Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Help to understand the sensor output and interfacing

Status
Not open for further replies.
D

doncarlosalbatros

Full Member level 2
Joined
Nov 26, 2016
Messages
143
Helped
1
Reputation
2
Reaction score
1
Trophy points
18
Activity points
1,758
I have 0-5V analog ratiometric pressure sensor called --->HSC DRRN004DAA5, the details are on page 13 here:
https://sensing.honeywell.com/honey...oard-mount-pressure-sensors-50099148-a-en.pdf

Is this part differential signalling?
I want to send the voltage output of this sensor 75 meters far away to a truly isolated diff. input of a data acquisition board by using STP cable.
Do I need a diff line driver or is it already diff signalling?
And any recommendations for the power supply part?

I found another data-sheet: https://www.tme.eu/gb/Document/6366ba3cef7c9cebdf13e5093e192285/HSC-analog.pdf
 
Last edited:

It is not differential, it is an analog output.

Good luck transmitting that voltage 75 meters. If you'll notice, these are identified as BOARD MOUNT pressure sensors. That should give you a clue that they are not intended for transmission over 75m.
 

barry said:
It is not differential, it is an analog output.

Good luck transmitting that voltage 75 meters. If you'll notice, these are identified as BOARD MOUNT pressure sensors. That should give you a clue that they are not intended for transmission over 75m.
Click to expand...


Thanks for the reply. What do you recommend for transmission?(unfortunately for some reason the transition has to be analog so I cant use ADC):

For analog transmission which one do you recommend at the sensor output?:

1-) Differential line drivers?

or

2-) Converting to current by such module?:
https://docs-emea.rs-online.com/webdocs/15aa/0900766b815aa3c7.pdf
 

These components are available with an SPI or I2C interface. I would recommend one of those with appropriate line drivers.
 

That requires a microcontroller and/or another interface like RS484.
Max length for I2C is 18meter and SPI 10m.
 

You want to send an analog voltage over 75m not digital data. Provided the current drawn is low you should be able to use a screened cable and preferably a low-pass filter at the measuring end to remove any residual noise pick up. The response of that unit is very slow (1mS) so the frequencies down the cable will be low. There may be an issue with it driving a capacitive load (the cable inner to screen capacitance) in which case a buffer amplifier may be needed at the sensor end.

If you want to convert it to serial data format, differential is the way to go. RS485/422/432 should all work to >1Km with no problems.

Brian.
 

betwixt said:
You want to send an analog voltage over 75m not digital data. Provided the current drawn is low you should be able to use a screened cable and preferably a low-pass filter at the measuring end to remove any residual noise pick up. The response of that unit is very slow (1mS) so the frequencies down the cable will be low. There may be an issue with it driving a capacitive load (the cable inner to screen capacitance) in which case a buffer amplifier may be needed at the sensor end.

If you want to convert it to serial data format, differential is the way to go. RS485/422/432 should all work to >1Km with no problems.

Brian.
Click to expand...

Thank you for your reply. If I convert voltage to current by a module like this: https://docs-emea.rs-online.com/webdocs/15aa/0900766b815aa3c7.pdf and send the signal as 4-20mA current, wold that be better than sending as voltage?
 

Hi,

If you really want to transmit analog signals, then I recommend "quasi symmetric" signalling over the cable.
Both signals should be transferred via twisted pair shielded cable. 2 twisted pairs: one for supply, the other for signals.

Sensor output --> R-C-R filter --> cable --> common mode filter --> RC filter --> difference amplifier
Sensor GND --> R-C-R filter --> cable --> common mode filter --> RC filter --> difference amplifier

The key to success is the symmetric signal flow with identical impedances.
Then external influence should be equal to both signals ... and thus eliminated by the difference amplifier.

Klaus
 

KlausST said:
Hi,

If you really want to transmit analog signals, then I recommend "quasi symmetric" signalling over the cable.
Both signals should be transferred via twisted pair shielded cable. 2 twisted pairs: one for supply, the other for signals.

Sensor output --> R-C-R filter --> cable --> common mode filter --> RC filter --> difference amplifier
Sensor GND --> R-C-R filter --> cable --> common mode filter --> RC filter --> difference amplifier

The key to success is the symmetric signal flow with identical impedances.
Then external influence should be equal to both signals ... and thus eliminated by the difference amplifier.

Klaus
Click to expand...

For the R-C-R filter would resistors R = 100Ohm and cap = 10nF work? (Max freq of interest is 100Hz)
 

Hi,

at the source (sensor) side:
Yes, 100R / 10n / 100R is a good sart.
The output impedance of the sensor should be much smaller than the 100R.

at the reciver side:
it could be higher impedance, maybe 10k.
And if you want to limit bandwidth to about 100Hz (and thus increase S/N ratio) then you may use about 100nF

Klaus
 

KlausST said:
Hi,

at the source (sensor) side:
Yes, 100R / 10n / 100R is a good sart.
The output impedance of the sensor should be much smaller than the 100R.

at the reciver side:
it could be higher impedance, maybe 10k.
And if you want to limit bandwidth to about 100Hz (and thus increase S/N ratio) then you may use about 100nF

Klaus
Click to expand...

Thanks and do you have any suggestions regarding the voltage for the excitation/supply for the IC?
I plan to send voltage to the sensor from 60 meters far away and put a 5V linear voltage regulator to filter out the noise.
I will measure the excitation anyway, but for Wheatstone bridge in this case should I use a voltage regulator or voltage reference for less noise and better stability?
 

