your time. As always, we ask you for your feedback!

The Basics

First, the basics. Sensors go in to the control system, and actuators—the action of the control system—go out. That’s input/output or I/O. Electronic I/O systems have been around for more than 50 years in every factory and large building. If we look at the electrical properties of sensors and actuators, we see a very wide range. There are tens of thousands of sensors and actuators, many with unique electrical formats.  Some of these electrical formats are on/off types--which we call digital I/O--and some are continuously variable--which we call analog I/O.  Sensors that measure temperature--for example thermocouples--are generally analog whereas sensors that detect the presence of objects--for example switches--are digital.  To keep things simple in this overview, we will focus on Digital I/O.

Configurable Digital I/O

One of the most common digital sensors is a proximity sensor. In a modern semiconductor fabrication plant, you might literally count a million of them. Proximity sensors often sense the presence of metal, so they are good at verifying that doors are closed or latches sealed. They produce an on or off signal to the control system. A proximity sensor has a transistor in its output stage that can be configured in two basic ways: there is no standard since there are benefits to each configuration. We say that some of these "source" current and and some "sink" current. Well, guess what? The input system hardware needs to be different for the sourcing and sinking switch. Actuators which provide the action or output of the control system have a similar plethora of formats. A contactor to turn on a pump might require a very different interface than a valve to control pressure in a diaphragm. The list of sensors and actuators goes on and on, making connecting sensors and actuators to computers a task requiring skill and knowledge. Oh, and a lot of different hardware modules. Let’s look at some of them.

Figure 1 shows a generic digital or on/off sensor. It could be sensing anything, such as proximity of a metal latch. Or the sensor of Figure 1 might be a pressure switch that closes when a given process pressure is reached. For this discussion, we don’t care what the sensed value is. Our focus is on the interface of that sensor to the I/O system.

The sensor of Figure 1 has three terminals labeled Power+, Power- and Output. To hook this sensor up to your I/O system, you connect power and ground. Although that may sound simple, we will show that the I/O industry does not make it simple for you. Then you hook the output of the sensor to an input on your I/O system, right? Well, we say, “Maybe.” Note that the output terminal is connected to the emitter of the final stage transistor in the sensor. That means that when the sensor causes the “turn on”, the final stage transistor conducts and current is sourced to the output terminal.

No problem, you say, you merely connect the sensor of Figure 1 to the input circuit of Figure 2. What could be simpler? The circuit of Figure 2 expects that its input will source current. "No problem" becomes "Yes, problem" when we try to connect the sensor of Figure 3 to the input module of Figure 2. The sensor of Figure 3 has a so-called sinking output circuit because the output terminal is connected to the collector of the output transistor.

Hooking the sensor of Figure 3 to the input module circuit of Figure 2 will not work. Countless people over the past 50 years have made that mistake. They ordered the wrong sensor or the wrong input module and then were frustrated to see that they would not work together. That’s a fact of life, pre-Soft-I/O. Now we look at Figure 4 which is an input circuit designed to operate with a sensor configured with a sinking output stage. When we connect the sensor of Figure 3 to the input circuit of Figure 4, it works.

Some Examples

The purpose of this note is to contrast the old way with the new way. In order to show the benefit of configurability, we will hook up one each of these sensors to a conventional I/O system such as the one you would find on any PLC or PAC. We will then hook up the same sensors and actuators to Soft-I/O and show the benefits of the configurable connector of Soft-I/O. The result is a dramatic reduction in labor, panel space and hardware parts. In order to make the complexity more manageable, we will do the old-style I/O system in pieces. Otherwise it will be too complex for us to describe. We begin with Figure 5 where we hook up one sourcing sensor and one sinking sensor to on old-style I/O system.

We need two input modules, one for sinking sensors and one for sourcing sensors. If you were thinking that you just “hooked the sensors up” to the input modules, you would be sorely disappointed. Instead, you have to add a second power supply which we call the Device Power Supply. "Why do I need a second power supply?" you ask. Well, it’s a long story, but if you don’t believe us, just search online for I/O modules and see sample hookup examples. Note that they won’t show the detail that we show here because it’s so ugly.  They will only show you small examples without all the real connections.

