Relay Fundamentals

Open an HVAC control panel, and you’re likely to find at least one relay mounted inside. Sometimes you find lots of relays! Even though relay technology has been around since the 1800s, there is still a need for relays in control panels.

What components make up a relay? How are relays typically shown in wiring diagrams? What functions can a relay perform? Stay tuned for some answers…

What Components Make Up A Relay?

A relay is constructed from the following components:

  • Electromagnet (coil of wire with metal core)
  • Armature (arm that moves when attracted by the energized electromagnet)
  • Spring (keeps the armature retracted when the electromagnet is de-energized)
  • Contact(s) (one or more electrical contacts that open or close when the armature moves)
  • Electrical Terminals (so you can make connections to the electromagnet and contacts)

How Does It Work?

You might have built an electromagnet in high school science class and used it to pick up small metal objects.  An electromagnet is a coil of wire wound around a metal core.  When the coil of wire is hooked to an electrical power source and current flows through the coil, a magnetic field is produced which surrounds the coil and the core:

While the electromagnet is energized, any metal objects which come close will be physically pulled towards core.

When the switch is opened, the current flow stops and the magnetic field disappears.  When the magnetic field disappears, nearby metal objects are no longer attracted to the core:

We now place a pivoted metal arm (armature) above the electromagnet, and we connect a spring to keep the armature pulled away from the electromagnet under “normal” (electromagnet de-energized) conditions:

If we now energize the coil, the magnetic field will overcome the spring force and pull the armature down:

So… we have an armature which we can move up and down by de-energizing or energizing the electromagnet coil.  What do we do with it?

We use the moving armature to open and close electrical contacts which are separate from the coil circuit.  We can do it like this:

In this example, we’re using the metal armature itself as the “Common” or “Wiper” for a double-throw switch.  The tip of the armature makes contact with one of two possible points depending on the position of the armature.

When the relay is de-energized, the armature is up and the wiper makes connection with the “Normally Closed” (NC) contact.

When the relay is energized, the armature is down and the wiper makes connection with the “Normally Open” (NO) relay contact:

This example relay would be known as a “Single Pole, Double Throw” (SPDT) or “1 Pole, Double Throw” (1PDT) relay.  It has one set of contacts (1 Pole) with both Normally Open and Normally Closed (Double Throw) connections.

Relays can have more than one set of contacts (poles).  In the next example, the metal armature is not part of the electrical circuit but instead moves a plastic connector that activates two sets of contacts:

This would be known as a “Double Pole, Double Throw” (DPDT) or “2 Pole, Double Throw” (2PDT) relay.

The number of poles (contact sets) can be expanded (3PDT relays, 4PDT relays etc.).  Typically 4 poles is the largest relay we would see in an HVAC control panel.  If you need more than 4 poles in a relay, you would probably use two relays with the coils wired together in parallel.  Shown below are a DPDT relay and a 4PDT relay wired together to create 6PDT relay action:

 

Electrical Ratings of a Relay

The coil of a “classic” relay (no electronics) is typically rated for a specific voltage and either AC or DC operation but not both.

There are some modern relays (like Functional Devices “Relay In A Box”) that contain electronic circuits that allow the coil to operate over a range of voltages and either AC or DC operation.

If in doubt, the relay’s datasheet should clearly spell out the acceptable voltage range for the coil and whether it’s AC only, DC only, or AC/DC compatible.

Relay coils are generally tolerant of some variation in the coil voltage, a typical datasheet specification is 80% – 120% of rated coil voltage.  So if a 24V transformer is “running a little hot” (putting out maybe 27V) or there is a bit of voltage drop in the coil connecting wires (maybe the coil is only getting 22V) the relay will still function just fine.

The relay contacts will have ratings for the maximum voltage allowed and the maximum current allowed.  Sometimes there are multiple ratings depending on the type of load being switched (for a motor load sometimes the contacts are rated in horsepower for example).  The circuit being switched by the contacts should not put more voltage or current on the contacts than allowed by the datasheet.

Relay Symbols Used in Wiring Diagrams

The relay coil may be drawn several different ways:

The relay contacts are typically shown one of two ways:

Relays go by several possible names on an electrical drawing:

‘R’       (not a great idea since resistors can have this designation too)
‘RL’     (better)
‘RLY’   (better still)
‘K’      (Where did this come from?  We don’t know.  But it’s widely used!)

If there are multiple relays, the name will be followed by a number or a letter as in RLY1, RLY2, RLY3 or RLYA, RLYB, RLYC.  There might be a hyphen (-) between the name and number or letter too.  There is no specific standard on how relays must be named.

Sometimes the relay is shown as a single assembly with coil and contacts arranged in a rectangle:

But sometimes the coil and the contacts are scattered around the drawing:

Relay Terminal Numbering

Unfortunately there does not seem to be a standard for the terminal numbering on relays.

Most of the relays we use in HVAC panels plug into a base socket.  The pins/blades of the relay have assigned numbers and then the bases they plug into have assigned numbers on the screw terminals.  If the same brand of relay and base are used together, the blade numbers usually match the terminal numbers on the base.  If, however, different brands of relay and base are used together, the blade numbers may not match the base terminal numbers even though the two parts are physically compatible (example:  IDEC SH3B-05 relay plugged into Omron PTF11A base).  It’s a good idea to use the same brand of relay and base to avoid confusion.

Some electronic relays such as Functional Devices “Relay In A Box” have color-coded flying leads rather than screw terminals.  Consult the device datasheet to determine the function of each color-coded lead.

Functions a Relay Can Perform

So what useful functions can a relay perform?