Hi,

but for Wheatstone bridge
Click to expand...
is it a passive wheatstone bridge? I don´t think so. It has built in amplifier with single, buffered output.

Yes, you should use a linear regulator at the sensor side of the cable.
Select a regulator with high initial accuracy and low temperature drift.
Filtering is required and LowESR capacitors to suppress noise. Use a bulk capacitor + ceramics capacitor at the input side of the regulator and a ceramics capacitor at the output side of the regulator. Mind that the regulator error will have 100% impact on the sensor output. --> 5% in input voltage change will result in 5% sensor reading change.

Your part number is not complete
DRRN004xDAA5
Please select the correct type.

Klaus
 

KlausST said:
Hi,


is it a passive wheatstone bridge? I don´t think so. It has built in amplifier with single, buffered output.

Yes, you should use a linear regulator at the sensor side of the cable.
Select a regulator with high initial accuracy and low temperature drift.
Filtering is required and LowESR capacitors to suppress noise. Use a bulk capacitor + ceramics capacitor at the input side of the regulator and a ceramics capacitor at the output side of the regulator. Mind that the regulator error will have 100% impact on the sensor output. --> 5% in input voltage change will result in 5% sensor reading change.

Your part number is not complete
DRRN004xDAA5
Please select the correct type.

Klaus
Click to expand...

Yes 5% in input voltage change will result in 5% sensor reading change. But I will make radiometric connection so I will monitor the voltage output of the regulator(IC's excitation voltage) as well. The ratiometric measuring should cancel out that error, don't you think so?

- - - Updated - - -

KlausST said:
Hi,


is it a passive wheatstone bridge? I don´t think so. It has built in amplifier with single, buffered output.

Yes, you should use a linear regulator at the sensor side of the cable.
Select a regulator with high initial accuracy and low temperature drift.
Filtering is required and LowESR capacitors to suppress noise. Use a bulk capacitor + ceramics capacitor at the input side of the regulator and a ceramics capacitor at the output side of the regulator. Mind that the regulator error will have 100% impact on the sensor output. --> 5% in input voltage change will result in 5% sensor reading change.

Your part number is not complete
DRRN004xDAA5
Please select the correct type.

Klaus
Click to expand...

Yes 5% in input voltage change will result in 5% sensor reading change. But I will make radiometric connection so I will monitor the voltage output of the regulator(IC's excitation voltage) as well. The ratiometric measuring should cancel out that error, don't you think so?

And being voltage regulator or voltage reference does it matter in this case if has enaough current for the IC?
 

Hi,

Yes, then you need an extra pair of wires and the extra filters ... but how do you process the two analog signals at the receiver side...to get the true sensor value?
Now you need 4 channels of very symmetric characteristic...

As others mentioned above: I also recommend to go to digital communication.
Otherwise you need high effort to generate, transmit and process analog signals .... but they will never be that exact as the digital transmitted values.
I can´t see the benefit. But for sure you are free to what you want.

Klaus
 

KlausST said:
Hi,

Yes, then you need an extra pair of wires and the extra filters ... but how do you process the two analog signals at the receiver side...to get the true sensor value?
Now you need 4 channels of very symmetric characteristic...

As others mentioned above: I also recommend to go to digital communication.
Otherwise you need high effort to generate, transmit and process analog signals .... but they will never be that exact as the digital transmitted values.
I can´t see the benefit. But for sure you are free to what you want.

Klaus
Click to expand...

Digital communication is not possible from the whole system point long story. Yes for each channel 2 wire. 3 STP pairs. One pair for power supply and its ground, one pair for signal our and ground, and one pair for Vexc and ground. All will be in a CAT6 cable which is F/FTP- All grounds are isolated from each other at receiver end.

I will use F/FTP cable STP type with 3 pairs very good noise immunity. Processing is just about ratio of (Vexc/Vout) times a constant. Equation can be find from the data sheet.

The signals will go into signal conditioning modules with LP filter and truly isolate the input and output with a isolation buffer. You can see the board here https://www.mccdaq.com/usb-ethernet-data-acquisition/SC-1608-Series.aspx Each channel is truly isolated. Each channel is differential ended inputs.

But what do you think about this voltage reference at the sensor end for sensor excitation/power?:
https://docs-emea.rs-online.com/webdocs/077f/0900766b8077ff69.pdf



REF195GPZ
 

Hi,

But what do you think about this voltage reference at the sensor end for sensor excitation/power?
Click to expand...

You
* either need a stable excitation voltage
*or an excitation voltage readback.
Both is overkill in my eyes.
Now as you read back the excitation voltage .. there is no need for an excessively stable excitation voltage anymore.
The ratiometric calculation at the receiver side will compensate excitation voltage fluctuation.

Klaus
 

Consider also sending the signal as an AC frequency. You'll want a voltage-to-frequency converter. The receiving end has a frequency-to-voltage converter. This is similar to FM modulation. (Or amplitude modulation might be a possibility).

Or as a voltage lasting for a certain length of time. Example, if the sensor reads 3 V, send a voltage lasting 0.3 seconds. Then drop the line to 0V, to wait until the next sensor reading.

For both the above methods, it is okay if the amplitude drops. Impedance is not an issue. A certain amount of noise is not a problem.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top