Pneumatic Controls, First Cost Advantages, and Retrocommissioning Opportunities

Recently, a reader posted a question in response to my post titled Retrocommissioning Findings: Make Up Air Handling System Simultaneous Heating and Cooling – The Clues – #2 – The Controls May be Pneumatic.  Specifically, the question was:

What are cost advantages if any, under energy codes such as ASHRAE 90.1-2010 and IECC 2012 to installing a new pneumatic control system in a New School ?

That’s an interesting question and I can’t say I had ever really considered pneumatic controls in the context of energy codes and cost advantages.  So what follows is a sort of “stream of conscious” response. 

Answering “What Time Is It?” by Explaining the History of Timekeeping

It turns out that even thought the question seems fairly simple and straight-forward, there are actually a lot of things that need to be considered if you really want to make the best decision.  So, as I am prone to do, I have answered a question about the cost advantages of pneumatic controls by going into the history of pneumatic controls, or at least a bit of the theory. 

The post does end up including a lot of basic pneumatic control information (and pictures of same)  if you are interested in that.  So maybe that is a plus if my longevity drives you crazy, as I know it does some people.

Basically, the more I wrote, the more I realized I needed to explain some nuances and details to be sure you understood the points I was trying to make.  I suspect this was partly because I just did a lab class on pneumatic actuation last week, so all of the questions that were asked there were fresh in my mind, causing me to bring them into this discussion since they seemed relevant.

To help with those who wish I could be more brief, I have used headings to separate topics and (hopefully) make it easy for you to find a particular point of interest.  I suspect there is a way to hyperlink inside the post, but I have not figured out how to do that, so for the time being you will have to page down and look for the heading of interest.

The topics include the following, which are the headings for the various sections in the order they occur.

Taking a Closer Look at a Pneumatic Actuator

  • A Typical Pneumatic Actuator
  • Spring Ranges
  • Sequencing HVAC Processes via Spring Ranges
  • Another Force in the Balance of Forces
  • Sequencing Things at the Central Systems Level
  • Why This Matters if you are Considering a New Pneumatic Control System

Pneumatic Actuation vs. Pneumatic Control

  • Pneumatically Actuated DCC Controls at the Central Plant and Systems
  • Pneumatically Actuated DCC Controls at the Zone Control Level
  • The Cost Implications of Using Pneumatic Actuation in a New Project
  • Deciding Which Way to Go

The Energy Implications of Pneumatic Control Air Piping Leaks

  • Excessive Compressor Run Time
  • Potential Simultaneous Heating and Cooling
  • Why This Matters if you are Considering a New Pneumatic Control System
  • The Cost Implications of Installing a Leak Resistant Piping Network

Implementing Complex Strategies

  • Ease of Implementation
  • Flexibility
  • Gene Pool Issues
  • Persistence
  • The Cost Implications of Implementing Complex Strategies with Pneumatic Controls

Since there are both first cost and ongoing (life cycle) cost implications to using pneumatic controls I will discuss each of them in the context of the various topics I bring up.  This is important because frequently, decisions made on a first cost basis are short sighted in terms of life cycle costs, especially if you are talking about substituting pneumatic controls for DDC controls.  That’s not a good thing financially most of the time.  And I’m pretty sure that its not a good thing holistically for us or the earth all of the time.

Substituting pneumatic zone controls for DDC zone controls is a good example of this.  Typically, you can install the pneumatic zone controls for less money than DDC controls, especially if you use one pipe thermostats.  But, using pneumatic zone controls will tend to increase ongoing annual maintenance costs, reduce the Owner’s ability to do diagnostics and react to occupant concerns in a timely manner, and reduce the potential for zone control settings and optimized zone control strategies to persist.  You also give up having the ability to easily optimize the control strategies in the central systems based on what is going on at the zone level.

Ultimately, it may mean that at some point in the future, you scrap the whole thing and upgrade to DDC.  Meaning you throw away the resource represented by the initial investment and consume more resources trying to achieve what you originally set out to do. In the interim, the system wasted resources that could have been saved had the life cycle perspective been taken in the first place.

Taking a Closer Look at a Pneumatic Actuator

For some of the things I am going to discuss to make sense, I need to be sure we are on the same page about how a pneumatic actuator works and interacts with the control system and the field conditions.  So,  I will start with a discussion about that topic and how pneumatic control systems use actuator spring ranges to create a control sequence.

A Typical Pneumatic Actuator

Most pneumatic actuators, especially the ones out on zone devices like VAV terminal damper and reheat coils, are piston actuators. Here is a picture of a piston actuator on a replacement reheat valve from a recent project.

Pneumatic valve 01

If you take the top off, it looks like this.

Pneumatic valve 04

The black rubbery looking thing inside the valve cap on the left is the diaphragm. Air pressure on the top of it pushes on the piston, which is the silver cap shaped part sticking up out of the valve body on the right. The red spring pushes back against the piston.

The little plastic “button” in the foreground fits through a hole in the cap, into the hole you see in front of the valve spring and holds the cap on. If it breaks, then the air pressure cause the cap to push itself off of the valve instead of opening the valve; so there is one potential thing that could go wrong.