  • Use a low voltage and/or low current signal (on the coil) to switch a large voltage and/or large current circuit (on the contacts).
  • Use a DC signal (on the coil) to switch an AC circuit (on the contacts) or vice-versa.
  • Use one signal (on the coil) to switch multiple circuits (multi-pole contacts).
  • Invert the sense of a signal (using the Normally Closed contact).
  • Use a signal from one circuit (on the coil) to switch another circuit (on the contacts) where the two circuits are not allowed to have any direct electrical connection.
  • Use one signal (on the coil) to switch between two alternate loads (on the contacts).
  • Various combinations of the above.

The example below demonstrates a relay performing a combination of several of the above functions.  A low voltage, low current control signal from a controller is switching a high voltage, high current load on and off.  The relay is also allowing a DC control signal to switch an AC load.  There is also electrical isolation between the drive circuit and the load circuit so that the controller is protected from the high voltage 240V power source:

In the next scenario a smoke detector signal is being sent to three different systems simultaneously while maintaining isolation between the systems to make sure there is no unintended interaction between the systems:

Here is an application where we want to reverse the sense of an electrical signal.  We have a foil alarm loop glued to a glass door.  When the foil is intact (current flowing through foil) we want to hold off the Alarm light and horn.  When the foil is broken (current flow stops) we want to energize the Alarm light and horn.  We can get the reverse action we want by using the Normally Closed contact on the relay:

Let’s see how we can use both the N.O. and N.C. contacts to get “alternate action” between two loads.  We decide that in addition to the ALARM light, we also want a NORMAL light for our alarm circuit that is on when the foil connection is intact.  So we want the action of the NORMAL and ALARM lights to alternate, one or the other is on at any time, but never both at the same time.  We can get this alternate action by adding the NORMAL light to the N.O. relay contact:

You may begin to see how versatile relays are.  Even with all the high-tech electronics in control panels, relays will be with us for a long time to come.  The operation is easy to understand and they can be used for many different applications.

Magnetic Latching Relays

There is a special class of relays known as “magnetic latching relays.”  These relays contain one or more small permanent magnets which hold the armature in the last commanded position when the coil drive is removed.  When a coil drive signal is applied, the magnetic field from the coil is stronger than the permanent magnets and can thus move the armature to the opposite position in spite of the permanent magnet’s presence.

Magnetic latching relays must have two coil drive commands, Latch and Unlatch.  Some relays accomplish this with two separate coil windings and two command leads:

Other magnetic latching relays have just two coil wires, but the polarity of the applied DC drive determines whether the command is Latch or Unlatch:

The Latch and Unlatch drive signals are typically meant to only be momentary, you normally would not leave either signal on the coil continuously.

Magnetic latching relays are frequently used in lighting circuits.  One example is the popular General Electric RR7 lighting relay.

Wrap-Up

Some points to remember about relays:

  • Make sure that the relay coil voltage and AC/DC rating matches the coil drive signal in the application circuit.
  • Some relays are available with a “coil energized” light built in. If any troubleshooting has to be done on the panel, these lights can be very helpful.
  • Some relays are available with a manual override button built in. If any troubleshooting has to be done on the panel, these override buttons can be very helpful.
  • Make sure that the relay contact voltage and current ratings are at least as high as the volts and amps that will be applied by the application circuit. If possible, use contacts rated higher than the volts and amps that will be applied.
  • If you need more poles than are available on one relay, you can use multiple relays with the coils connected in parallel.

Divide & Conquer Those Hard-to-Read Flow Meter Pulses

Some HVAC applications require reading and totalizing pulses from flow meters.  This sounds simple enough, just take the pulse output from the flow meter, connect it to a Binary Input (BI) on your controller, and set up the program logic to count pulses coming in on the BI.  What could go wrong?

Unfortunately things are not always as simple as they appear.  HVAC controllers frequently operate on the “scan” principle where the inputs are not read continuously, but only once per controller scan.  Controller scan time could be fast or slow depending on the controller design.

The controller only “sees” a pulse if the input is low on one scan and high on the next scan.  This means that:

  1. A pulse whose duration is longer than the time between controller scans should always be reliably detected:

    upd1
  2. A pulse whose duration is less than the time between controller scans will sometimes be detected and will sometimes be missed:

upd2

upd3

NOTE:  some controllers have one or two “high-speed counter inputs” in addition to their regular Binary Inputs.  High-speed counter inputs have a much higher scan rate and may be fast enough to detect your flow meter pulses reliably with no additional hardware needed.  Always check the controller datasheet to see if your controller has any high-speed inputs you can use for connecting your flow meter.

What can I do if my flow meter pulses are too narrow to be reliably detected by my controller?

Kele sells a Universal Pulse Divider (UPD-2) that can divide the incoming meter pulses by a selectable divisor and output a wider pulse that your controller can reliably detect.

What kind of pulse signals can the UPD-2 accept on its input?

The UPD can accept pulses from a simple contact closure, a transistor switch, or a driven 0-5VDC signal:

upd4

upd5

upd6

In the Low state, the output of the flow meter must be able to sink approximately 2.3 mA of current.  This will be compatible with most flow meters.

What kind of output does the UPD provide?

The UPD output is an optically-isolated electronic switch.  It is polarity sensitive.  In the High state it is open-circuit, in the Low state it closes the circuit.  In the Low state, the output can sink up to 6 mA of current.  In the High state, it can tolerate up to 30VDC.  The UPD output is compatible with most controller Binary Inputs which are typically configured as an internal reference voltage with a pull-up resistor as shown below:

upd7

What is the minimum high or low pulse time the UPD will detect on its input?

The UPD will reliably detect any pulse with a high or low time of at least 10 milliseconds.  This limit is programmed into the UPD’s microcomputer firmware, the hardware could actually detect shorter pulse durations.  But HVAC equipment lives in electrically noisy environments, and we don’t want to start counting noise glitches as flow meter pulses, so we set the minimum valid pulse duration at 10 milliseconds.

How do I set the Divisor value on the UPD?

