Pressure measurments... How to???

[jaydee]

What would be the ideal pressure measurement to take and what kind of equipment can do it? What kind of gauge or transmitter would work?

This is one of the main missing pieces to my bench that I have little knowledge about...

[MaxxxRacer]

Well i got a very nice gauge from a british company. sadly i cant find the company anywhere online except for a few exerpts. This one has .001 resolution measured in bars.. it also has two inputs and does pressure differential. if you dont hook one up u just get pressure relative to ambient which is nice becuase it is self calibrating to ambient.

um back to something useful to yourself. I noticed that pH uses a pressure differential meter. the one he uses looks alot like the one below. either way rosemount sells units for this purpose.

http://cgi.ebay.com/ws/eBayISAPI.dll...513380956&rd=1

[Incoherent]

I would note that it is important that it is a differential measurement. Relative to atmosphere will not give you the info you need about pressure drop unless you measure on both side of e.g the block relative to atmospheric and take the difference. Then you need two gauges

[MaxxxRacer]

yah. like for taking pressures from the pump for PQ curves you dont want the absolute pressure but the differential between the inlet and outlet pressures.

[jaydee]

So with a unit like the one Maxx linked (pics attached) I would plumb water in on one side of the unit and water out on the other side of the unit? This gives a differential measurement?

[RaptorRaider]

I think you have to connect it parallel to the pump; just get a t-line next to the outlet and a t-line net to the inlet.

Not sure that thing is what you would need though.

Personally I'd get something like this. Probably less accurate but cheaper, smaller and shows PSI instead of bar.

[jlrii]

How much accuracy do you need?

[pHaestus]

The people telling you that analog gauges are the way to go are feeding you a line of crap Jaydee. If your instrument has the resolution to do this sort of testing then it will pick up the oscillatory nature of the pump's action. The dP xmitter you linked looks pretty much like mine. You use a dmm to measure current (4-20 mA) which linearly corresponds to dP. Need good DMM to do it. You can probably find one that has LCD output as well for extra $$. Make sure you choose one that has the proper range for your test setup, eh?

[bigben2k]

Yeah, water-in and water-out. You really want to measure the pressure drop across the block (the pump is irrelevant).

Once you have a flow measurement, you'll then be able to fully characterize the block, in terms of restriction. BTW, we tend to test at 0.5, 1.0, and 1.5 gpm, but there seems to be an increasing need to measure at 0.25 gpm; some of the latest blocks run well at low flow and high pressure.

As pHaestus put it, the analog gauges are useless. First, you have to find one with a decent resolution, then you'd have to have two of them (inlet and outlet) then calculate the difference. That of course doubles your error margin, and each unit usually starts off at ~5%, leaving you with a total error of 10% (pretty useless). Each unit measures the pressure relative to atmosphere, hence the calculation.

For the sake of convenience (because water block testing can be very involved), we use a digital unit, as pHaestus suggested. The units are available as absolute or differential (or both). We want differential. Again, for the sake of simplification, if you have the ability to log the data (which requires some equipment), it'll save you quite a bit of time.

pHaestus also made a reference to fluctuations; these are indeed induced by the pump. Bill worked his way through some fluctuations, but little of his work is documented (OC forums?!?). I'll share my work, as I get to it. These fluctuations will occur at various frequency ranges, as I remember in the 120 Hertz and below. That figure is important, as it defines a parameter when you pick your dP meter; it should have a sample rate that allows you to visualize these fluctuations. 500 samplings per second or more would be a minimum, IMO.

I'll be adapting a Honeywell unit to my rig. Let me know if you want more information. Digi-Key sells it for ~$50, as I remember.

[Brians256]

Why would you want to sample at that high of a frequency? You don't want to accurately reproduce the pressure fluctuations, do you? You want an averaged approximation to the steady state pressure differential.

[pHaestus]

All I was saying was that it's practically impossible to take average measurements accurately by eye if they are fluctuating as the pump cycles. It's even worse if the readout is analog, and still worse yet if you're using 2 absolute pressure gauges which are analog. I do an integration over 1 or 2 seconds via the DMM and dump that average value into a spreadsheet automagically.

[bigben2k]

The unit is digi-key p/n 480-1919-ND. Honeywell p/n is ASDX005D44R.

The specs show an input voltage of 5 volts and a response time of 8 ms. The Honeywell site puts the response time as 8 to 11 ms. This is a bit misleading though: 8ms is the time required to go from 0 to 5 psi (for this unit), but is only for "10% of 90% of rise time". It otherwise states nothing about a sampling rate, specifically. at 11 ms, this makes for a sampling rate of 91 per sec.

Output is in volts, with a max 2ma current draw.

Accuracy is 2% of span.

[pHaestus]

I use the Rosemount 1153 xmitter wired into my HP 3478A DMM. I think it's something like 0.25% accuracy over the span. It's basically the same instrument as the 1151 but built to nuclear spec. Seems like a pretty nice unit to pH

[MaxxxRacer]

sorry i didnt realize that one was a analoge unit. Personally i use a nice digi unit.

