Pressure signal from DAQ keeps changing daily.

[Xiaoting]

We setup a pressure transducer to our hydraulic tensile tester and hope that the output signal can't be converted to meansure the load applied, which seems like a good and cheap soluation at the time. We bought a separate simple power supply 12v for pressure transducer and connected to DAQ board, which connects to a laptop.

Pressure transducer: MSP-300, supply 10-30 vdc, output 1-5vdc (fixed), pressure up to 5000psi

DAQ: NI USB-6210, Pressure transducer links to ai1, BUS-powered.

The initial problems we had were 1) initial voltage when the tensile machine is turned on keeps on changing daily. We've been told by our sales engineer that this is caused by the pressure loss in the hydraulic system due to daily use, which is tolerable, and can be eliminated by set up a reference zero value when the machine ram is floating and ready to load. 2) the output is very noisy and has spikes. This problem seems to be solved by adding two filter steps and take a 'mean' as the final output in the labview software.

The calibration engineer came along to help us convert pressure to force and calibrate it as well, but refused to do so, because the problem occurs again the reference zero value is no long zero, it keeps moving ups and downs, and is out of our tolerence.

I did find that the power supply of PC will influence the reading, and the reading changes when I plug and unplug the power cable from the laptop. But it must be something else as well. The whole thing is set up in a workshop full of electronic equipments.

To conclude, my questions are:

1) does anyone have the experience using a pressure transducer and NI DAQ unit, who could show me how to setup the software properly?

2) Why does laptop power supply infuence the DAQ voltage output so much? I've been told that this is because of the noise let out by many switch mode power supply units, but can DAQ filter all the surrounding noise signals, and only process the signal through the given ports. I'm not experienced in that area, but seems like that DAQ is not doing a very good job for this purpose.

3) Sale engineer now thinks that the pressure is not changing and the output of pressure transducer is not changing, which confuses me with what he said before about the pressure loss. Is anyone familiar with old hydaulic tensile tester who can help me to understand this?

Thank you for your time and look forward to any reply. 

[NI-Bongo]

Hi Xiaoting,

I have seen noise issues when trying to work with a laptop and bus-powered USB devices because of grounding issues.  The fact that plugging and unplugging the laptop changes the measurement makes me suspect this is the case.  This also may explain the reason the pressure reading changes, as the ground planes of the different system components may be drifting and causing an offset. 

I would do some experimentation with the laptop.  I assume you are using a standard laptop power supply with 2 prongs (no ground pin).  If you have a dock for the laptop that as a 3 pin power cable, I would try that and see how things change.   Also, you may try running it on battery mode, although I believe this also introduces some issues because the DAQ device is bus-powered.

Are the power supply and the laptop on the same circuit?  You may want to try isolating them if they are not or putting them on the same circuit if the already are.

See if you can deduce anything further about the problem by playing with the grounding in these ways.

Also, are you measuring the voltage from the pressure transducer using the referenced single-ended, non-referenced single-ended, or differential terminal configurations?  The differential mode should work well for cancelling common-mode noise.  In that mode, connect the transducer power supply ground reference to the AIGND channel on the DAQ device in addition to the signal connections.

Let me know what new information these suggestions provide.

Field Wiring and Noise Considerations for Analog Signals
http://zone.ni.com/devzone/cda/tut/p/id/3344

Getting Started with NI-DAQmx
http://zone.ni.com/devzone/cda/tut/p/id/5434

[Xiaoting]

Thank you for your response, John. I tried 'Differential Terminal' mode without modification of wiring. I read the information on the website, I think that I can understand why this mode would work better. However, the measurement I read from the Labview is around 10v, which can not be right, since the output range of pressure transducer is between 1-5vdc, unless I misunderstand something here. Could you please help me out here.

I would like to explain the wiring setup around the NI USB-6210.

The pressure transducer has three wires, white goes to Ai1 of DAQ for signal connection, red goes to positive of power supply, black goes to both AiGND of DAQ as ground and ground of power supply. The power supply of pressure transducer is just a simple transformer with a range of vdc supply bought from the shop and we are using 12vdc range. We set up wiring by ourselves. The cables are all less than a metre, not shielded.

I couldn't find a dock station for my laptop, but am going to try a desktop today to see whether that would make a difference.

Thank you very much and look forward to your reply.

[NI-Bongo]

Hi Xiaoting,

In order to do a differential measurement, which I recommend, you need to change the wiring.  The differential configuration uses 2 analog input channels and measures both with respect to ground, then takes the difference between the two.  For the USB-6210, when measuring in differential mode your signals should connect to the adjacent AI channels.

The White signal wire should go to AI0 (AI+) and the ground should connect to AI8 (AI-).  You can use the other pairs if you want as well (AI1 and AI9, AI2 and AI 10, and so on).  You can then also provide a return path to ground by adding resistors between both the AI+ and AI- to AIGND.  See the field wiring document linked in my previous post for more information on this.

You can also use NRSE measurement mode if you want by connecting the ground to the AI Sense on the DAQ device.  Again, refer to the field wiring document for details on this.

Let me know if this helps!

[Xiaoting]

Hi John

Thank you very much for your reply. Here is what we had done.

We changed to 'differential' mode.

We changed the wiring so that Output of Pressure Transducer is connected to Ai0 and Gound is connected to Ai9. Red wire and black ground wire are connected to power supply. Then we realized that there were lots of noise coming from THAT power supply who works like a bridge sending the noise from power line and laptop to the DAQ.