The UPD has an 8-slider dipswitch for setting the divisor.  Each slider has a divisor value assigned thus:

  • Slider 1 = divide-by-1
  • Slider 2 = divide-by-2
  • Slider 3 = divide-by-4
  • Slider 4 = divide-by-8
  • Slider 5 = divide-by-16
  • Slider 6 = divide-by-32
  • Slider 7 = divide-by-64
  • Slider 8 = divide-by-128

The total divisor value is the sum of all the sliders which are turned on.  For example, if sliders 2 and 3 are turned on, the total divisor is 2+4 = 6.

** After changing the dipswitch sliders, be sure to cycle the UPD power as the dipswitches are only read by the microcomputer chip at power-up.

How does the division logic work on the UPD?

It’s important to understand how the UPD division logic works.  The UPD counts all the up-and-down signal transitions on the incoming pulse.  For every N up or down transitions on the input signal, the UPD makes 1 transition on the output signal (where N is the total divisor selected).

This is one of those times where a picture is worth a thousand words.  Shown below is the division action when the total divisor is set to 3 (sliders 1 and 2 On, all others Off):

upd8

How do I know what the duration of the output pulse will be?

That totally depends on the durations of the high and low portions of the original input signal.  Let’s say that, on the diagram above, the input signal high portion is 15 msec long and the input signal low portion is 25 msec long.  If we add those values to the diagram, it’s easy to calculate the duration of the output signal high and low portions:

upd9

You will note that the durations of the low and high parts of the output are not identical.  That’s because we used an odd divisor.  If you use an even divisor, the output low and high durations will be identical.  If you use an odd divisor, the output low and high durations will be close, but not identical.

You should be able to find some documentation about the flow meter output pulse timing on the flow meter datasheet.  Typically the high portion of the pulse is fairly constant and the duration of the spaces between the pulses changes as the media flow rate changes (faster flow = shorter spaces between the pulses).

The flow meter datasheet may specify a nominal pulse width and have a calculation to give the output frequency versus flow rate.  This is no problem, if we know the nominal pulse width and the frequency we can calculate the duration of the low part of the signal as follows:

Period of signal = 1/frequency

Duration of low part of signal = period of signal – duration of pulse

For example, say the nominal pulse with is 20 msec and the frequency at max flow rate is 20 Hz.

1/frequency = 1/20 Hz = 0.050 sec = 50 msec period

50 msec – 20 msec pulse width = 30 msec duration for low part of signal

So the overall approach to using the UPD is this:

  1. Use flow meter datasheet and anticipated maximum flow rate to figure the shortest duration pulses/spaces that should be coming from the flow meter.
  2. Use controller datasheet to find the shortest pulse/space that the controller is guaranteed to reliably detect.
  3. Draw out the flow meter pulse train on paper, including the durations of the pulses and spaces.
  4. Look at the drawing and start adding together the high and low durations of the meter signal until the sum exceeds the minimum pulse duration required by the controller.  If the sum should fall exactly on the minimum spec for the controller, add one more section from the picture as a safety margin.
  5. Count the number of up-and-down edges for the sections selected in step 4 ignoring the initial rising edge.
  6. Set the UPD divisor to the number of up-and-down edges counted in step 5.
  7. In your controller logic, multiply the raw count collected on the Binary Input times the UPD divisor to arrive at the true meter pulse count.

How about an example?

  1. The flow meter datasheet says the nominal pulse width is 15 msec and the calculated frequency at our max flow rate is 15 Hz.We figure the duration of the low part of the signal as:
    1/15 Hz = 0.067 sec = 67 msec period for the signal
    67 msec period – 15 msec pulse = 52 msec duration for low part of signal
  2. The controller datasheet says that the minimum pulse width/space that can be reliably detected is 100 msec.
  3. We draw a picture of our flow meter signal showing the 15 msec and 52 msec times:upd10
  4. We start adding high and low durations together until we exceed the 100 msec. minimum pulse detect time specified for the controller:upd11
    So we see that the output pulse duration will need to be a minimum of two high portions and two low portions of the input signal.
  5. Now count the number of up-and-down edges for the selected sections ignoring the initial rising edge:
    upd12
  6. In step 5 we counted 4 pulse edges (ignoring the initial rising edge) and so we set a divisor of 4 on our UPD by turning on slider #3.  And this is the final result of our efforts:
    upd13
  7. Keep in mind that every pulse counted by the controller input actually represents four pulses from the flow meter.  So to get the true flow meter total counts for your application program, you need to multiply the controller counts x 4.

What happens if you set the Divisor = 1?

You simply get a replica of the original input signal.  But remember, the output is electrically isolated from the input so you could use the UPD as a simple pulse signal 1-for-1 isolator.

What else should I know about the UPD-2?

The UPD-2 has an on-board 24VAC isolation transformer for its power supplies so you can get your 24VAC from any convenient source without worries about ground interactions.

Each UPD-2 contains two completely independent divider sections electrically isolated from each other.  One UPD-2 can serve two separate flow meter/controller setups or you can parallel the UPD-2 inputs from one flow meter and get two electrically isolated output pulses:

upd14

Conclusions

Flow meter output pulses are sometimes too narrow to be reliably detected on a controller’s binary input if the controller scan time is slow.  Inserting a UPD-2 Universal Pulse Divider between the flow meter and controller creates wider pulses that the controller can reliably count.  There is a well-defined method for deciding the correct divisor to use on the UPD-2.

The UPD-2 is dual-channel device which can serve two separate flow meter/controller setups or a single flow meter can drive both UPD inputs to create two separate electrically-isolated pulse output signals.

“Currently” Playing: Connecting A 4-20 mA Constant-Current Signal To Multiple Loads

One of the most popular and long-lived methods for transmitting analog control signals in the HVAC and industrial control worlds is the “constant-current signal loop.” In this scheme, the value of the current flowing in the circuit is the control variable, rather than any voltages that may appear at different points in the circuit. The constant-current loop is very robust and noise-tolerant and can run for long distances. Even in this high-tech digital age, it remains a very popular analog signaling method.