If you look at the end of the valve …

Pneumatic valve 02

… you can’t quite see the valve plug because there is no pressure applied to the piston. The plug is up inside the cavity towards the top of the valve in this picture, resting against a stop.

If you took the valve plug out of the valve body, it would look like this.

New valve plug

Here is a slide from a class with a bit more information about how it all works together.

Slide Show 2122013 120324 PM

So, this valve would be considered “normally open” meaning it takes air pressure to close it. When I use my hand to take the place of air pressure in this next picture, and squeeze down on the piston, you can see the valve plug move towards the valve seat.

Pneumatic valve 03

Spring Ranges

Out of the box, the balance of forces on the actuator assembly is such that the piston tries to move the valve plug one way while the air pressure tries to move the plug the other way. If you look closely at the nameplate on this valve, you will notice that it talks about spring ranges.

Valve Nameplate

A spring range of 2-5 psi means that out of the box, the air pressure would have to change from 2 psi to 5 psi to cause the valve to go full stroke, in this case from fully (normally) open (the position with no air pressure) to fully closed. Manufacturing tolerances will come into play here, meaning that its not exactly 2 psi to 5 psi; rather, its about 2 psi to about 5 pse.

If you look closely, you will notice that the valve also can have a 3-10 psi spring range. This is accomplished by shipping the valve to the field with two springs and then allowing the installer to select the correct one; green for 2-5 psi (you can see it installed in the actuator in the close-up of the valve plug about 4 pictures back) and orange for 2-5 psi (you can see it in the plastic bag behind the actuator in the same picture and in the picture above). So, there is something else that could go wrong; the installer may not realize that there is a choice and install the valve with the wrong spring range.

Sequencing HVAC Processes via Spring Ranges

Pneumatic control systems use spring ranges to sequence multiple actuators. For instance, a pneumatic VAV terminal flow controller may be calibrated and set up to go from maximum flow to minimum flow as the signal from the thermostat controlling it drops from 15 psi to 10 psi.

By using a normally open reheat valve with a spring range of 10 psi to 3 psi, you could sequence the valve with the controller so that a single thermostat could control them both, relying on the fact that the valve would not start to open until the pressure signal from the thermostat dropped to 10 psi (the same pressure that should reset the flow to its minimum) to prevent unnecessary simultaneous heating and cooling. Thus in a perfect world, you would not have reheat until you were on minimum flow.

Slide Show 2132013 75326 AM

Another Force in the Balance of Forces

If you go back to your basic physics class, you may remember a thing called a free body diagram that was all about balance of forces. In general terms, the piston and the spring in a pneumatic actuator are a real world example of a balance of forces.

Slide Show 2122013 33104 PM

If the force produced by the air pressure exceeds the force produced by the spring, the piston moves towards the spring until the force generated by the spring being compressed is exactly equal and opposite to it, at which point the system comes to rest. Reduce the air pressure and the balance shifts the other way.

But, in a real valve in a real piping system with flow passing through it, the pressure of the water acting on the cross-sectional area represented by the valve plug also generates a force. Air flowing across a damper can do the same thing to an actuator serving it.

Slide Show 2122013 33116 PM

Depending on how the actuator is built, this pressure works either with or against the piston or spring and shifts the effective spring range of the actuator. If you go back to the VAV terminal with the normally open control valve with a 3-10 psig spring range, if the pump pressure were to shift the spring range 2 psig, a problem crops up.

Slide Show 2132013 75329 AM

Specifically, the reheat valve starts to open before the terminal unit has reached minimum flow and the system does unnecessary simultaneous heating and cooling.

One way to solve that problem is to pick the control valve spring range so there is a significant gap between the two operations that are to be sequenced. For instance, if we replaced the 3-10 psig spring in our valve with the 2-5 psig spring, the sequencing would once again be as desired, even with the shift to a 4-7 psig range created by the added pressure from the water on the valve plug.

Slide Show 2132013 75333 AM

Sequencing Things at the Central Systems Level

Pneumatic control systems also use spring ranges to sequence things at the central systems level.  For instance, you could sequence the preheat valve (a heating process) with the economizer and chilled water valve (cooling processes) to maintain a discharge temperature requirement by appropriate selections of spring range and normal valve and damper position (normally open vs. normally closed).

Slide Show 2132013 80628 AM

And, as you might imagine, all of the issues I discuss in the preceding sections can come up.  In fact they are more likely since the valves and dampers are larger and thus, the forces generated by their interaction with the process flows are larger.

Adding to the complexity is the fact that if you are sequencing multiple elements, you run out of options in terms of spring ranges that have some spread between them.  A common solution to that issue as well as other issues is to apply a positive positioner a.k.a a positioning relay a.k.a. a positioner. 

Basically, this is a device that sits in-between the control signal and the actuator.  It has a source of air supplied to it and knows if the actuator moves and how much.  Its basic function is to look at the control signal and then use its main air supply to do what ever it takes to move the actuator as much as required.

Positioners also tend to make the actuator respond more quickly, essentially eliminating lag and dead time from the control process, which is a good thing.  David St. Clair has a good discussion of lags and why minimizing them is important in his book if you are interested.

Here is a picture of the positioner on the lab demonstrator I was using last week with a bit of info about what it does.

