Testing Accuracy of a 0.25 Water Manometer Using a 10:1 Inclined Mount

[paulmars]

I need to test this gauge. It reads -0.25 to 0 to +0.25 inches of water. I am using it in the 0.02 to 0.03 negative and positive ranges. What of known value can I use to deflect this gauge maybe .1 or .05 in either or both
directions to test the gauges accuracy?

it gives the option of selecting 'no prefix' however doing so gets error message "Please select a prefix."

pa

 

[berkeman]

I need to test this gauge. It reads -0.25 to 0 to +0.25 inches of water. I am using it in the 0.02 to 0.03 negative and positive ranges. What of known value can I use to deflect this gauge maybe .1 or .05 in either or both
directions to test the gauges accuracy?

it gives the option of selecting 'no prefix' however doing so gets error message "Please select a prefix."

pa

Welcome to the PF. What kind of manometer? Can you post a link to the datasheet? And what kind of accuracy are you looking for?

 

[berkeman]

it gives the option of selecting 'no prefix' however doing so gets error message "Please select a prefix."

I'm assuming this "it" is the forum software, and not the manometer, right?

 

[paulmars]

clicking quote does not work. as accurate as possible, let's say 0.005wc.

https://www.ebay.com/itm/Dwyer-Magnehelic-2300-0-Differential-Pressure-Gauge/263459109078?epid=12013719917&hash=item3d5762f8d6:g:4D4AAOSw4DxaajFz

 

[paulmars]

+/-0.005

 

[berkeman]

clicking quote does not work.

Yeah, the "Quote" feature is not very intuitive to use. Instead, click "Reply", which automatically quotes the post for you. Or click-drag the sentence you want to quote, and then click the "Reply" popup option.

https://www.ebay.com/itm/Dwyer-Magn...719917&hash=item3d5762f8d6:g:4D4AAOSw4DxaajFz

How are you going to use this gauge? What pressure differential are you wanting to measure?

 

[paulmars]

"Im using it in the 0.02 to 0.03 negative and positive ranges."

 

[berkeman]

"Im using it in the 0.02 to 0.03 negative and positive ranges."

I mean, what equipment will it be mounted to? What is creating the pressure difference? Do you have an already-existing and calibrated gauge that you can use to check the accuracy of this new gauge?

 

[paulmars]

"I mean, what equipment will it be mounted to?" A flue. Is this relevant?

"What is creating the pressure difference?" An oil heater. Is this relevant?

"Do you have an already-existing and calibrated gauge that you can use to check the accuracy of this new gauge?" I am not going to answer that, because I don't want to be sarcastic.

 

[berkeman]

So you are looking for a way to check the calibration and accuracy of that differential pressure meter, without having any calibration equipment?

If you buy it new instead of off of eBay, the calibration may be guaranteed for some period of time (like a year). BTW, since this will be used in a heater/flue application, have you checked the temperature specs for the gauge? What temperatures will the two sides be near?

 

[NTL2009]

What if you connected it to a container of a known volume with two ports. Have the container half filled with liquid. It seems you could calculate how much liquid to add or remove to create a known pressure difference.

 

[paulmars]

and how to add or remove such a small and precise amount of wtr?

 

[NTL2009]

and how to add or remove such a small and precise amount of wtr?

How small depends on the size of the container. Syringe?

 

[Asymptotic]

it gives the option of selecting 'no prefix' however doing so gets error message "Please select a prefix."

What gives this option?

 

[sophiecentaur]

What gives this option?

This is all about how to start a new thread on PF, I think. Some forums have the option and others don't. But this other issue is injecting an air of confusion into the thread.

The actual question is about how to produce a (fairly small) reference test pressure difference. Ignore any of the following that's 'blooming obvious' apologies where appropriate but we should be singing from the same hymn sheet..

A U tube of water (manometer) is all that's needed, with one side connected (well sealed) to the gauge with a bleed tap / screw. You open the bleed screw and close it when the levels are equal. You then add enough water on the other side to get a level difference of whatever you choose (or just raise / lower one tube). Using a level and a scale, measure the difference in water levels in the centre of the water surface. That will have calibrated the scale on the gauge. (Needless to say, the water needs to be fresh so you can be sure of its density)

The bore needs to be quite large so that surface tension effects can be eliminated and to ensure that a reasonable amount of water flows. The "Tube" need not have uniform bore, all the way round. Two wide vertical glass tubes can be joined by a thinner tube at the bottom. (they can be of different bore too, if that's all you have. If the two vertical tubes are right next to each other and you take your measurement from a couple of metres away then parallax error can be largely eliminated. If you put the scale (ruler) on a mirror, behind the tubes , you can move your head so that the level and its mirror image are at the same level level and read the scale. Ignore the meniscus and use the lowest point of the liquid surface. I would think that you could achieve +-1mm repeatability (40 thou) which would be at least as good as your gauge.