This article will deal with how to connect multiple loads to a constant-current signal loop. Connecting multiple loads to a constant-current signal loop is totally different than connecting multiple loads to a voltage signal source (which is covered in a previously-published article for those interested).

In the HVAC world, the most common range of current used for constant-current signaling is 4-20 mA (1 mA = 1 milliampere = 0.001 amperes of current). 4 mA represents 0% of the variable’s value and 20 mA represents 100% of the variable’s value. Other current ranges can be used, but today we’ll work exclusively with 4-20 mA current loops.

Current loop load “resistance” or “impedance”

Every load to be connected in the current loop will have an electrical characteristic called “resistance” or “impedance.” The load’s datasheet might use either term. Don’t worry about whether the datasheet says “resistance” or “impedance” we’ll take them to be the same thing. In this article we’ll use the term “impedance.”

Impedance is measured in units called “ohms.” If you don’t know what ohms are, don’t worry. All you have to do is use the numbers in some simple calculations later on.

Basic constant current loop configuration

Before we discuss connecting multiple loads to a constant current loop, let’s get familiar with the basic constant current loop containing a single load:

currently figure 1

On the left we have a constant-current source that is regulating the current to some desired value (the desired value is determined by a calculation in a controller, measured physical value in a sensor, etc.). The constant current leaves the source, flows through the load, and returns to the source.

Simple, right? What could go wrong? As usual, “the devil is in the details.”

Two different styles of current sources

There are two different styles of current sources you should be aware of. We’ll call the first style the “locally powered” current source. It has its own source of power for the 4-20 mA output and the only other item needed to complete the loop is the load:

currently figure 2

You will usually find this type of current source on the output of a powered controller.

We’ll call the second style the “loop powered” or “2-wire” current source. It does not have its own power source built-in, but rather depends on a separate power source (usually 24VDC) wired into the loop:

currently figure 3

You will find this type of current source in many 4-20 mA loop-powered temperature, pressure, and humidity transmitters.

Current sources have a “maximum load impedance” they can drive

Current sources like small values of load impedance, in fact they can drive their current straight into a short circuit (zero ohms load impedance)!

Current sources run into trouble when the load impedance gets too large. When the load impedance gets too large, the current source will not be able to make the higher values of current (near the 20 mA end) flow. The current value will plateau at some value less than the desired value.

How do I know the “maximum load impedance” my current source will support?

For a locally-powered current source, you will need to find the maximum load impedance value on the product’s datasheet.

For a loop-powered current source, you may find a parameter called “maximum load impedance (with 24V supply)” on the datasheet. If you do not, you will have to perform the following steps:

1. Find a parameter called “minimum operating voltage” (or some very similar wording) on the current source datasheet.

2. Calculate maximum load impedance as follows:

Maximum load impedance = (Loop power supply – minimum operating voltage) / 0.020

For example, suppose we are using a 24V loop power supply and we see on the datasheet that the 2-wire current source needs a minimum of 11V to operate. Then we would calculate:

Maximum load impedance = (24V – 11V) / 0.020 = 650 ohms

Now that we know something about how a single-load constant-current loop works, let’s see what kind of complications we get into when trying to drive multiple loads!

Connecting multiple loads to the current loop

Rule #1: All loads must be connected in series around the loop, never in parallel.

Below are sketches of the right way (series) and the wrong way (parallel) to connect loads in a current loop:

currently figure 4

currently figure 5

At first glance this requirement seems simple enough. That’s because we’ve drawn all the loads as 2-wire loads that don’t require power supplies. A not-so-obvious problem may occur if the loads do require power supplies. The current will always take the “path of least resistance” back to the current source whatever that path might be. If one of the loads before the last load has a power connection back to the current source, this connection will be the “path of least resistance” and the 4-20 mA will bypass the rest of the loads in the loop:

currently figure 6

If the powered load is a standalone module, you may be able to use a dedicated “floating” transformer or small DC power supply to power the load and resolve the issue:

currently figure 7

 

If multiple loads in the series connection are powered, each load would need its own power supply that is floating with respect to all the other supplies. For example, if Load 1 and Load 2 were both powered, Load 1 would need a power supply that is floating with respect to Load 2 and the current source, and Load 2 would need a power supply that is floating with respect to Load 1 and the current source. It can get messy quickly if there are several powered loads in the loop!

Consider also that If the powered load is a controller input and the controller is serving other inputs and outputs, it wouldn’t be practical to “float” the entire controller subsystem’s power connections.

In cases where a floating power supply for each load is not practical, it’s time to employ a signal isolator module such as the DT-13E from Kele:

currently figure 8

 

The DT-13E signal isolator has its inputs isolated from its outputs, and both inputs and outputs are isolated from the power terminals. So with the DT-13E installed, it just doesn’t matter where Load 2’s power source comes from, it will not corrupt the 4-20 mA current in the loop.

Rule #2: The sum of all the load impedances must be less than the maximum load Impedance the current source can support.

When the mA loads are wired in series (as they must be) the total load impedance seen by the current source is the sum of the individual load impedances. So if each of the three loads in the drawing below is 250 ohms, the total load impedance seen by the current source would be 250 + 250 + 250 = 750 ohms:

currently figure 9

 

Suppose your current source datasheet shows that it can only drive 650 ohms maximum impedance? Houston, we have a problem…

What can you do if your total load impedance adds up to more ohms than the current source can handle? This is another way that the DT-13E signal isolator can sometimes help. Many HVAC 4-20 mA loads have impedances of 250 to 500 ohms, but the DT-13E mA input impedance is only 125 ohms. So by isolating one or more loads using DT-13Es, not only do you remove grounding concerns but you may also lower your total loop impedance to a value the current source can handle:

currently figure 10

 

Now the current source only sees 250+125+250 = 625 ohms which is lower than its maximum load impedance, and Rule #2 is satisfied.