Slide Show 2132013 95349 AM

I  could go into more detail but that would make this post even longer than it already is.  So I will save that for a different post or provided it sooner if someone asks for more information.

Why This Matters if you are Considering a New Pneumatic Control System

It may seem like I have gone “way out in the weeds” as they say in the context of the original question.  And I will be the first to admit I can do that.  But I think the information matters for a  number of reasons.

The bottom line reason is that (I think) its important to understand the pros and cons of the technology you are considering so you can make an informed decision.  Many of the field problems we see as commissioning providers come back to decisions that were made with out fully understanding the details and implications associated with them.  And the results can be annoying (like the chiller nuisance head pressure trips I mention in the next section) or somewhat alarming (like the bent elevator door frames I also mention in the next section).

In terms of relevant, real-world, new construction issues relate to what I have discussed so far, here are a few examples.

DDC Controls with Pneumatic Actuation

Pneumatic actuators are still quite viable, desirable even, for current technology systems.  This is the topic of the next session so look there for more details.  But just because we are moving away from pneumatic controls does not mean we are (or should) move away from pneumatic actuators.

Pneumatic Controls are Still Offered as “Value Engineering” Alternatives

As mentioned in the introduction, people are still “saving money” by installing pneumatic zone controls instead of DDC controllers.  This is usually a first cost benefit at the captured at the expense of a system/building life-time cost penalty.  By understanding the details, you will be in a better position to argue in favor of the better life-cycle cost solutions when challenged by a value engineering recommendation.

Renovations and Tennant Improvement Project Replicate Existing Conditions

In a similar vein to the preceding, it is not unusual for a renovation or tenant improvement project to simply replicate the current approach.  Sometimes, this is the best way to go.  But other times, it is a lost opportunity to leverage budgets to improve things.  By understanding the pros and cons of the various options, you will be well positioned to recognize an opportunity when you see it and take advantage of it.

Pneumatic Controls are Often Upgraded to DDC using a “Replace the Existing Pneumatics with DDC” Approach

It is not uncommon for a control system upgrade to have limited definition of scope, especially if it is being handled in-house by the operating team in concert with a vendor.  Frequently, the agreement is that the vendor will simply replace the existing pneumatic panels with DDC panels and re-use the existing actuating system to control costs. 

There is nothing inherently wrong with this approach, but it is important to realize that simply substituting a DDC panel for a pneumatic panel will not guarantee all the benefits are achieved.  An amusing but occasionally accurate way to think of it is that with out some engineering that considers the entire system, the microprocessor in the DDC controller will simply allow you to do stupid things much faster and with greater precision that the old pneumatic controller it replaced.

For instance, if the existing control system used one output and spring ranges to sequence and HVAC process, and the process loads are shifting the spring ranges so that simultaneous heating and cooling is occurring, then the same thing will happen with the DDC system unless steps like providing independent outputs or adding positioning systems to the actuators are not included in the upgrade project to address the problem.

Value Engineering Eliminates Independent Outputs

It is not unheard of for a value engineering process to save first cost by proposing that sequencing be accomplished with a common output serving actuators with appropriately selected spring ranges instead of providing independent outputs for each actuator, even though the facility is a new building with a new DDC system.  Again, there is nothing inherently wrong with this in that it can be made to work.

But you need to recognize the limitations up front and accommodate them as a part of the value engineering.  Meaning that you probably want to be sure that positioning relays are provided for each of the actuators that will be sequenced (which adds some first cost and operating and maintenance cost back into the equation and may make the value engineering proposal less attractive).

And, if you become involved in such a discussion, you may want to point out that part of what you buy with a DDC system is nearly infinite flexibility in terms of making changes or modifications via typing instead of running conduit, wire, and tubing.  Providing independent outputs maximizes that flexibility, so if you agree to use a common output and sequence things in the field hardware, you are giving up some of the benefits of the technology you are employing.

Pneumatic Actuation vs. Pneumatic Control

I want to re-emphasis the point I made early in the referenced blog post about pneumatic control vs. pneumatic actuation.  For the purposes of my discussion, pneumatic control implies a control process that is powered by pneumatic air and uses analog controllers with lots of levers, diaphragms and orifices to control pneumatic actuators.  This is in contrast with what I term pneumatic actuation, where the control process may or may not use pneumatic air for working out the logic and sequence, but ultimately moves the final control elements with pneumatic air pressure.  Frequently, DDC control systems are combined with pneumatic actuation to gain the benefits of both worlds.

Pneumatically Actuated DCC Controls at the Central Plant and Systems

There is a lot to be said for pneumatic actuation coupled with a computer based DDC system running the logic and measuring the parameters that are inputs.  As a designer and operator, I actually have a preference for it, especially at the central plant and central systems level because of its reliability, simplicity, speed and power. 

Most current technology pneumatic actuators can be provided with a positioning system that not only performs the positioning function but also converts the electronic DDC system signal to a pneumatic signal, as illustrated in these photos from a recent job site.

2-way Butterfly Valve on Pallet 02

These are pneumatically actuated butterfly valves with a double acting piston type actuator.  Here is a cut away of the actuator, courtesy Bray International.


