Field Pressure Transducer Calibration Setup

[Droshi]

Hi all,

I'm wondering what sort of setup is ideal for this sort of field calibration? The pressure transducers are for static duct pressure: http://www.mamacsys.com/PR-276features.htm

We currently have a big test bench with a large incline manometer, but for one, I don't trust the rigged setup (and I'm unsure how to correct it, what should be the pressure source etc) and I'm also sure there's a proper and elegant way to reset this test bench. But also I would like a setup that is easily carried in a tool-bag and can calibrate the sensors hopefully by just pulling them out of the duct and not having to disconnect them etc.

Any suggestions on either setup are very welcome, and pictures/diagrams will help my thick head too! I've also tried the search and came up with some related material, but nothing specific on this, so I thought I'd post.

Thanks!

[beer30]

I mount the Mamac transducer in the control panel, with the poly line running out to the ductwork. I install a tee with a plug in the line in the panel. I plug a mag gauge into the tee to check/calibrate the transducer. I may put the vfd in hand while I disconnect the poly line to prevent ramp-up/high duct static.

[Vortech]

I use this for my Pharma calibrations.....

http://ashcroft-gauges.com/pdf/ate_100.pdf

-V-

[Droshi]

Beer:
That sounds like a nice idea, how far can you run the poly line? In some places our panels are a bit of a distance from the pressure sensors, but for future it's nice to keep in mind.

Also, my idea for proper calibration would be to zero the device, and then measure at a few different pressure points with gage connected, using the trim-pot to find the best calibration point. Is that the standard procedure for these devices? My question then becomes, how do you hook up a pressure source and what do you use? Small bicycle pump? Etc? Anyway, I had planned to jury rig up something, but it's probably better if I find out what others use :)

[Vortech]

some calibration devices provide the air pressure or you can use a small hand held calibration pump.
fluke has a nice dp calibration setup with a built in pump
http://us.fluke.com/usen/products/Fluke+718.htm?

catalog_name=FlukeUnitedStates&category=PRESSCAL(FlukeProducts)
what type of industry are you working in?
a proper calibration on a DP sensor is the typical 3 point low mid high
ie on a bi-directional .25 it would be -.25 0 +.25
making sure you always did the Zero first of course

[Droshi]

I'm doing building management for a number of buildings. Most of the HVAC is standard office, but we have a couple special rooms as well as other special applications that are on our DDC system.

Our sensor calibration PM type work is basically non-existant or not standardized. My idea is to plan not only for PM work on the DDC system and sensors, but to extend a bit further to make sure everything is working properly for the HVAC system overall. Right now things are quite divided between mechanical, electrical and controls...but with many things I've been reading on this site I think a more integrated approach is necessary. The good thing is that I'll be learning more mechanical systems along the way.

The fluke meter with a built in pump looks quite handy...but it only goes up to 12 psi? It looks like maybe they have a module that extends that...is that right? It also mentions a pressure calibration kit...I assume it's necessary to buy that as well to keep the meter in tune?

Much thanks for all the answers, sorry for my ignorance!

[Vortech]

we dont buy the calibration kit due to needing a N.I.S.T. Cert from a lab for the calibration of the meter, yes they have diff modules you can get for a number of pressure ranges

-V-

[beer30]

Beer:
That sounds like a nice idea, how far can you run the poly line?

I have run the line hundreds of feet. At the time, it seemed like way too far, so I checked the pressure at the duct, and again at the panel, a few hundred feet away. My mag gauge showed no difference in pressure. Also, I have only checked calibration on these devices at zero, and at the range it is expected to run. Who cares if it is accurate in the middle, it only briefly passes through those pressures.

[osiyo]

I have run the line hundreds of feet. At the time, it seemed like way too far, so I checked the pressure at the duct, and again at the panel, a few hundred feet away. My mag gauge showed no difference in pressure. Also, I have only checked calibration on these devices at zero, and at the range it is expected to run. Who cares if it is accurate in the middle, it only briefly passes through those pressures.