Conclusions

Multiple loads inserted in a 4-20 mA current loop must be connected in series. If the series-connected loads are powered devices, “sneak paths” between the load power supply connections and the current source may misdirect the 4-20 mA so that not all the loads receive the signal. Isolating the series-connected load with a DT-13E or similar signal isolator will solve this problem.

The sum of the series-connected load impedances may be too high for the current source to handle. The DT-13E signal isolator’s impedance of 125 ohms is less than that of many common HVAC loads. Isolating a load with the DT-13E may lower the total loop impedance to an acceptable value.

Shedding Some Light On 0-10V Dimmable Lighting Fixtures

 

A vdimmable lightingery popular way to decrease energy usage these days is to use dimmable lighting fixtures and throttle back on the electrical lighting when outdoor light is available through windows or skylights. A light sensor such as the Kele MK7 family can feed light level information into a building automation system (BAS). The BAS can then use an intelligent algorithm to vary the electrical lighting level with changing outdoor light levels to maintain a constant level of indoor illumination while saving energy.

In order for the BAS to command the dimmable lighting fixtures to the desired light level, some sort of control interface must exist between the BAS and the light fixtures. There are several types of light dimming systems out there in the world with different control interfaces.   The one we want to discuss today is known as the “0-10V current sinking” dimming system. We will also briefly mention several other types of light dimming systems, but they are not the focus of today’s article.

Classic Phase-Chopped High Voltage Light Dimmer System

The first light dimming system we’ll briefly touch on is the classic phase-chopping system. These dimmers connect in series with the high-voltage line to the lighting load and perform the dimming by removing part of each half-wave of the AC cycle:

dimmable lighting figure 1

This dimming system is typically limited to small-to-medium incandescent loads although some of the newer CFL and LED light bulbs will work with it also.  These dimmers are typically manual-adjust units without any control interface to a BAS.

Networked Digital Light Dimming Systems

DMX is a networked digital light dimming/control system used in theaters and at rock concerts.  DALI is a networked digital light dimming/control system that is popular in Europe and has found some use in the USA.

“0-10V Current-Sinking” Light Dimming System

This is the dimming system we want to discuss today.  It is formally defined in the standard IEC 60929 Annex E.

Although the interface is named “0-10V” it’s not like the 0-10V analog interfaces we are accustomed to in the HVAC world!  In the HVAC world the 0-10V is generated in the controller and is consumed by the load like this:

dimmable lighting figure 2

The classic 0-10V analog interface shown above is NOT the same as the 0-10V dimmable lighting interface!  The “0-10V Current Sinking” lighting interface is implemented as shown in the following diagram:

dimmable lighting figure 3

Wow, that’s quite a bit different than what we are used to!  The voltage source for the 0-10V signal is actually contained in the lighting fixture, not in the controller!

The voltage source is typically more than 10V, something in the 11-20V range.  A series resistor located inside the lighting fixture allows the light dimmer module to “pull down” the original voltage to the desired value.  The dimmer module does this by varying its own internal resistance until the desired voltage appears across its output terminals.  Those of you who have studied circuit theory will recognize the combination of light fixture resistance and dimmer module resistance as a classic “voltage divider” circuit.

You will notice that a small current flows around the loop from light fixture to dimmer module and back to the fixture.  The value of this small current is NOT the control signal, the voltage across the terminals is the control signal.  The small loop current is just a necessary evil to make the voltage divider circuit work as needed.

Hmmm… I’m getting the idea that a standard 0-10V output from a BAS controller may NOT work with a 0-10V dimmable lighting fixture.  Is that correct?

That is correct.  Your 0-10V BAS output might work with a dimmable lighting fixture if you are very lucky.  But probably, it won’t work.  If you’re unlucky, you might burn up the 0-10V output on your BAS controller.

So… I need a specialized dimmer control to drive these 0-10V lighting fixtures.  Where can I get such a dimmer control? 

We’re glad you asked. J  Kele sells the LDIM2 light dimmer module which is specifically designed to interface with 0-10V current-sinking dimmable lighting fixtures.  The LDIM2 can accept standard 0-10V or 2-10V or pulse-width input signals from your BAS controller and provide the necessary current-sinking 0-10V output for the light fixtures.  The 0-10V current-sinking output to the light fixtures is electrically isolated from the BAS signal inputs to prevent any interference between the two systems.

Can one LDIM2 dimmer module control multiple lighting fixtures?

Yes it can, just wire up the wire pairs from multiple lighting fixtures in parallel like this:

dimmable lighting figure 4

The total current flow through the LDIM2 output will be the sum of all the individual lighting fixture currents.  Different makes and models of fixtures may supply different current values.

How many lighting fixtures can I attach to the LDIM2 output?

That depends on the control current flow from each lighting fixture.  The maximum load current allowed on the LDIM2 output is 0.5 amps.  So you can add lighting fixtures until the total from all the fixtures reaches 0.5 amps, but you can’t go further.  If, for example, each fixture supplied 1 mA of current, you could attach 500 fixtures to one LDIM2 (0.5 amps / 0.001 amps = 500).

If you have so many lighting fixtures that the total control current exceeds 0.5 amps, wire them up in “banks” where each bank is 0.5 amps or less and is controlled by its own LDIM2 dimmer.

How do I find out how much current a particular model lighting fixture puts through the LDIM2?

The IEC 60929 Annex E standard specifies that the control current value should be between 10 uA (microamps) and 2 mA (milliamps).  However, there’s absolutely no guarantee that the lighting fixture manufacturer adhered to these guidelines.

If you’re really lucky, maybe the lighting fixture data sheet will tell you the value of the control current.  If you cannot find a published value for the control current, please don’t just assume a value.  Also don’t mistake the lighting fixture’s supply current for the fixture’s control current.  The fixture’s supply current will almost always be on the data sheet, but will be a much higher value, possibly several amps.