The actuator can be equipped with a pneumatic positioner, like the Valve Accessories and Controls V200P in the picture …

2-way Butterfly Valve on Pallet 03

… or it can be equipped with a positioner that also converts the electronic signal from a DDC system to a pneumatic signal in addition to providing the positioning function like this one from Bray.


Pneumatically Actuated DCC Controls at the Zone Control Level

Once you get out to the zone level (terminal equipment like variable volume boxes, constant volume regulators, fan coil units, etc.), my preference switches to purely electric/electronic control systems.  For one thing, at the terminal equipment, the actuating loads are small and the thermal inertia of the zone and elements it contains generally means things tend to not change quickly ( with “quickly” meaning “in seconds” like they can back at the AHU or central heating or cooling plant). 

That means the speed and power I gain by using pneumatic actuators in the central plant may not be warranted at the zone level.  By using an all electric system at the zones, I avoid one of the big pitfalls of pneumatic actuation, that being air leaks (see the section that follows with that title for more on the issues).

The bottom line is that it’s one thing to keep the pneumatic piping in a central plant leak free, where it is:

  • Generally exposed and accessible
  • Could be soldered copper piping (vs. poly tubing) if you wrote the specs that way, and
  • Is and confined to a relatively small area (say 20,000 to 50,000 sq.ft. for the central plant and equipment rooms serving a 600,000 sq.ft. facility). 

It is quite another thing to keep the pneumatic piping serving the remaining 550,000  – 580,000 sq.ft. of facility leak free when it is:

  • Concealed behind ceilings and walls, which may or may not be accessible, and
  • Run with poly tubing to control costs.

I’ve dedicated a separate section titled The Energy Implications of Pneumatic Control Air Piping Leaks to discuss why leakage is an issue.  For the remainder of this section, I will focus on the cost implications of using pneumatic actuation.

The First Cost Implications of Using Pneumatic Actuation in a New Project

Whether or not there is a cost advantage to using pneumatic actuation for the control system for a new project will likely depend on a number of factors.

Expanding an Existing Control Air Source

One of the obvious costs associated with using pneumatic actuation is that you need a control air system, which typically includes compressors, a storage tank, an air dryer or two, and a piping network with associated pressure reducing stations, valves, and fittings.  So, if the new project expands an existing facility and the existing control air system has reserve capacity, then it is likely that there will be a cost advantage associated with using pneumatic actuators since for a given amount of torque, they are generally less costly and all that will be required to support them will be to extend the control air piping to the required location.

If large actuators with position controllers using high pressure air are to be supported, the piping network may need to include a low pressure (under 30 psig) main and a high pressure main (typically 80 psig or what ever pressure the position controllers have been selected to work with).   If the system does not include large valves with position controllers and/or the position controllers have been selected to work with 30 psig air, then only one piping main will be required.

Installing a New Control Air Source

If there is not an existing control air source, then it will be necessary to install one, which would include the costs of all of the items listed above along with the piping network.  When contrasting this with an electrically powered system, its important to realize that using electrical actuators is not totally free of cost in terms of providing the power required to support them.  However, the investment required to provide compressors and dryers is not insignificant and would need to be considered if the infrastructure was not already in place.

Installing an Electric Control Power Network

It is important to recognize that for most control systems, the network cables that carry the data passed around in the system are not capable of carrying the power required by the controllers and actuators.  That means you will need an independent power network in addition to the network cabling.

Granted, you will not need compressors, dryers, etc. like you would for a pneumatic system.  But, you will still be installing some hardware in the form of conduit, wire, circuit breakers and panel space in distribution panel boards along with low voltage transformers, control panels to house the low voltage transformers, etc.  If the actuators are powered by 120 vac, then the wiring will need to be installed in a manner that meets all of the requirements associated with 120 vac power, including conduit, a grounding system, branch circuit protection, etc. 

If the wiring is low voltage (24 vac is common for this application) then, even though it can run with out conduit, it will need to comply with the requirements of National Electric Code (NEC) article 725, which will tend to mean that the number of actuators powered from any given transformer will be limited.  That in turn will typically mean multiple power circuits and related transformers and distribution panels for areas with a lot of actuators (like the zone controllers in a VAV system, which will often have two actuators each, one for the air flow damper and one for a reheat coil control valve).  And, the low voltage source will, at some point, need to be fed from a 120 vac system, with all of the related requirements.

Pneumatic Actuation in the Central Plant with Electric/Electronic Actuation at the Zone Level

The preceding implies that for a new facility with no control air system, there would be a cost premium associated with my preferred design approach.  That is because you would have to install a control air system for the central plant but also would have to run the power required for the zone level equipment.  In my experience, the long term benefits will outweigh the first costs for most projects, but it probably pays to assess each situation on a case by case basis. And it is important to recognize that a life cycle cost perspective must be taken in this instance for it to make sense.  It will just about always loose on a first cost basis.

Deciding Which Way to Go

So, if my closing statement in the preceding paragraph is true, why would you decide to use pneumatic actuators in your central plant if you were planning to use electric/electronic actuators everywhere else in your facility?  Sometimes, you do it because of an existing standard or Owner preference, or because the Owner respects your preference.  But there is actually a technical reason to make the decision, at least in my experience.