Just a couple of points to consider.

Yes, if you set up the VFD to run at a steady speed while you are checking the system (and I also set any VAV boxes to lock their dampers into a fixed position at the same time) you will not get any differences between measurements taken at the point where the pickup tube connects to the ducting as compared to measurements at the remote point where you've mounted the pressure sensing device. Because you have a steady state condition, and things equalize.

The problem (possible problem) comes into play when the system is live and demand is varying more or less constantly. Under such conditions, very long runs of small tubing can act as a restrictor/metering device. That is it can produce a time delay into the measurements. For instance, you can get a lag time between the point at which the duct pressure drops, and the sensing device detects this.

This may, or it may not be significant. Depends on a number of factors.

Friction loss (pressure drop) over long tubing lengths may also may come into play, depending upon the type of sensor used. One should always check with the device manufacturer. Especially if using a new type sensor with which you are unfamiliar. This bit us in the a** on one project. Not one of my projects, if it had been I would have probably checked beforehand as I've run into this sort of thing before. Customer (his HVAC design engineer, actually) specified a particular make and model of sensor, no substitutions allowed. Our controls design engineer didn't alert on this and do proper investigation before laying out his plans, and our installers didn't question anything, they just did what our engineer laid out. Nobody read the manufacturer's literature in any detail.

Net result, things didn't work out right. We were having pressure control issues. I was sent out. Looked over the system, scratched head, etc. Everything looked okay. At least it was not significantly different from what we'd done successfully before. But our pressure readings were way off as compared to locally made measurements by independent instrument. Then I looked at the actual sensors installed. Brand new item to me. Looked up the manufacturer's manual for them and read it.

Ahhh, Geez. Manufacturer had very specific guidelines for lengths of tubing runs versus size of tubing to be used. This was a very large entertainment complex, actual transducers mounted in machinery rooms near the controllers, with long tubing runs from sensing points to transducers. IAW manufacturer's suggestions, the tubing size we'd used was way undersized. Net result, we had to redo tubing runs completely. And by this time the building was done. And it was a COSTLY mistake. A large complex, long lengths, many air handlers. Several of the duct attachment points were 25 foot or better high, and over occupied areas with booths, desks, counters, and in one case over an auditorium area with bolted down spectator seating.

Cost us big time.

So check the sensor device manufacturer's literature. Just in case there are some special requirements.

That said, where I work opinions vary. Some prefer to mount the transducer's in panels or some other centralized, easy access location and then make long tubing runs.

I personally prefer to mount the transducers right at the duct tapping location. Most of the time. Not always.

If the tapping point is going to be at a really difficult to access location after the building is occupied will put the transducer in a cabinet in a machinery room or wherever so it's easy to calibrate and check or replace if necessary. Then, if the device manufacturer has no instructions otherwise, I have installers use extra sized tubing. 1/4 inch for short runs of 100 foot or less. 3/8 out to about 200 foot. 1/2 inch for longer than that. Everyone seems to have their own opinion about this, that's just mine. Works for me, never had a problem. I HAVE had problems related to 300 foot or so runs where somebody used 1/4 inch all the way.

However, usually I prefer to have transducer mounted at the tapping point. The reason? Besides any possible issue with pressure drops or time delays induced by long tubing runs. Poly tubing is a bit more troublesome to pull as compared to an 18/2 or 18/3 cable. And since my installers are pulling cables anyway, it's just another cable in the bundle. In addition, many a times when troubleshooting problems with pressure control I've tracked it down to excessively bent of crimped poly. Sometimes the initial install and check was okay. But later on somebody was doing something and that was went the crimping occurred. Seen it happen far more often than I care to remember.

As concerns the use of copper tubing for such sensing air pressure (duct or space) lines. Well, my installers can pull cable a whole lot faster than they can put in copper tubing. We only use copper when its dictated. Which it sometimes is. In some specs, they specify you can use poly for short, local connections, must use copper for long runs.