If you have access to the physical light fixture(s), you can measure the control current with your DC mA meter.  Just put it across the two signal wires coming down from the fixture(s).  But beware, the mA meter resistance is less than 1 ohm.  It will pull the voltage down very close to zero volts, and the lights will go dark, so don’t do this during work hours on an occupied space unless the people are warned first!

What happens if the fixture wires are connected to the LDIM2 with the polarity reversed?

If the lighting fixture wire polarity is hooked up backwards, the voltage will go to about 0.7V which is near 0% light level.  Nothing will be damaged, but the lights will go out.

How can I test the LDIM2 on my workbench if I don’t have a dimmable lighting fixture available? 

You can use a standard 24VDC supply and a pull-up resistor like this:

dimmable lighting figure 6

The catalog description of the LDIM2 is “fluorescent dimming control.”  Will it work with dimmable LED lighting fixtures?

Yes, it will work with any dimmable lighting fixture that uses the 0-10V current-sinking interface.  You just need to figure out what control current the fixture puts through the LDIM2’s output so you don’t overload it by attaching too many fixtures.

Conclusions

The 0-10V current-sinking interface used by dimmable lighting fixtures is not compatible with the standard 0-10V outputs used in HVAC/BAS systems.  You should use a specially-designed dimmer control module such as Kele’s LDIM2 for dimmable lighting fixture applications.

Networking 101: BAS Network Cabling

The various devices comprising a building automation network are either connected by cables or talk wirelessly to each other, or possibly a combination of both.  In Networking 101, we are going to concentrate on network devices connected by cables.  Let’s take a look at the different types of network cables typically encountered in building automation networks.

Network Cables 

The most common type of network cable used in building automation networks is the “twisted pair” cable.  Two wires are twisted around and around each other at a uniform number of twists per inch to form the “twisted pair”: Screen Shot 2014-09-24 at 2.19.53 PM Why twisted pair cable?  Without going into a lot of technical detail, let’s just say that with the proper type of transmitters and receivers using twisted pair cable greatly reduces noise pickup from noise sources such as motors, fluorescent lights, and radio stations.  This results in more reliable communications with less corrupted data and retransmissions.

Not all twisted pair cables are created equal.  There are electrical characteristics such as resistance and capacitance per foot that vary between different brands and models of twisted pair cable.  Generally the lower the resistance and capacitance of twisted pair cable, the better it performs in networks.

It’s a good idea to select a network cable that’s specifically recommended by the manufacturer for use in your type of network.  If network-specific cable is not available, substitutions can be made.  If the resistance and capacitance per foot of the proposed substitute cable is as low as or lower than the recommended cable, the substitute cable should be acceptable.

Some styles of network cable are available with or without a metallic shield around the outside of the wires.  Shielded cable generally helps with reducing noise pickup, but the cable will cost more.

Now let’s get more specific about the types of physical layers typically found in building automation networks…

RS-485  

RS-485 is used extensively in building automation networks.  Modbus RTU, JCI Metasys N2, and BACnet MSTP protocols all use RS-485.

RS-485 uses a single twisted-pair to send the data.  It also uses a third “reference wire” in addition to the twisted pair.  There is debate among users as to whether the reference wire is always needed, and some devices do not include a terminal for connecting a reference wire.  But to be safe it’s always best to include it in the cable: Screen Shot 2014-09-24 at 2.22.13 PM RS-485 devices typically use screw terminals for connecting to the cable.

The two wires of the twisted-pair do have a polarity assigned to them.  The same wire of the pair must be connected to the RS-485 “+” terminal of all the devices on the network.  If an RS-485 device is connected to the network with the “+” and “-“ wires attached backwards, it will not talk on the network!

RS-485 cable should be run as one continuous “trunk line.”  The RS-485 standard specifies that as many as 32 devices can share the same trunk line, and the line can be as long as 4000 feet.

It is bad practice to “tee” into the middle of an RS-485 cable and run a “stub” line off to another RS-485 device:

3

Instead, the RS-485 line should be daisy-chained to each device on the line like this: 4

If an RS-485 cable needs to be longer than 4000 feet a “repeater” module can be inserted in the middle of the cable run.  The repeater will refresh the electrical signals and allow for longer runs.

You can also expand an RS-485 network to more than 32 devices by creating multiple sub-networks, each with its own cable, and connecting the sub-networks together with a “bridge” or “router” device.  A bridge or router will have two or more separate RS-485 electrical ports.  Each port on the bridge/router connects into one of the sub-networks and data can then flow between the sub-networks:

5 A “Bridge” forwards all the traffic on either sub-network across to the other sub-network.

A “Router” has the capability to learn (or to be told through a user-generated lookup table) which devices reside on each sub-network.  Now if two devices on the same sub-network are talking, the conversation is not passed to the other sub-network.  This eases overall network traffic congestion.

FT-10 “Free Topology” (Lonworks)  

The FT-10 “Free Topology” physical layer was invented by Echelon Corporation as part of their Lonworks networking technology.

Similar to RS-485, FT-10 uses a single twisted pair to carry the data, but FT-10 does not use a third  “reference wire”:

6 The FT-10 twisted pair has no polarity – the wires can be attached to each Lonworks device without worrying about which wire is “+” and which is “-“.

Like RS-485 devices, Lonworks FT-10 devices also typically use screw terminals to connect to the network cable.

FT-10 is much more relaxed as to how the cables can be arranged – you can pretty much wire different FT-10 network segments together any way you want to.  You can “tee” into an FT-10 line with a “stub” headed in another direction, or you can make a “star” with a center point and multiple lines radiating out from it:

7 Lonworks FT-10 documentation recommends that the total cable run be no longer than 1640 feet and you put no more than 64 devices on a network.