Specifically, if there are valves or dampers that you will need to move quickly and/or that are large and will require a lot of torque, then the cost advantage will likely shift to using pneumatic actuation.   In fact, if you need something to move faster than about 15 to 30 seconds full stroke, pneumatic actuation may be your only practical choice with out going to some fairly exotic and costly (in the context of  HVAC control) technology.

Condenser Water Bypass Valves

One example of a common situation where this contingency comes up is with regard to the bypass valve on a condenser water system.   Typically, one of the functions of the bypass valve is to ensure that the minimum allowable condenser water supply temperature is provided to a chiller during start-up.  Even though lower condenser water temperatures will typically result in improved chiller efficiency, for most machines there is a limit to how low you can go because the temperature in the condenser will impact things like lubrication pressure. 

So, in an environment where you might shut down a plant over-night or for the weekend and where low ambient wet bulb temperatures can drive the idle water in the cooling tower basins down below the lower limit for the chiller entering condenser water temperature, you need a way to recirculate water at the chiller and use the heat rejected by the machine to warm up the loop. 

If you want to understand the issue in more detail, I describe it in an article I wrote for Networked Controls back in 2003 so I won’t repeat myself here since this is already getting pretty long!

Dampers in Systems Where Sudden Changes Cause Large Pressure Swings

A slightly less common example of where actuating speed matters is related to managing air flow and pressure relationships in systems where a sudden change in the system could generate a damaging pressure relationship.  One way to understand this contingency is to consider a clean room with large volumes of make up and exhaust air and a requirement to maintain the clean area at a positive pressure relative to the surroundings, but not so positive that you can’t open the doors and get out in an emergency. 

For the sake of our discussion and put some numbers on things, imagine that the make-up air is provided by two, nominal 50,000 cfm, 100% outdoor air units while the majority of the exhaust is provided by two nominal 44,000 cfm process exhaust fans.  The balance of the exhaust is provided by smaller fans serving things like process gas distribution cabinets, etc.  Because of the need to maintain the clean room at a positive pressure, a lot of attention has been focused on building the envelope to be leak free and air tight.

In a situation like this, if one of the process exhaust fans were to fail suddenly (say a motor overload trips, or the belts break for example), then there would be a sudden (in seconds) loss of significant exhaust flow.   To compensate for that with out over or under pressurizing the clean room envelope, the make-up systems need to be able to react and manage the make-up air flow just as quickly.  This will likely involve moving dampers and shutting down fans in about the same time frame (seconds).

The bottom line is that a field test of this failure mode on a recent project  with make up systems that had dampers that took 90 seconds to move full stroke resulted in:

  • A near implosion of the clean room structure (the peak observed static across the clean room was was about 1-1/2 inches),
  • Bent elevator door frames from the force generated when the elevator doors were slammed against them by the pressure difference, and
  • Failure of the test, which was required to receive an occupancy permit. 

Similar issues can come up in more conventional systems if they are required to operate to manage pressures for smoke control. 

The Energy Implications of Pneumatic Control Air Piping Leaks

Leaking pneumatic tubing can have a number of energy/resource implications.

Excessive Compressor Run Time

Pneumatic control air compressors are typically sized for a 30-50% duty cycle, at least back when I was selecting a lot of them.  Meaning that for any given hour, you would expect to see the compressor running 30-50% of the time.  If the compressor assembly was a “duplex” unit (meaning it has two compressors which alternate every other cycle) as illustrated below …

South tower pneumatic control air compressor

… then the run time of any given compressor will be a relatively small portion of any given hour.

Thus, in a retrocommissioning environment, one of the clues of opportunity is to go into a facility and discover that its control air compressor(s) are running most of the time if not all of the time.  The implication is that there is enough leakage that the compressors have to run a lot of the time to keep up.

That means the first savings opportunity is associated with eliminating the air leaks and thus, reducing the compressor run time.  Even though the compressors might be small relative to some of the other motors in the plant, continuous operation of the two duplex compressors on a typical unit vs. a 30% duty cycle can add up as is illustrated below (the numbers are from a project a couple of years ago).

Microsoft Excel - Pneumatic Leak Compressor Energy.xlsx  [Compatibility Mode] 2122013 111555 AM

Potential Simultaneous Heating and Cooling

But wait, there’s more as they say.   To understand this one, you need to think back to our discussion about spring ranges and extrapolate some of the concepts that were presented there.  And, you need to think about what happens in a piping system when there is a leak, especially in terms of the flow related pressure drop that will occur.

Leaks Create a Pressure Drop in the Mains Due to the High Flow Rate Through the Leak

Here is a chart of pressure drop vs. flow for air in a 1/4” tube, the most common size used for pneumatic control piping;  25 psig is a typical distribution pressure for control air for use at the final control elements.  Higher pressures might be used for distribution mains.  But the pressure generally has to be stepped down to below 30 psig because most pneumatic equipment, especially in the commercial market, is not rated for a pressure higher than that.


Here is a chart showing flow through orifices (leaks) of various sizes. 


If you study the two charts for a minute or two, you will realize that it does not take much of a leak in a pipe with 20 psig air in it to create a flow in the line that results in a significant pressure drop. 