As concerns calibration checks, just how accurate do you need to be, in all reality? For things like duct pressure control in a HVAC system?

Now we have times we've got to be very accurate for space pressurization control and similar, in special facilities with special requirements. For instance, with a few of our customers in the medical products and pharmaceuticals biz. Where they have special clean rooms and so forth. And must meet special requirements that require the use of NIST certified instruments and such.

But normally, absolute accuracy is not a real factor. Close enough is good enough. I go through the procedure to ensure the sensor is zeroed. Then do a check at normal duct pressure and at the high end (near where high pressure limits would be set). I just tee off the same connections that are used for the transducer, and connect my magnahelic. (I have several, I use the proper range for the application) Heck, for normal apps, if the transducer is within 5% or so, even 10%, of what the magnahelic is saying, I consider it good to go. Not worth the time to go through more trouble.

The fact is that at the end of the job we're gonna have a balancer check over the system. And the actual duct pressure setpoint we're gonna set will be based upon his readings and data. What he determines to be the lowest we can go and satisfy the worst case VAV.

Yeah, I know about demand reset of duct pressure. Tried it, didn't like it. A problematical solution. That I won't go into here. Just suffice it to say I don't do it unless it's unavoidable. Let's just say it worked at first, but was a high maintenance tactic that created its own problems long term.

Just my thoughts and opinions, worth no more than that. Which means ... they're not worth even a cup of lousy coffee.

[Droshi]

Good points.

I don't need something extremely accurate, as you've said, most times close enough is good enough. But I would like something quick and easy to check the pressure transducers against while in the field. Taking out the sensor and bringing it back to a test bench isn't my idea of a great time.

T-Junction and a magnahelic would probably work for us on all of these sensors. Do you measure only at zero and operating pressure as well? Or is it necessary to use a pump to measure at other points too? The only issue now is the mechanical setup, which I'm pretty clueless about, but I'm sure I'll figure that out.

[osiyo]

Good points.

I don't need something extremely accurate, as you've said, most times close enough is good enough. But I would like something quick and easy to check the pressure transducers against while in the field. Taking out the sensor and bringing it back to a test bench isn't my idea of a great time.

T-Junction and a magnahelic would probably work for us on all of these sensors. Do you measure only at zero and operating pressure as well? Or is it necessary to use a pump to measure at other points too? The only issue now is the mechanical setup, which I'm pretty clueless about, but I'm sure I'll figure that out.

I zero the sensor. Done with all tubing disconnected from it's input ports. Then check readings at or near normal expected operating pressure, then somewhere near high limit safety trip. i.e For a typical system, I do a check at 1.0 inWC (normal running point, close enough) and another at 2.5 inWC since we normally set trips (hardware and software) at 3.0 inWC. I'm more interested in accuracy at that midpoint than I am at the high end.

Now, sometimes I've got another guy on site who is at the front end or whatever. Who is telling me what the controller is seeing as I'm looking at my magnahelic.

But usually not. Unless its a really big project, and we're pressed for time, we rarely have two techs on a job site at once. The way we work where I'm employed, we have installers and techs. Installers install. Techs do testing, commissioning, programming, etc. In fact, the initial programming is usually done back in the shop (not always but usually). And a tech just simply gives the programs a test run and evaluation. Then modify as needed, if needed.

When working alone, besides my magnahelic, I've have my multi-meter with me. Typically we'll use milliamp signal output sensors, especially if there is any long cable run, for more accuracy and protection from EMI/RFI influences. My meter (one of them, I carry several in the truck) reads milliamps directly. I just hook it up in series with the signal output. Compare that to my magnahelic readings.

For instance, if duct pressure is 1.0 inch, I should see 9.3 milliamp on the meter (or close enough). At 2.5 inch duct pressure, I should see approx 17.3 milliamps.