As was the case with RS-485, you can extend the FT-10 cable distance with an FT-10 repeater.  You can also extend beyond 64 devices by creating sub-networks and connecting them together with an FT-10 bridge or router.

Please note that even though RS-485 and FT-10 both use a single twisted pair to carry the data signals, they are totally incompatible electrically!  You cannot ever tie an RS-485 network cable directly to an FT-10 network cable.  

If you want to communicate between an RS-485 network and an FT-10 network, you must use a “Gateway” device which has an RS-485 port on one side and an FT-10 port on the other side:

8 Ethernet  

Ethernet was invented back in the 1970s by Xerox and is extremely popular in the Information Technology world.  BACnet IP and Modbus TCP protocols use the Ethernet physical layer.

There are several different versions of “Ethernet” networks, but the two we deal with in building automation are the 10BASE-T and 100BASE-T versions.  The only significant difference is the speeds.  10BASE-T runs at 10M (10 million bits per second).  100BASE-T runs at 100M (100 million bits per second).

10BASE-T and 100BASE-T devices can be mixed on a network.  If a 100BASE-T device needs to talk to a 10BASE-T device, it simply drops its speed to the 10BASE-T rate temporarily.

The two styles of Ethernet cables you’re likely to encounter in building automation are designated as CAT5 (Category 5) and CAT5e (Category 5 Enhanced).  Both cable styles can be used with either 10BASE-T or 100BASE-T, but CAT5e cable gives superior performance.

The 10/100BASE-T cable contains four twisted pairs, but only two of them are used to carry the data: 9 The other two twisted pairs may be left unused or used for some other purpose.

Standard length off-the-shelf 10/100BASE-T cables typically come with RJ-45 plugs pre-installed on the ends of the cable.  RJ-45 plugs look like modular telephone plugs except they are wider and have positions for 8 connections.  If you need to make a custom length 10/100BASE-T cable, you can buy a bag of loose RJ-45 plugs and a special wire stripper/crimper tool and attach RJ-45 plugs yourself (but it is a bit tedious to do).

Unlike RS-485 and FT-10, you can only connect two devices with a 10/100BASE-T cable!  And the recommended maximum cable length is only 330 feet.  So how are we able to network many devices together?  The answer is in the use of Ethernet cable-sharing devices such as “hubs” and “switches.”

These hubs and switches have multiple RJ-45 data jacks for plugging in Ethernet cables.  They are used to build out an Ethernet network “like a tree”:

10 An Ethernet ‘Hub” forwards everything it receives on any port out all the other ports on the hub.  So if an Ethernet network is built using all hubs, every end device on the network must listen to every conversation on the network whether it’s involved or not.  This creates a lot of unnecessary traffic and slows down the operation of the network.

An Ethernet “Switch” learns where the different end-devices are located in the network and restricts the traffic flow to only those network segments that are needed to reach the specified end device.  This greatly reduces unnecessary traffic on the network and speeds up response time.

An Ethernet “Router” is used to connect multiple Local Area Networks (LANs) together to form a Wide Area Network (WAN).  There are also Ethernet “Gateways” to connect Ethernet networks to RS-485 or FT-10 networks.

RS-232  

RS-232 is a physical communications layer that is used for making point-to-point connections between 2 pieces of equipment.

You may find that some network devices include an RS-232 port for connecting to a computer to perform device configuration.  This RS-232 port is not for the network connection, only for configuring the device prior to using it on a network.

RS-232 ports on older equipment used 25-pin connectors like this: 11 Modern equipment providing RS-232 ports use 9-pin connectors like this: 12 The connectors come in both male and female versions, so when connecting two pieces of equipment you must the sure that the cable has the correct gender connectors.

A “normal” RS-232 cable will have a male connector on one end and a female connector on the other end with “straight-through” wiring (pin 1 connects to pin 1, pin 2 connects to pin 2, etc.).  But sometimes a custom cable must be purchased or fabricated with male-male or female-female connectors and possibly with crossed wiring between the pins.  The details of these special RS-232 cables are beyond Networking 101.

Connecting A Voltage-Output Signal Source To Multiple Loads

The Tech Support crew here at Kele frequently gets quizzed by customers who have a voltage-output signal source which must drive multiple voltage-input loads. Requests for “voltage signal replicators” come in frequently. Sometimes extra hardware is needed, but sometimes it isn’t! So we thought it might be a good idea to write an article addressing this topic. Here we go…

*** A note about the terms “resistance” and “impedance”

When looking at datasheets for HVAC control modules, you may see the terms “resistance” or “impedance” used in describing signal input/output loading characteristics.  Strictly speaking, “resistance” and “impedance” are different entities (impedance = resistance + reactance) but for a slow-moving control signal which is what we usually have in the HVAC world, the two terms can be used interchangeably.  So don’t worry about whether the datasheet says “resistance” or “impedance” we’ll take them to be the same thing.  In this article we’ll use the term “impedance” because it’s one letter shorter than “resistance” and we’ll save keystrokes! J

Impedance is measured in units called “ohms.”  If you don’t know what ohms are, don’t worry.  All you have to do is poke the numbers into some simple calculations later on.  You may see a suffix “k” or “K” appended after the ohms value on a datasheet.  This is important to note!  The “k” or “K” means “times 1000” so an impedance of “100K” means 100,000 ohms and that’s the number you would put into your calculations.

On with the show…

The first thing we must do is answer a few questions about the signal source and loads:

  1. Do the Signal Commons on the different loads have to be isolated from each other?
  2. Does the Signal Common on the voltage source have to be isolated from all loads?
  3. What is the “minimum load impedance” the signal source can drive (datasheet item)? (Alternately, this could be specified as “maximum output current”).
  4. What is the “input impedance” of each of the voltage-input loads (datasheet item)?

The first two questions are sometimes difficult to answer, and are application-dependent.  If you cannot find concrete answers, you can always assume that every device must have isolated signals.  This approach will always work, but it will also cost more money as signal isolator hardware is required.