For example, if you made a 1/16” hole in a air main with 25 psi in it, the result would be a flow in the range of 1.7 cfm through the pipe to the location of the leak.  In a copper pipe, that would result in a pressure loss of 10 psi per hundred equivalent feet of pipe.  For a plastic (poly) pipe, it’s off my chart.

Of course the pressure drop to the leak location would drop the pressure at the leak location, meaning the actual leakage rate would be lower than the 1.7 cfm associated with the 25 psi available at the source.  But I suspect you get the idea;  a small leak (or multiple little tiny leaks) can drop the available pressure at remote points in the system.

Leaving an open branch on a tee in a 1/4” line could cause big problems.


Low Pressure = Less than Full Actuator Motion

The bottom line is that if there are enough small leaks in a system or if somebody left a branch line on a tee open, or if a barbed fitting popped loose because it was improperly installed in the first place or because the tubing got hot and soft, the available pressure at a controller located at a remote point in the system might not be high enough to fully actuate the valve or damper actuator it serves, no matter what the controller did.

For instance, with only 8 psi available, the pneumatic VAV controller we discussed previously (that used reheat until the output from the thermostat reached 10 psi) would never stop reheating and never get off of its minimum flow setting.

It Gets Worse

Actually, the VAV controller might get off of its minimum flow setting, but not because the thermostat is asking for more flow by delivering a signal in excess of 10 psi.  Rather, it would happen because the VAV controller would be using the same 10 psi of supply pressure to power its internal logic. 

Depending on the design requirements, the VAV controller could be set up to work with a direct acting or reverse acting thermostat, a normally open or a normally closed volume damper and a spring range for the volume damper that could be just about anything from 2 to 5 psig to 15 to 2o psig.  If you take a look at the application guide  and data sheet for a KMC CSC 3000 reset volume controller, you will see what I mean.

As a result, depending on the details of the mechanism, a lower than required air pressure could cause the output from the VAV terminal controller (a.k.a the volume reset controller) to the damper it controls to drop enough that the damper would actually open up.  Stated another way, because of the low supply pressure, the controller would not be able to close the damper even though it wanted to.

A Signature of Dysfunction in the Utility Bills

The combination of the terminal unit damper not being able to close along with the reheat valve not being able to close leads to unnecessary reheat that is “masked” by unnecessary air flow.  In a retrocommissioning environment, the signature clues are:

  • VAV fan systems that run at fairly high flow rates with little if any shape to the load profile,
  • Higher than anticipated thermal consumption due to the excessive reheat, especially in the summer, and
  • Higher than anticipated electrical consumption due to the higher than necessary air flows, especially during lower load conditions.

Bear in mind that if the reheat coils are electric, then the thermal consumption I mention above will show up in the electric bill, not the gas bill.

Because the excess reheat is masked by the extra cold air flowing through the damper, there may not be occupant discomfort complaints.  And, since we tend to run facilities by responding to occupant complaints, the problem often goes undetected.

Thus far, I have been talking about this issue in the context of a VAV terminal unit flow controller. But its important to recognize that similar things can happen in central systems located at the far end of the control air distribution network.

Its also important to recognize that leaks can evolve over time (retrocommissioning opportunity), or they can be the result of an oversight during construction., meaning inadequate control air pressure at remote points in the mains is there from the start.

In either case, the signature in the utility bills will be similar.

Why This Matters if you are Considering a New Pneumatic Control or Actuation System

The reason the issue of air leakage matters when you are considering a new pneumatic control or actuation system is that you will want to install the system in a manner that prevents the leakage in the first place and facilitates identification of the problem if (when) it happens.

Copper vs. Poly Tubing

There are a number of reasons why copper piping systems will tend to be less prone to leakage when compared with poly piping systems.

Copper Piping will be More Durable

Because it is metal vs. a soft material like plastic, copper is less likely to be damaged by activities during construction, operation, or renovation.  For example, pneumatic lines often penetrate firewalls and smoke separations.  All penetrations through separations need to be sealed to maintain the integrity of the separation.  So, when someone is using some sort of tool, like a trowel, to seal around a pneumatic line, it is much less likely that a copper line would be nicked or otherwise penetrated than a soft poly line.

A common approach used to provide better durability while retaining the lower installation cost associated with poly tubing is to pull the tubing into a conduit system, usually fabricated from thin-wall conduit (EMT or Electro Metallic Tubing) and fittings.  When using this approach, the issues discussed next, related to the fittings used for the tubing, still apply.

Copper Piping will be Joined with Fittings Less Prone to Leakage

Copper tubing will likely be joined using solder …


… compression …


… or flare fittings. 


In contrast, poly tubing is typically joined using barbed fittings …


… although compression fittings are an option.


(All of the fitting pictures are screen shots from the Parker electronic fitting catalog except for the solder fittings, which are from the Mueller Industries HVACR catalog.)

All of the fitting styles have their pros and cons, but the solder, compression, and flare type fittings typically used with copper pipe are less likely to come apart.