If yah don't know how to figure the math out, do a search. This kind of thing has been discussed on this forum before. If yah can't find the answer, ask here again. One or more of us will help yah out.

[Droshi]

Easy enough, we usually work with voltage signals (Honeywell products), but also I usually have someone in the office that can view it from the software as well.

To generate the higher pressures do you use a pump of some sort? The nice thing about that Fluke meter is the built in pump, but if it's not really necessary I like the idea of the simplicity of just a magnahelic and tubing.

[osiyo]

Easy enough, we usually work with voltage signals (Honeywell products), but also I usually have someone in the office that can view it from the software as well.

To generate the higher pressures do you use a pump of some sort? The nice thing about that Fluke meter is the built in pump, but if it's not really necessary I like the idea of the simplicity of just a magnahelic and tubing.

Don't usually use anything but the unit fan, running the VFD in hand. Have used a small bottle of nitrogen, a regulator, hose with both an inline needle valve and a tee after that with a "bleeder" needle valve that'd allow for very precise pressure control. Along with a larger faced, high accuracy gage that I have periodically checked. Such items are normally in my truck for a miscellany of uses.

Haven't found a hand pump that works well at such low pressures.

But usually I just use the AHU fan. If I can't build up enough duct pressure, I command some of the VAV's to a damper shut condition.

The important part is to have a steady state condition. So VFD in hand, constant speed. VAV's commanded to full open or full shut (some of each in order to be able to build up enough duct pressure). Doesn't take long to do all that and it's easier than rigging a separate pressure source and hauling it around. Given that I work solely on DDC systems, so commanding the system to the condition I want isn't difficult, nor does it take significant time.

That Fluke meter looks nice, tho I've never seen one such in reality much less used one. However, in my mind, hard for me to justify if it costs much. i.e. More than $100 or $150 bucks. The magnahelics I have are LONG ago paid for, reliable, don't require batteries, and don't much care if they are left in the truck on a -20'F day. Just got to let em defrost enough so the glass face stops fogging up.

[crab master]

The magnahelics I have are LONG ago paid for, reliable, don't require batteries, and don't much care if they are left in the truck on a -20'F day. Just got to let em defrost enough so the glass face stops fogging up.

FYI - I asked my dentist what they use to keep their mirrors from fogging up and they said simply dip the mirror in some water with a bit of dish soap. I tried it and it definitely worked. The trick is to either let it simply air dry or dab it dry, but don't wipe it as you'll wipe off the thin soap film that is keeping the fog from forming on the mirror. Anyway a simple, cheap solution.

[osiyo]

FYI - I asked my dentist what they use to keep their mirrors from fogging up and they said simply dip the mirror in some water with a bit of dish soap. I tried it and it definitely worked. The trick is to either let it simply air dry or dab it dry, but don't wipe it as you'll wipe off the thin soap film that is keeping the fog from forming on the mirror. Anyway a simple, cheap solution.

Thanks, I'll have to try that.

[jjackson7]

Yeah, I know about demand reset of duct pressure. Tried it, didn't like it. A problematical solution. That I won't go into here. Just suffice it to say I don't do it unless it's unavoidable. Let's just say it worked at first, but was a high maintenance tactic that created its own problems long term.
I was reading a spec book recently for a remodel at my campus and the engineer wants to reset duct pressure down "until the primary air flowrate to one of the associated terminal units has been below the required value for more than an operator established period of time" and then back up if vice versa. It seems to me this added dimension to pressure control will require many calculations (INET remodel), be unreliable and unnecessary. any thoughts?

[Droshi]

I was reading a spec book recently for a remodel at my campus and the engineer wants to reset duct pressure down "until the primary air flowrate to one of the associated terminal units has been below the required value for more than an operator established period of time" and then back up if vice versa. It seems to me this added dimension to pressure control will require many calculations (INET remodel), be unreliable and unnecessary. any thoughts?

In my opinion, it's not really necessary. It saves a bit on fan speed, but really how much actual electrical savings is that? Especially for all that programming. I would say it's a much better idea to use resets on things that take up chilled water (chw temp from the chillers, duct temp setpoint, etc).