With regards to question #3, the signal source datasheet may specify a “maximum output current” value instead of “minimum load impedance.”  The “maximum output current” will be expressed in units of mA (milliamps).  No worries, we can calculate what we need by this equation:

Minimum load impedance = Largest output voltage required/maximum output current

As an example, suppose the signal range of interest is 0-10V and the maximum output current available from the signal source is 2 mA (0.002 amps):

Minimum load impedance = 10V / 0.002 amps = 5000 ohms

Calculating “equivalent load impedance”

When connecting voltage loads in parallel, you’ll need to calculate the “equivalent load impedance” of all the loads connected together.  There are a couple of ways to do this:

If all load impedances are the same—

Equivalent load impedance = impedance of one load / number of loads

If the load impedances are different –

Equivalent load impedance = 1 / (1/Load1 impedance + 1/Load2 impedance + …)

The second calculation is most easily done using the 1/X key on your calculator.  Just do a 1/X calculation for each load impedance and add the terms together.  Then do a final 1/X calculation on the first result.  Here’s an example—

Load 1 = 10,000 ohms (10K)

Load 2 = 50,000 ohms (50K)

Load 3 = 100,000 ohms (100K)

Equivalent load impedance = 1 / (1/10000 + 1/50000 + 1/100000)  =  7,692 ohms

The two rules that must be satisfied

There are basically two rules that must be satisfied when driving multiple loads from a voltage source:

  1. If electrical isolation is required, you must add a hardware signal isolator module to each device to be isolated.
  2. At each voltage output-to-input interface, the equivalent load impedance of the loads connected together must be higher than the minimum load impedance of the signal source

Case 1:  It’s OK for all devices to have their Signal Commons tied together 

This case can be done using direct wiring (NO extra hardware) if the equivalent load impedance (calculated above) is greater than the signal source’s minimum load impedance:

direct connect method

That’s great UNLESS… your equivalent load impedance is smaller than the voltage source’s minimum load impedance.  Then what are you going to do?

When the equivalent load impedance is too small for the signal source to drive directly, you will need to insert a “signal booster” device of some sort between the original signal source and the loads.  The signal booster doesn’t need to boost the voltage (you typically want the same voltage out that’s coming in).  It needs to boost the available current so there is enough to drive all the loads connected together.

The Kele UAT is a good device to use as a signal booster in this application.  A look at the datasheet shows that its voltage output can drive up to 20 mA of current.  Let’s assume that the signal range of interest is 0-10V so the maximum voltage we ever need to output is 10V.  Then the minimum load resistance supported by the UAT would be:

UAT minimum load resistance = 10V / 0.020 amps = 500 ohms.

This is plenty of drive power for most parallel-connected voltage loads.

Now when we insert a signal booster device, we must insure that the input impedance of the booster itself is not too low for the original signal source.  We see that the UAT input impedance on the 0-10.9V range is 156K ohms (156,000 ohms).  If your UAT datasheet shows 156 ohms input impedance, don’t panic!  It’s really 156K, some datasheets have a typo.  Our apologies.  So the UAT input impedance of 156K is far higher than the signal source’s minimum load impedance of 5K, and it’s all good:

signal booster method

Now you should understand that the UAT does not provide signal isolation as it has one Common terminal for both signal input and output.  If you require isolation between devices, read on…

Case 2: Load Signal Commons can be tied together, but voltage source needs isolation

In this case you will need a signal isolator device such as the Kele DT-13E.  On the DT-13E, the input signal terminals are completely isolated from the signal output terminals and both of those terminal sets are completely isolated from the power terminals.

We look at the DT-13E datasheet and see that the input impedance on the 10V input range is 13.3K.  This is higher than the signal source’s minimum load impedance of 5K, so there is no problem driving the DT-13E input from the original signal source.

Looking at the DT-13E voltage output, we see that the maximum current available is 6 mA.  For a 0-10V signal, the max voltage we need to drive out is 10V so the DT-13E’s minimum load impedance will be:

DT-13E minimum load impedance = 10V / 0.006 amps = 1667 ohms.

If we have the same three 10K loads paralleled as in the previous example (3333 ohms equivalent load) then the DT-13E has enough drive and this setup should work just fine:

isolated signal source

Case 3: Load Signal Commons must be isolated from each other

In this situation, each load will need its own signal isolator module like the DT-13E.  The most obvious setup to use is shown in the next figure.  When we do the load calculations, however, we find that three DT-13E inputs in parallel will present an equivalent load impedance of 13.3K / 3 = 4433 ohms.  Oops, that is lower than the minimum load impedance of 5000 ohms specified for the signal source:

individually isolated loads

How do we fix this problem?  Here are some possibilities:

  1. Use a different signal source with more drive capability.
  2. Use different signal isolators with higher input impedance.
  3. Try rearranging the wiring using the devices that we have.

Let’s see what we can do with option #3.  Realizing that inputs and outputs are isolated on the DT-13E modules, it’s perfectly legal to take one of the DT-13E inputs and drive it from another DT-13E’s output.  Let’s check the math to see if this would work.

With only two DT-13E inputs paralleled on the signal source, the equivalent output impedance would be 13.3K / 2 = 6650 ohms.  This is higher than the signal source’s minimum load requirement of 5000 ohms, so that works.

With the new arrangement, one DT-13E output will drive a parallel combination of the original 10K load and one DT-13E input (13.3K).  The equivalent load impedance will be:

Equivalent load impedance = 1 / (1/10000 + 1/13300) = 5708 ohms.

This is higher than the DT-13E output’s minimum load impedance of 1667 ohms so that works too!  And this is the final configuration:

idividually isolated 2

Conclusions

By following the two interface rules discussed above, multiple voltage-input loads may be successfully driven by a common voltage-output signal source.  The math is simple algebra and can be done on any calculator.  Happy interfacing!