For poly systems, barbed fittings are the fitting of choice due to low cost and ease of use.  However, one of the issues with poly tubing is that if it is run in a warm environment (like in a ceiling cavity under a roof or in an industrial plant) and barbed fittings are used, then, as the tubing softens with heat, there is a tendency for the joints to blow apart, especially if the tubing has been used as a high pressure (above 25 psig) distribution main. 

Using compression style fittings can alleviate this to some extent, but at a significantly higher cost per fitting.  It is also important to recognize that the strength of the poly tubing itself is reduced at elevated temperatures.

Assessing Integrity

There are a number of ways to assess the integrity of a pneumatic tubing system.  The obvious one is to pressure test it at the time of installation.  While it is probably not cost effective to pressure test every piece of tubing in a pneumatic control system, taking the time to test long runs and the distribution network, especially when it will be concealed by subsequent construction, can be time will spent.

Once the system is up and running, monitoring pressures and pressure trend at remote points in the system can provide an indication that a leak has developed in addition to helping to pin-point the leak.  Some manufacturers actually have this feature built into the pneumatic output circuit boards. Monitoring compressor run time and run time trends can provide similar insights.

Finding a leak once it has been discovered can be harder than it sounds.   One approach is to walk the facility late at night when things are quiet.  Ultrasonic leak detectors can be handy in pinpointing the exact location of a leak and may even identify ones that you can’t hear unaided.

Frequently, you can narrow the area you have to search in by isolating portions of the distribution system and observing what happens to the pressures other places in the system.  For instance, if you valve out a section of the system and there is no change in pressure at the remote locations, then the leak is probably in a different part of the system.   In contrast, if you valve out a section and the pressures rise back up at the remote points, then it is likely that there is at least one significant leak in the section of the system that you isolated.

Including isolation valves in the distribution system as part of its initial fabrication can help facilitate this process when it becomes necessary.  Taking the time to note were the valves are located on the “as built” drawings is also highly desirable.

The First Cost Implications of Installing a Leak Resistant Piping Network

The bottom line is that spending a little money up front for a piping system that has good resistance to leakage will yield long term benefits and savings over the life of the system, both in terms of energy and in terms of the labor it takes to locate and repair any leaks that do occur.

Using copper tubing or poly tubing in EMT will provide resistance to damage due to day to day operations and renovation activities.  Using compression fittings instead of barbed fittings on poly systems will generally minimize the potential for a failure at a joint. Taking the time to verify critical piping runs are free of leaks to start out with and then providing monitoring and isolation valves to detect and help pinpoint leaks when they occur will reduce the labor required to restore functionality

But all of these features come at a cost that should be assessed when weighing your options in terms of pneumatic vs. electric/electronic controls and/or actuation systems and reflected in the budget if you elect the pneumatic option.

Implementing Complex Strategies

Frequently (but not always) optimizing energy efficiency and performance will tend to involve control strategies that are more complex than what it would take to simply maintain a comfortable and safe building.

Ease of Implementation

In my view, implementing complex control strategies will be easier in a DDC system (Direct Digital Control system; i.e. digital, non-analog, microprocessor/computer based control) than in a pneumatic (analog, discrete element) control system.   To gain some insight into this, contrast this picture of the wiring inside my relay logic controlled Jeopardy Game board

Jeopardy Game 02  Jeopardy Game 04

… with the ladder diagram behind the wiring (the link takes you to .pdfs of you really are interested in the details).


I realize that the ladder diagram probably looks pretty complex.  But if you think about trying to figure out how to make the game work our diagnose a problem, I think you might conclude that it would be easier to work with the ladder diagram than the actual physical wiring.

Expanding the analogy to control systems, contrast this pneumatic control panel, which controls a VAV air handling unit …

S18 Receiver Controller Panel

… with this logic diagram, which also also controls a VAV air handling unit (the link takes you to a better resolution .pdf).

VAV2 Logic Diagram

While complex, the logic diagram is probably a better tool for developing and troubleshooting a VAV control sequence when contrasted with a bunch of parts wired and piped together to produce a sequence of events (or a bunch of parts that you have to figure out how to wire and pipe together to produce the sequence).

Control programs are very much like working with the logic diagrams, especially graphic type programs like Eikon (ALC’s programming language) or the ladder logic used in programmable controllers.  Working with discrete hardware like a pneumatic control system is literally working with a pile of parts you have to wire and pipe together to make a sequence.  To troubleshoot, you have to try to trace through the bundled wires and tubes in the panel or follow them when they leave the panel to go to the field (and disappear behind walls and ceilings).

So from that perspective, a DDC system will win in terms of ease of implementation when contrasted with a pneumatic control system. 

Note that from a purely technical standpoint, implementing a complex strategy in a pneumatic control system would likely be more interesting and challenging, especially if you are like Matthew in the recent Portlandia Rube Goldberg skit.


In light of the preceding, I suspect you can see how making a change to a software based system is much easier to do than making a change to a hardware based system.  In the former case, you are typically moving electrons around via the programming tool.  In the latter case you are moving wires and tubes and may even need to go buy a different part if you want to change the way something works.

In my mind, one of the things you buy when you buy DDC is flexibility.  That flexibility can come in very handy when you are trying to implement complex strategies to save energy and other resources or optimize them.  So, this is another area where I think DDC will “win” when contrasted with a pneumatic control system.