 

[Merlin3189]

First, I know absolutely zilch about this! But was intrigued by the thought of measuring the pressure if a few thousandths of an inch of water. So, FWIW:

The only context in which I've ever come across measuring small pressures (albeit several orders of magnitude greater than this) is, as the OP is doing, measuring gas flow. So could you build a pair of venturis in the same flow, one with a very small change in flow speed and the other with a much larger change. A 10:1 change in radius should give 10^4 greater pressure difference than a 1.2:1 change. So you might compare your device with a more normal device.

I don't know if a single taper tube with (very small orifice) taps at different points along it would work as well as traditionally shaped venturis, but it might be easier to construct.

Edit:

I would think that you could achieve +-1mm repeatability (40 thou) which would be at least as good as your gauge.

Is he not wanting to measure 0.3 - 0.5 mm ?
Which is why I thought it necessary to multiply the pressure to make it measurable by conventional means.

 

[paulmars]

quoting still don't work.

"What gives this option?" starting new topic on this site

 

[paulmars]

this would be a VERY small amount of water! The inaccuracy of the water measurement would invalidate the results.

 

[paulmars]

once inch of water equals 1" of water height in a glass liquid manometer. So, 0.05 inches of water equals 0.05". try and measure that accurately w/o special equipment.

 

[CWatters]

"Im using it in the 0.02 to 0.03 negative and positive ranges."

So you are using it to measure pressure between 0.02 and 0.03 (or -0.02 to -0.03)?

That's only one division on the gauge scale. Even if the gauge is accurate you probably can't read it to within 10% Eg typical reading might be 0.02 +/- 10%.

Or have I missunderstood something?

Edit: I've had another look and think 10% is optimistic.

 

[CWatters]

once inch of water equals 1" of water height in a glass liquid manometer. So, 0.05 inches of water equals 0.05". try and measure that accurately w/o special equipment.

+1

Surface tension might affect the result.

 

[paulmars]

+/- 1/2 division.

 

[sophiecentaur]

If we're not going for a practical solution then we are on a wild goose chase. Let's face it, the picture of the device he wants to calibrate is of a second hand item on eBay.
What is the relevance of such an accuracy of a steady state (Aka DC) value for pressure. What is the recommended instrument for measuring the flue pressure? In the interest of safety and proper combustion, it could be a pretty important measurement. If it's to ensure that the flue has 'draw' then doesn't the pressure just need to be at least 0.02 in the negative direction? That's not far off the width of the needle on your proposed device.BTW I pressed the "Quote " button (bottom right of the post you want to quote) and then, in the edit box where you have been writing, there is a button under the edit window and selected the quote in the box that appears. This is what it gave me.

quoting still don't work.

"What gives this option?" starting new topic on this site

Try again. It is a tad confusing, to be sure. You can also lift other quotes from other posts if you lassoo a passage and press the little quote box that appears underneath the highlighted passage - but you may have to lassoo forwards an not backwards to be sure it's been captured.

 

[CWatters]

+/- 1/2 division.

Might not be that bad. One of my university lecturers 30+ years ago claimed that a study showed people were reasonably good at estimating to within 1/10th of a division hence my initial comment about +/- 10%. But in this case the pointer is quite fat compared to the division.

 

[sophiecentaur]

Perhaps we should all start again and behave like real Engineers and Physicists. We are not dealing with precision equipment here, we are dealing with an Oil Heater and it just needs sufficient draw in the flue to extract the gases fast enough to ensure total combustion and not to generate CO. There could be a pretty wide tolerance in the requirement for this but can we be sure? The Heater will have a type number and a manual plus spec. @paulmars Are you getting the pressure information from the maker's documentation? The couple of boiler manuals I found specify a flue gas 'velocity' (?) (figures such as 4m³/s ihttp://www.pottertoncommercial.co.uk/downloads/Rapido_F100_installation_guide.pdf ) rather than a flue pressure.
I am beginning to think that your DIY approach to checking the flue performance may not be advisable unless you can use an approved pressure meter. The cost of the right device may not be too much compared with the cost of the boiler and a year's worth of fuel. Could it be time to consider a change of plan? CO is not stuff to play about with.