[crab master]

I've done a simple reset and it has worked very effectively - cost savings and improved comfort, especially on VVT type systems. On VAV systems it simply is more of a cost savings since the VAV box is controlling to an airflow setpoint.
Simply reset your static setpoint based on outdoor air temp and make that setpoint adjustable. For example say below 55 deg OAT reset your duct static to 1.0" w.c and then above say 58/60 deg OAT reset it to 1.5" w.c. You want to make sure your programming/PID loop is setup correctly to handle this change when it happens. Most places are set just as stated above but a few I've had to tweak. You could look at the highest terminal load of a vav box or others, but likely you'll have an OAT readily available as an input to your AHU.

[Vortech]

In my opinion, it's not really necessary. It saves a bit on fan speed, but really how much actual electrical savings is that? Especially for all that programming. I would say it's a much better idea to use resets on things that take up chilled water (chw temp from the chillers, duct temp setpoint, etc).

you will see more of this with buildings trying to get LEED certified every little "point" add's up and the differance from a normal morden building -vs- a leed cert building is miniscule and some engineers are actually saying the some Leed goals hurt performance.

as far as calibrating end devices there is the right way that meets N.I.S.T or GMP standards and then all other ways thats just "good enough", I prefer to take the time to do each calibration with a proper SOP so there can never be any question if it was done right and we also look more professional.

-V-

[osiyo]

you will see more of this with buildings trying to get LEED certified every little "point" add's up and the differance from a normal morden building -vs- a leed cert building is miniscule and some engineers are actually saying the some Leed goals hurt performance.


Can't disagree. And what I've also seen is POI getting extended way out there.


as far as calibrating end devices there is the right way that meets N.I.S.T or GMP standards and then all other ways thats just "good enough", I prefer to take the time to do each calibration with a proper SOP so there can never be any question if it was done right and we also look more professional.

-V-

I can't entirely disagree with you. If I had my druthers, given that I come from a background in the industrial process world, I'd do a proper calibration and certification on everything.

HOWEVER ... one must be realistic. Most customers don't have endless funds. And one must consider the principles of diminishing returns.

Percentage-wise, just how much additional energy savings can you GUARANTEE that a customer will realize by having a job done where all sensors are checked and calibrated IAW NIST standards. As versus having sensors checked and adjusted if necessary by the use of "good enough" methods? Given the additional costs of labor, the necessary tools and instruments, etc.

Numbers, please. Both the raw data and your claimed results. The raw data because I'd like to check your math.

And keeping in mind that I have done installations that had to meet NIST standards, so I do know what it takes in additional time and effort. Of course those were installations done for pharmaceutical and medical materials producers. Who had to meet very exacting FDA requirements. Places that were turning out from a million to several millions of dollars per day of product. Very sensitive product. Where deviations of a couple degrees at some controlled temperature point, for more than N period of time, meant yah had to toss the whole batch and start all over again.

Just did such an installation last year. Took 4 technicians 3 weeks to do their thing JUST to get the calibration and adjustments done to meet NIST standards. Took 20 to 30 minutes per temperature sensor. One guy at the sensor with his gear, another at a laptop observing readings, recording them, filling out the darned required paperwork, adjusting offsets for the inputs, etc. Longer for pressure and flow sensors. In the case of this customer, a pharmaceutical producer, the whole check and calibrate process will need to be repeated each year.

But in their case, it's worthwhile. Just ONE fault or error will cost em far more than that calibration check process.

But, as a percentage, how much REAL energy savings is gonna be realized by a customer with an ordinary office building that has an energy bill of, say, ... $100,000 per year if they go the NIST standards calibration check route as versus just having the technicians go the route of using due diligence in that they DO check all sensors, and use reasonable and "good enough" means to check and calibrate them.