“Gene Pool” Issues

I borrow the “gene pool” term from Jay Santos, one of FDE’s principles.  What I mean by it is that as pneumatic controls are replaced by DDC, the number of people who are familiar with them and understand them is diminishing.  The reality is that anybody with an interest and an aptitude for mechanical things (like Matthew from Portlandia) could learn how they work and apply them.  But the number of people actually doing that or who have experience doing that is diminishing. 

One of our field engineers was regaling us just the other day of an incident on a job site where the control technician charged with installing a new pneumatic system (an Owner preference) was at a total loss on what to do with the pile of parts he had received  to work with.  That was because in his entire career with the company, he hand never had to work with a receiver controller before. 

That doesn’t mean he didn’t figure out how to do it, but I bet it took him some time and he may not have been as confident in the outcome as he might have been had he been able to implement the controls using the DDC technology he was familiar with.

So, if you were contemplating installing a new pneumatic control system, you might have to ask yourself if you think the people that will be installing it and maintaining it will understand it.   Given where technology is headed and where the new people entering the industry are coming from, it is probably more likely that a DDC system would be understood than a pneumatic system.


In my view, the key to all of this, in terms of making things more sustainable, is persistence of benefits.  Meaning that it is one thing to implement a complex strategy that saves energy.   Its another to have that strategy remain fully functional over time. 

If you read between the lines a bit in the preceding paragraphs, you are likely to conclude that there will tend to be persistence issues with a pneumatic control system that might not exist in a DDC system.  That’s not to say there are not persistence issues with DDC systems.  Most folks out in the field know of several instances where someone accidentally loaded an old program over a new program, eradicating the improvements that were contained in the latest version. 

And then, there are the hand-off-auto switches.  No matter how dedicated you are to optimizing the schedules in the operator work station, the effort will yield no benefit if the selector switches at the starter are positioned to remove the automation system from control of the load served by the starter.

But issues like that aside, DDC systems are less prone to failures due to the wear and tear of motion.  Pneumatic controllers are full of levers, pivots, bellows, diaphragms, springs, tubes, and orifices.  These elements can bend, wear down, take a set, rupture, pop loose and plug up. 

So even the highest quality devices will tend to require some expert attention (see Gene Pool Issues above) to recalibrate and adjust them every 6 to 12 months, all other issues aside.  If the air dryer serving the pneumatic system fails or the compressor starts throwing oil out into the system, things can go down hill pretty fast as the little tiny orifices fill with oil and water and then corrode and pick up debris. 

And it is no easy thing to bring the system back when that happens.   You probably will end up throwing away a lot of parts and spending a lot of time purging tubes with something to clean them up.

For all of these reasons, in my experience, control strategies that are implemented and commissioned in a DDC system are more likely to persist than those that are implemented in a pneumatic control system (or any control system built up from discrete elements with moving parts.  (Again, I want to emphasis that I am saying pneumatic control system, not pneumatic actuation system.  Pneumatic actuation systems, in my experience, will probably be more persistent than electric/electronic ones if the air source is well maintained, especially for larger actuators.)

Bottom line, in the context of efficiency codes and sustainable operation goals, DDC systems will probably show better persistence assuming that they are set up and made to work properly in the first place (i.e. commissioned).

The Cost Implications of Implementing Complex Strategies with Pneumatic Controls

From a first cost standpoint, pneumatic controls are likely to win out over DDC controls unless someone takes the time to really understand what they need and then writes and enforces a specification that will procure it.  Given the way the market is heading, to get a really robust pneumatic control system, you would need to use process grade equipment and my guess is that the people out there specifying pneumatic controls in commercial buildings are not writing that type of spec.

That means the systems will be fabricated from the parts that are commercially available and that field is narrowing as demand drops off due to DDC taking a larger portion of the market share.  So, with out a strong spec, a job that is competitively bid will likely have the lowest cost (usually also lowest quality) parts available, something that probably works against long term success and thus long term savings.  It may also work against achieving the design intent in the first place if the parts can not deliver the precision and repeatability required by a complex, nuanced sequence.

The bottom line is that a pneumatic control system would likely look attractive on a first cost basis to someone unaware of the potential pitfalls.  That is probably the reason you still “idea” come up as an alternative in a value engineering session, especially as an alternative for DDC zone controls.

But because of the pitfalls, the pneumatic system will likely not deliver on the intended functionality, especially in the long term.  So that means that an attractive first cost will likely lead to a very poor life cycle cost, with the ultimate penalty being a system that does not work and has to be upgraded to DDC at some point in the future to achieve the desired functionality.  That scenario means you basically throw away the initial investment and start all over again.  But now you do it with a facility full of people, production processes, walls, and ceilings.

So, despite having a bit of Matthew from the Portlandia skit in me, I have to raise my hand in favor of moving away from pneumatic control (not actuation) and towards DDC if at all possible in a new facility or in a retrofit or repair situation.  If you can’t tell  that after reading all of this, then maybe I’m just upside down.

David Sellers
Senior Engineer – Facility Dynamics Engineering

Click here for a recent index to previous posts

This entry was posted in Controls, Pneumatic Controls, Retrocommissioning Findings. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s