 

[Merlin3189]

Having looked at the Dwyer site to see what I could find out about this machine, the obvious solution is to send it to them for re-certification.
But being a tightwad myself and liking a challenge, I've had another thought (though I still prefer the Bernouilli idea.)

If a water manometer is too difficult to read precisely to 0.001" (which it obviously is!) and we can't have a significantly less dense fluid, can we amplify the displacement to make it easier to read?
If the two vertical tubes of water were perhaps 1cm diameter and connected by a capillary tube of say 1mm diameter, then the displacement of liquid in that capillary would be 100x greater than that in the vertical manometer tubes. If we could read that displacement, by say a small bead of mercury in it, we could read it 100x more precisely.
To avoid the mercury's weight biasing the result, the capillary would need to be adjusted very close to level, then readings taken both ways and averaged.

Re. Sophie's comment. When I had a gas heater installed, the fitter obviously had to test for gas leaks and for flue draw. The first he did by painting all the joints with a special fluid (soapy water, but approved for gas testing and costing many times more than Fairy liquid) and the second by placing a small pyrotechnic in the mouth of the flue and observing that the smoke went into the flue rather than out. He assured me that in pre-CORGI days, a smouldering handful of crumpled up newspaper had sufficed.

 

[paulmars]

"Im using it in the 0.02 to 0.03 negative and positive ranges." Actually adjusting the flue to allow a little air in is to slow down the exhaust so more heat gets into the house and less out the chimney. Also to cool the exhaust (nobody wants a chimney fire).

 

[Asymptotic]

it gives the option of selecting 'no prefix' however doing so gets error message "Please select a prefix."

This is all about how to start a new thread on PF, I think. Some forums have the option and others don't. But this other issue is injecting an air of confusion into the thread.

Perhaps so. However, 'please select a prefix' suggested to me the OP may have been describing the readout of a pressure calibrator, and in this context would be relevant to the discussion.

 

[Asymptotic]

A U tube of water (manometer) is all that's needed, with one side connected (well sealed) to the gauge with a bleed tap / screw. You open the bleed screw and close it when the levels are equal. You then add enough water on the other side to get a level difference of whatever you choose (or just raise / lower one tube). Using a level and a scale, measure the difference in water levels in the centre of the water surface. That will have calibrated the scale on the gauge. (Needless to say, the water needs to be fresh so you can be sure of its density)

Ignore the meniscus and use the lowest point of the liquid surface. I would think that you could achieve +-1mm repeatability (40 thou) which would be at least as good as your gauge.

I think what @sophiecentaur described is about as close as is possible without spending more than what the gauge originally cost. If the eBay link the OP provided is the gauge in question, it's a used Dwyer Magnehelic model 2300-0 (-0.25 to +0.25 inch water column, 3% full scale accuracy, 140°F max.) that cost twenty dollars. Don't know what Dwyer charges, but sending it back and having them check & correct calibration and provide a NIST calibration certificate is probably at least twice the $20 purchase price. Checking the 0.25" w.c. full scale gauge reading against a 1/4" displacement in a U-tube manometer will at least tell you whether the gauge is wildly out of spec.

It isn't any more accurate (also 3%, 140°F max.) but may be easier to buy a Dwyer Mark II inclined leg manometer ($30-$40) and check the Magnehelic against it.

In addition to the resolution issues @CWatters outlined, 3% of full scale is +/- 0.008" w.c. (0.008" in 0.25") so obtaining accurate measurements down to 0.02" w.c. is problematic in either case.

Taking liberties with the drag race shop credo ... "Accuracy costs. How close do you want to spend?". Generally speaking, calibration requires a test instrument with better specs (that is, a gauge that costs more) than the one under test, in conjunction with a low-drift, high resolution pressure source such as the Dwyer model LPCP calibration pump (add another thousand bucks or so).

Take a look at National Bureau of Standard monograph #114 (1970), "Survey of Micromanometers". I never knew there were so many variations on this theme! If I had to build something capable of high accuracy without spending a lot on components it would be along the lines of an index point detector as described in section 2.2.

 

[sophiecentaur]

Perhaps so. However, 'please select a prefix' suggested to me the OP may have been describing the readout of a pressure calibrator, and in this context would be relevant to the discussion.

The OP seems to hop seamlessly from manometers to editing posts but I think he probably has the right tools for posting now.

I just love that "Inclined Leg" job!