In all actuality, most existing installations I've checked over when we're taken over an account from some previous contractor, or we're doing a retrofit, etc ... I usually find myself doubtful that anyone ever did any MEANINGFUL calibration check, of any kind, ever.

The exception to that is mostly in the case of the certified factory startup reps for commercial/industrial grade chillers and boilers. Most of whom seem to be thorough, cautious, and have both the knowledge and necessary tools to get it done right. Not all of em and not always. But mostly.

As concerns duct pressure sensors. I presume you know that yah can take readings at a dozen, or two dozen different points along a duct and get some variations between all those readings, and that is perfectly normal given that the flow is at least somewhat turbulent as versus laminar at many, maybe most, locations. And that at best, a single point tap and its associated sensor is just giving yah a ROUGH average for that point, and that point only. In fact many modern pressure sensors commonly used for sensing duct pressure have built into their circuitry a means of filtering (averaging) the many fluctuations that are constantly occurring (and which are quite normal). So at any particular instant of time, even if you're using a very accurate instrument to double check that sensor, variations in the readings produced by the two are quite normal and to be expected. With your hand held instrument showing the live, right now reading, while the installed sensor is showing a filtered average.

I don't think meeting NIST standards for the average, normal installation is worth either the time or the energy involved. Not in further real, measured, energy savings to the customer.

I DO, however, believe that any good automation person should exercise due diligence and actual check/adjust all sensors. Using reasonable, cost effective means, AND instruments that are reasonably accurate and regularly checked.

Heck, most I've run across rarely check their portable instruments for accuracy at all.

Chuckle, it's like last week. Got a complaint call from a customer. The installation was within warranty. So I went to take a look. Customer has an in-house HVAC specialist. And he told me one of our space sensors was unacceptably deviated from the real temperature.

Then went on to "prove" it by pointing his infrared temperature sensor at it, taking the reading, and then showing what was displayed on the front end. <Shrug> I broke out my stick type electronic temp sensor. A lab grade item, which I routinely double check. Yah know, the old glass of CRUSHED ice and water trick (don't use cubed ice, it makes a difference). Stuck it on top of the space sensor housing, and let it settle out for 10 minutes. It does the automatic averaging/filtering thing. It produced a reading within a half degree of the space sensor. Then, just to make my point, went to my truck and selected a couple spare space sensors and temporarily installed em in place of the existing one. Neither varied significantly from the existing sensor (one read about half degree different one way, the other about 3/4 degree the other way).

That guy's infrared sensor was reading consistently 5 degrees high.

Chuckle, he turned red, groaned about the fact that it was a nice, new sensor and that he'd had his employers pay a pretty penny for it as he'd bought the best (according to the literature he'd read) he could find.

I suggested he us it just for troubleshooting. Something is getting hotter, or something is getting colder. But not for taking temperatures where accuracy was needed.

Same goes with pressure gauges. I've got some for rough work. Don't care much if they get banged around. For use when I just want to know if I'm in the ballpark so to speak. Then I have some that are much more accurate (and more pricey), that I keep in hard cases for protection. And which, from time to time I put on the dead weight tester we have back in the shop, just to be sure. Those I break out only when I need accuracy.

[ckartson]

You may find this to be to simple and just blow it off but I'm going to say it anyway and risk being laughed at. If you are bench testing your sensors at a control panel or in your shop, use a glass of water and a piece of 1/4 tubing (I like tygon tube it's very flexible and clear) mark it whith a marker in 1/2 inch intevals. Push the tubing in the water 1/2 inch at a time and read your signal. We are reading In w.c. so it is easy to test with water.

[ckartson]

Why would you need more than 12 psi to calibrate static sensors? I think you would need 5 in w.c. max.

[Droshi]

Yeah, 12 psi will be plenty for me. I was confused by conversion as I don't think in psi really.

I've rigged up a good magnahelic setup now, so I'm pretty happy with it. I just need to get a couple more for better resolutions.

Thanks for all the help everyone! Brilliant forum.