Back to Figure 5. We hook up the Device Power Supply to the two sensors and then to the input modules because we need to reference the input modules to the power supply. Finally, we hook up the one wire from the sensor to the input module, being careful to hook the sourcing sensor to the sourcing module. In Figure 6, we hook up two actuators to two output modules. One actuator is designed to be driven by a sourcing output module while the other is designed to be driven by a sinking output module. So, the drill is the same for the outputs as the inputs. We hook up power, reference the modules and then hook up the actuators to the proper output module. When we put the old style system together, we have Figure 7.

The old-style I/O system of Figure 7 has four I/O modules, a Device Power Supply and a number of terminal blocks. There are no connectors used because everything is custom wired. Before we leave Figure 7, we ask that you look at I/O module utilization. What percentage of the I/O modules channels are used? The answer is 12.5%. Believe it or not, this low utilization is a major reason that old-style I/O systems tend to be centralized with long wire runs of large cable bundles. In the modern world, we think that centralized I/O is backwards. You will see that Soft-I/O encourages a distributed architecture.

The Soft-I/O Solution

Now, let’s look at Figure 8 in which we have hooked up a single Soft-I/O module to all the previous devices. And, oh, yeah, we threw in a thermocouple just to show how much flexibility we have ready. (Figure 8 shows only 15 pins whereas Soft-I/O has 25 pins or 11 extra for you to work with!)  Please note a few things about configurability. We just start with pin number one and hook up a wire. For a three-wire sensor, that’s three wires and three Soft-I/O pins. Notice that we did not distinguish sourcing from sinking. Soft-I/O can handle either. You just configure Soft-I/O. Now, where’s the Device Power Supply that we had with the old-style PLC I/O? Soft-I/O provides the device power. Soft-I/O can supply up to 100 Watts of power on its pins. Any pin can produce ½ amp of 24VDC. Or ½ amp of 5VDC. So, there is no second power supply and no tedious wiring. Note how simply the devices hook up to Soft-I/O. We put the first three-wire sensor on pins 1, 2 and 3. We could have chosen pins 5, 10 and 15. It does not matter because all pins are created equal.

With our sourcing sensor hooked up, we proceed to our sinking sensor.  Just pick out three more pins and then configure the Soft-I/O Soft-Device for a three-wire sinking sensor. Done. Next, we go to the actuators and hook them up. Sourcing and sinking are simply hooked up to more pins. Finally, we gently rub it in to the PLC guys. Many PLC’s cannot handle a thermocouple. Those that do will have a module that you can purchase for perhaps $500. You will need an empty slot in your PLC backplane and then install the thermocouple module. With Soft-I/O you simply pick two pins and hook up the thermocouple. Soft-I/O supports all popular thermocouple types including J, K, N, T, S, R, E and B.  (Oh, and it gives you the actual temperature--not some hex code--in degrees C, F or Kelvin.  But that's another story for another time.)

Note that we used 12 Soft-I/O pins out of 25 pins to hook up all four devices that the old-style I/O struggled with. We used two more for the thermocouple. We have 11 more for any other sensors or actuators. In summary, configurability is something that Soft-I/O invented. Old-style input/output systems are fixed configuration requiring that your wiring conform to its rigid structure. Power is not available at the input/output module. When we compare configurable Soft-I/O to the fixed-configuration old-style I/O, this is why we say that XiO reinvented Input/Output.

Wiring has always been the necessary evil of control systems. We all take for granted that thousands of wire points need to be carefully prepared before a machine can be turned on. A technician cuts a wire, strips a wire, perhaps installs ferrules, lables the wire and then inserts the ends into terminal blocks and carefully torques down the screws.  When automation professionals look at the figure to the right, they are not surprised. It’s part of the territory.

Control Wiring Properties

In this brief piece on wiring and the profound changes that Soft-I/O brings, we are going to talk about this figure a bit. Before we go too far, let’s list some remarkable facts:

1. Consider the amount of panel or enclosure area to hold your control system. Fact: More area is taken up by wiring and terminal blocks than by the controller itself. Often it’s a factor of three! Note the ratio at the right. It’s more than two to one.
2. A normal rule of thumb is that each wire point takes 30 minutes during control system build, including drawings, labor, labeling and debug. That’s a huge cost.
3. A recent study by the semiconductor industry found that the main cause of failure of wafer processing tools was wiring.