 

[Asymptotic]

The OP seems to hop seamlessly from manometers to editing posts but I think he probably has the right tools for posting now.

I just love that "Inclined Leg" job!

My first experiences with using manometers was with a Dwyer Mark II. Fond memories, except for what happens with the red oil if one attempts to store it upside down :). They are also a favorite with the coal burning crowd where they are used pretty much for the application I'm imagining the OP has in mind.

 

[paulmars]

they charge 99$
"Don't know what Dwyer charges, but sending it back and having them check & correct calibration and provide a NIST calibration certificate"

 

[paulmars]

Dwyer Mark II (model 25 shown) accuracy is +/-0.09, so a reading of -0.02 could actually be anywhere between -0.11 to +0.07. I confirmed that with dwyer.

 

[paulmars]

"outlined, 3% of full scale is +/- 0.008" w.c. (0.008" in 0.25") so obtaining accurate measurements down to 0.02" w.c. is problematic in either case." The gauge I referenced is FS accuracy 2% which is 0.005. I am ok with +/-0.005.

 

[paulmars]

yes many flame heating (oil and otherwise) forum ppl suggested the dwyer mark II model 25 and I almost bought one. However, the accuracy makes it really useless. I even tried to get the manu to tell me if it is really effectively more accurate that then, just in the range that I need, after I zero it. they would not go there.

Another idea is the inclined tube manometer. 10:1

 

[Asymptotic]

"outlined, 3% of full scale is +/- 0.008" w.c. (0.008" in 0.25") so obtaining accurate measurements down to 0.02" w.c. is problematic in either case." The gauge I referenced is FS accuracy 2% which is 0.005. I am ok with +/-0.005.

Could be I'm looking at this incorrectly, but not per Dwyer bulletin A-27, and specs at the web page for series 2000 gauges. The model xxxx-0 is 3%.

Accuracy:
±2% (-HA model ±1) of FS (±3% (-HA ±1.5%) on -0, -100PA, -125PA, -10MM and ±4% (-HA ±2%) on -00, -60PA, -6MM ranges), throughout range at 70°F (21.1°C).

 

[paulmars]

maybe u right. I am unhappy with dwyer email support on other issues. I asked what the accuracy was, since their page lists 2,3, and 4%. response was for that specific unit 2%. Still, I don't like used and new costs too much. Interesting is that dwyer also lists the same 3% for mark ii mdl 25 and for mark ii mdl 40-1 of 3% FS. So, 40-1 would be more accurate. I wonder why both have same stated FS accuracy.

3% of 1.0 is 0.03 and that is not accurate enough and no one is recommending this unit and I wonder why.

 

[sophiecentaur]

However, the accuracy makes it really useless.

I cannot believe that a simple oil heater can require the sort of measurement accuracy that you suggest. If the concern is the rate of flow of exhaust gases then there will be many other factors involved than pressure. Tolerances can't be particularly tight. The flue layout spec for most boilers has some recommendations about length and bends which are pretty non demanding.
I suggest you visit a plumbing and heating forum and ask what they all do. Be prepared for a less friendly response when you actually suggest DIY for anything to do with heating. They will probably tell you to get it done 'professionally' and then go on to explain things in impossible non-Science terms. The words "idiot" and worse are to be expected if you suggest taking the bread out of their mouths but you may get the name of a suitable instrument from someone before the sky falls in.

You wrote "
" I wonder why both have same stated FS accuracy."
I think the answer could be that they just thought it up from nowhere.

 

[paulmars]

mount it inclined 10:1 for 10 times the resolution. https://www.amazon.com/dp/B0038QHTK2/?tag=pfamazon01-20

its graduated at 0.25 intervals and you can easily judge 1/5 of that. So, 0.25 included 10:1 the 0.25 would then be 0.025 and I could fairly accurately judge 1/5 of that or 0.005.

thoughts?

 

[sophiecentaur]

mount it inclined 10:1 for 10 times the resolution. https://www.amazon.com/dp/B0038QHTK2/?tag=pfamazon01-20

its graduated at 0.25 intervals and you can easily judge 1/5 of that. So, 0.25 included 10:1 the 0.25 would then be 0.025 and I could fairly accurately judge 1/5 of that or 0.005.

thoughts?

That will magnify the scale but its accuracy will depend on things like the bore being parallel with the sides and distortion through the glass sides at an angle. (And thickness of the glass)
I ask again where you got the information about the required accuracy. Or was it just the claimed accuracy of a recommended instrument? Those are two different things. I am sure you are worrying needlessly about this.