To eliminate the noise, we switched to a battery power supply (it is just 2# 9v batteries in series). This works really well. There are still spikes, but most of them are eliminated after filtering and time averaging. Here are the steps we used:

1 Acquire Analog Voltage - a simple calibration is done in this step to make sure the final reading at zero force is at minimum, here is the table

Calibrated	Uncalibrated
1	1.0604
4	4.0604

 This makes all the signals below 1.0604 = 1.0

2 Filter - Lowpass Cutoff at 1hz. I tried 100m, which doesn't make any differences.

3 Amplitude - to get DC reading

4 Time averaging - to get rid of the spikes

5 Formula - to convert the voltage reading to force

You may remember the problem I had, the reading at zero force keeps changing daily. This still occurs. The initial reading changed from 1.061v (morning) to 1.059v (afternoon).

I still can't find the reason for it. The machine has a original dia gauge which is reading the pressure and converts it to force, therefore any pressure friction loss within the hydaulic system should be compromised. The pressure transducer has a maximum error of 0.1%, therefore the 1.061-1.059=0.002>0.001.

I'm wondering whether you could help us with this, also have any comments on the setup steps we used. Thank you and look forward to hearing from you.

Xiao-ting

[NI-Bongo]

Hello Xiao-ting,

Just to be sure, the differential mode requires you to use AI8, not AI9, for the negative signal from the pressure transducer.  Definitely double check this.

Also, you want to set the input range to minimum to 1 V and the maximum to 5 V to ensure you are using the most resolution you can get out of the DAQ device.

How much pressure does the 2mV fluctuation correspond to?  Is the setup in a configuration where the ambient pressure could change slightly and cause a difference in reading?  Can you measure the voltage output of the pressure transducer with a DMM and confirm that it is indeed the same from morning to afternoon?

[Xiaoting]

Hi John

Thank you so much for your reply. We are using Ai1 and Ai9, not Ai0 and Ai9, sorry about the typing mistake.

We set the input range to 1v-5v.

2mv fluctuation corresponds to pressure of about 2.5psi.

The oil tank is not sealed, and we don't think the ambient pressure's change would cause a difference in reading. I will check the temperature's influcence later and let you know the results.

I checked the voltage output of the pressure transducer with a DMM and it is the same as the labview reading from the Amplitude and Time Averaging step. Though I can only read minimum 1mv from the DMM. I checked the voltage reading again today and Table shows the comparison with what we got last week. The pressure transducer is still runing on the batteries (2# 9v), and I just check the output voltage is 17.8v.

	Initial	Ram Floating
20 Mar am	1.011	1.061 / 1.062
20 Mar pm	1.008	1.059
25-Mar	1.002	1.053

Could you please let me know what you think about this? We are really desperate here. We cannot get the whole thing calibrated to ISO qualification if the zero-force reading has such a offset, which leads to a untolerable readings of forces.

Thank you so much for your help and look forward to hearing from you.

Kind regards,

Xiao-ting

[Xiaoting]

Hi John

I checked the temperature of oil pipe and don't think that it changes much. Thank you and look forward to your comment.

Kind regards,

Xiao-ting

[NI-Bongo]

Hi Xiao-ting,

A few ideas here:

One idea that I recommend you try first is to add some bias resistors between the differential inputs and ground (10k to 100k is usually appropriate).  See the document linked below.  The bias resistor configuration is shown Figure 11.

Field Wiring and Noise Considerations for Analog Signals
http://zone.ni.com/devzone/cda/tut/p/id/3344

Do you have another pressure transducer you could try just to ensure that the current transducer isn't effecting the problem?

Does the battery supply voltage influence the measurement at all or does it remain consistent as the batteries are drained?  It looks like the voltage trends downward, although there are only 3 data points there.  Maybe you could try using the benchtop power supply again if it is variable just to see how changing the transducer supply affects things.

Also, have you been able to remove the laptop from the system and try this with a desktop machine?  It would definitely be interesting to see if this changes your situation.

The hope is that we can isolate the source of the noise in the measurement system and deal with it.  Right now, we are still unsure of which part of the system is giving us these spikes, which can throw our averages off.  How large are the spikes compared to the signal?

[Xiaoting]

Hi John

Thank you so much for your reply. I tried your suggestions and please take a look at the results and look forward to your comments

1 Bias Resistor

I tried bias resistors (2# 1W 47k RR) and wired them just like what is shown in Figure 11. And it still shows lots of noise in raw voltage reading, fluctuation (spikes) is about 0.01v. If using the Filtering and Time Averaging, the fluctuation reduced to about 0.001v. This is when the tensile tester is off, the average voltage reading is about 1.02v-1.026v.

The reading went up from 1.002v to 1.02v after added resistors.

2 Unfortunately I don't have another pressure transducer to try on.

3 Battery power supply

I did run a trial to find out the battery supply voltage's influence, here is the readings (tensile machine is off) which were taken within 15mins:

Battery Power Voltage	Labview reading voltage
17.73	1.026
17.62	1.025
17.57	1.0245

Then I used a benchtop power supply and changed it from 12v to 23v very suddenly, the readings are 1.023v and 1.026v, which indicates the influence of 1v power supply = 0.0003v reading, which is small compared to the data in the above table.

I think I do need to find a more stable power supply, but the main error comes from somewhere else.

4 Desktop

We have been using a desktop PC since yesterday. I didn't notice any changes. Significant changes occurs when 1) change to differential mode; 2) battery power supply 3) Labview filtering and time averaging.

5 Other information

Bad news is that we just realized the labview we got is only a limited edition and the company may not be able to pay for the full license, which mean I won't be able to use Filter and Time Averaging of Labview when the time is up. Therefore, I really need to find a way to get rid of both noise and offset of average reading.

I did notice a reading change about 0.002v when I was using the multimeter, which shows how sensitive this is. I feel like that I'm losing confidence in this whole setup, but don't want to give up. I'm sure there is something I can do to find out what is wrong. Thank you for your help and look forward to your reply.

Kind regards,

Xiao-ting