[jaydee]

Thanks guys. I wasn't originally intending to go with such a dp meter but the prices on e-bay are pretty decent. IF I ger the flowmeter I am bidding on then I will get something with the 4-20ma output as the flowmeter has the same. Then I just aswell use the output on my thermometer and aquire another thermometer with the same output and get some kind of logging software and move towards automation.

[jaydee]

Originally Posted by pHaestus

Make sure you choose one that has the proper range for your test setup, eh?

Speaking of that, what specs should I be looking for? What ranges?

[Long Haired Git]

S'funny, was just trying desperately to find PQ graphs for common components.
Looking forward to you getting your guages, jaydee...

[MaxxxRacer]

0-15 psi is more than ull ever need. like my WMD-20RLT does ~7psi akaik so u wont ever use a pump with more than 15

14psi = 1atm..

[jaydee]

Ok, thanks. That was about what I was thinking but wasn't 100% sure.

[dnkroz]

Hi.

Firstly, sorry my bad English. I cannot be expressed as it wanted.

I am using 4 sensors of Honeywell, concretely is a 40PC series (0-15PSI, Vout is 266.6mV/PSI) is compensated and amplified. Previously it had used another model nonamplified, 26PC series.

With 40pc series, I have only needed an a power supply of 5V and a I read the Vout with a datalogger of 10bit. I don´t need more amplification in order to increase the resolution. The system still is in phase of verification

The sensors are this:



I calculate the differential pressure by sofware, but it´s posible reading Vout with a multÃ*meter and to obtain the subtraction between the sensors by means of op amps

I have obtained good resolution, aprox 0.002 mH2O, but since it has commented, the measurement fluctuates enough according to my measures. I have needed to make average of several values. For example:

This is a graph obtained with a normal pump of aquarium, 220V ac. The sensor are in water out of pump.

value of the signal



amplitude of the wave, 10 msec aprox



In this, the sensor are in the water out of WB, the singal has diminished and modified.



In this, is with a mcp350 in water out of pump. With a diferent rotor and motor of pump, the value of the signal diminishes.



Graph or resolution

[MaxxxRacer]

nice work there. looks interesting. From the looks of it ur sensors boards are handmade. any way to ensure the accuracy all components?

[dnkroz]

Hi.

If you refer to the components of the sensor, is not very important, these only they have condensers as filters, for sensor and Vcc regulator. The Vout of sensor go directly to the ADC.

If you refer to the components of the complete system, all the channels for pressure, flow or temperatures, they can be calibrated for a linear or not linear response by sofware. Also it is possible to calculate other parameters in real time, for example, sums, subtractions of channel, convert different units, calculations between different channels, for example to calculate the power lost or generated in W with a channel of flow and differential pressure, heat in w in the water (unsuccessful experience) velocity of the water, and other things that they can imagine and that the program supports, frecuencymeter to measure the sensor flow, measurement of the value every msec.

For example, the magnetic field of the punp without amplification y detected here.



Before testing, I calibrate sensors of pressure to 0. Normally when I begin the measurement´s of pressure, I can read approximately 0.002-5 mh2O out of scale

Later, I trust in the precision and lineality of the sensor and I have used a column gauge to compare the results.

On the beginning of the system, still I am verifying, it is possible to see information here, sorry in Spanish.

One question please. ¿Where can I see a pressure drop vs flow graph of a MCW6002A?

[MaxxxRacer]

i dont know of one actually. the mcw6002 that is. when i review the block i will create one though.

[redleader]

Originally Posted by dnkroz

Hi.

Firstly, sorry my bad English. I cannot be expressed as it wanted.

I am using 4 sensors of Honeywell, concretely is a 40PC series (0-15PSI, Vout is 266.6mV/PSI) is compensated and amplified. Previously it had used another model nonamplified, 26PC series.

With 40pc series, I have only needed an a power supply of 5V and a I read the Vout with a datalogger of 10bit. I don´t need more amplification in order to increase the resolution. The system still is in phase of verification

The sensors are this:



I calculate the differential pressure by sofware, but it´s posible reading Vout with a multÃ*meter and to obtain the subtraction between the sensors by means of op amps

I have obtained good resolution, aprox 0.002 mH2O, but since it has commented, the measurement fluctuates enough according to my measures. I have needed to make average of several values. For example:

This is a graph obtained with a normal pump of aquarium, 220V ac. The sensor are in water out of pump.

value of the signal



amplitude of the wave, 10 msec aprox



In this, the sensor are in the water out of WB, the singal has diminished and modified.



In this, is with a mcp350 in water out of pump. With a diferent rotor and motor of pump, the value of the signal diminishes.



Graph or resolution

Maybe I'm not understanding how you have the sensors connected, but why is there a 100Hz sin wave in there?

Also, you have a lowpass wired up before the ADC right? Lowering the cuttoff might help those spikes you're getting. That or you could do it in software.

[bigben2k]

He's in Spain, so he's running on 50 Hz. 100 Hz is a resonant frequency.