Network Wiring Properties

Now, let’s contrast the wiring required to build a control system with the type and amount of wiring required to build an office computer. Here are some facts about building and configuring office computers:

1. A wire stripper is never used.
2. All connections are made via connectors.
3. Software is used to configure the connections. DIP switches and jumpers are a thing of the past.

I know what you are going to say. The computer industry builds millions of computers. It’s unfair to compare the office computer with an industrial control system. Really? Really? Let’s see about that.

PLC Wiring

Let’s look more closely at the figure to the right. At the top of the photo is an old-style PLC, unchanged for the past 50 years save a plastic case and a bit smaller size. At the bottom of the picture is a valve manifold with 14 valve positions. A common 25-pin D-style connector mounted on the valve manifold serves to connect the valve manifold to the PLC. The purpose of the vast wasteland of terminal blocks between the PLC and the valve manifold is to do two things:

1. Adapt the signals on the valve manifold D-connector to the fixed-format of the PLC.
2. Bring in power to actuate the valves since the PLC does not supply power.

Soft-I/O Wiring Advantage

Soft-I/O profoundly changes how wiring is done, and we are going to show just how different the result is. And, yes, we are going to ask that you compare our solution to the office computer solution, a totally different problem but a solution just as elegant and not based upon high volume production.

Now let’s look at where all the wires and terminal blocks come from.

The figure to the right is a schematic of the valve manifold connection to the PLC. Actually, it’s a tiny part of the valve manifold in the photo above, because the real one would be too dense for a normal screen. What are all the wires?

Here are some of the factors that lead to complex wiring with old-style PLC’s:

1. The PLC does not supply power, so a power supply needs to be wired in.  That's the extra "Device Power Supply".
2. The power supply needs to be referenced to the PLC.
3. The valve manifold handles commons differently than the PLC, so the terminal blocks are used to sort things out.

Now, let’s look at the schematic for the valve manifold hooked up to Soft-I/O. It looks much simpler. How come?

1. Soft-I/O provides power on its I/O pins, unlike the old-style PLC, so there is no "Device Power Supply".
2. Soft-I/O is therefore automatically referenced to the valve manifold.
3. Soft-I/O configures itself to the pinout of the valve manifold, so there is no need to “sort things out”.  In fact, there are no terminal blocks at all.  A standard cable simply connects the connector on the valve manifold to the connector on the Soft-I/O module.
4. Every unused pin on the Soft-I/O module is spare and available to connect any other sensor or actuator.  Compared to the old-style PLC's where there is typically poor utilization of input/output channels, with Soft-I/O, every pin can be used.

The photo on the right shows Soft-I/O connected to the same valve manifold as in the first picture. Where are all the terminal blocks? Where is the power supply? Here are the details:

1. The power for the valves comes from the Soft-I/O module on the correct pins.
2. The return path from the valve manifold back to the power supply is through a common pin or multiple common pins which are configured on the Soft-I/O module.
3. Because the Soft-I/O module and the valve manifold use the same modern D-connector, we simply employ a standard cable. Nothing is custom!
4. Software is used to configured the Soft-I/O pins for their correct function. No DIP switches or jumpers.
5. Not a single terminal block is employed!

Please compare the photo to the right with the second photo above.  The two photos depict very much the same functional system.  The upper, old-iron system employs hundreds of wires requiring a week of wiring time for a technician.  The lower photo of Soft-I/O demonstrates how a standard cable and a single part number module can outperform the old-style PLC I/O systems.

Conclusion

Let’s go back to the outlandish claim that we made at the start of this note that we would introduce a method of wiring a control system that was as simple and effective as wiring an office computer. Yes, it’s a totally different paradigm.

• Office computers achieve their plug-compatible nature by building millions of units;
• Soft-I/O delivers the same connected simplicity via a configurable connectorized input/output system.

It’s so unique, it’s patented many times over!