Communication between sensor board and Microcontroller. What hardware to use?

[Erik_]

Hello everybody,
I'm in need to connect a sensor board with an SPI interface to a Teensy microcontroller. Unfortunately, the microcontroller is 3 meters away from the sensor board. SPI link seems not appropriate to handle such a distance.
What can I possibly use? New hardware?

Thanks in advance

 

[Nigel Goodwin]

Why do you think 3m is too far?, what is the sensor?, and how fast do you need to be reading it?

 

[Erik_]

Maybe it works, but I can't test it on hardware. For this project, I need a really robust data link.

It's a work in progress. Pressure sensors are 5x**broken link removed** by All Sensors + N.1 barometric pressure sensor. Data acquisition is carried out with an ADAS3022BSTZ multichannel data acquisition IC. The SPI communication interface is provided by the ADAS3022BSTZ.

The desired frequency is 50 Hz.

 

[MaxHeadRoom78]

I don't see why you cannot use it over 3m, just use the right kind of cabling and shielding.
Max.

 

[Nigel Goodwin]

Simply run the SPI at a low clock rate, and 3m should be no problem whatsoever.

 

[Erik_]

That's great news! I will proceed with a plain SPI.

Thank you all.

 

[gophert]

That's great news! I will proceed with a plain SPI.

Thank you all.

Teensy is 3.3V. Make sure your sensor is able to send 3.3V or use an appropriate level shifter.

 

[tomizett]

Also, consider buffering your outputs with proper line drivers and your inputs with Schmitt triggers (along with some filtering).

If you're doing the SPI in software, you could take multiple samples across the period of each data bit and check that they agree - or simply take the majority.
There's no scope for doing checksums (since it's just raw data) but you could do some simple validation checks - for example, if it's barometric pressure you'd not expect it to change quickly, so you could discard any samples that did not fit the general trand.

Just some ideas!

 

[Erik_]

gophert ADAS3022BSTZ is able to handle 3.3V comm
tomizett Have you an idea about dedicated IC/HW?

 

[Nigel Goodwin]

Also, consider buffering your outputs with proper line drivers and your inputs with Schmitt triggers (along with some filtering).

A lot of effort for no gain - he's only talking about 3m (not 3 miles) - the sensors are low speed, simply using a low speed SPI clock would allow far more than 3M with no bother (and is a simple software change - just change one value in the code). As Max suggested above, I'd use a screened cable - but I wouldn't ever have considered not doing so

 

[tomizett]

As I said, "consider" - just some ideas, certainly not mandatory. It depends of this is a one-off or a product, what the electrical/physical environment is and how often it is likely to be connected and disconnected.
But screened cable, yes, certainly.

 

[Erik_]

Thank you all for your suggestions.

Firstly, this device is not for a commercial application.

As tomizett report maybe the point is that some information is missing. Hypothetically, the device is a part of the next generation of ADC/ADS from Basic Air Data. You find the former release here. Applications of the device include performance biking, HVAC, breath metering, metering, failure detection, user custom applications, and airborne aerodynamic identification(the probe was tested within a wind tunnel).

In this winter period (northern hemisphere) we are reviewing the requirements for the next release.

Regarding the cable, a shielded ethernet cable can do the job? I think it will be a good choice for the cable under the aspect of availability.

 

[gophert]

If it is for an aircraft application, light weight may be helpful and Ethernet cable is not exactly weight efficient for two signal cables and ground to travel 3-meters (also, Ethernet has 8-wires). You gah get away with much lighter gauge wire. Check DigiKey for the great search filerting tool when you search cable or wire.

 

[Erik_]

gophert Yes, lightweight is better. I try to use only widely available and standard material.
We are now evaluating an alternative to the original circuitry. The idea is to give more calculation and communication power at the sensor board side, so we are evaluating a sensor board with an onboard microcontroller connected to the main ADC via a USB link. In this way, it is possible to continue to use the older AirDC units.