Post by thread:Barometric pressure sensing stability

Piclist Gurus, I am trying to interface an MPX5100 absolute pressure sensor to a PIC via an LTC1298 12 bit A/D, where I require an accuracy of +/-2ft between ref. ground level and 1000 ft.. Using the CCS compiler (can't cope with asm!) and the example prog. and header written for the LTC1298, I can't get a stable reading off the A/D. I have seen various examples of this on the 'net, especially 'Barom.c' (http://www.phanderson.com)which reads pressure and wind speed, but can't get this to compile. As an aside, I have tried voltage to frequency conversion, then measure pulses as well, which also proved to be unstable. I have no doubt that the problem is my maths/code, and not the hardware, as a DVM on the measured output on all the schemes I have tried is solid. Any help/pointers greatly appreciated - I need to sleep! Andy Holdaway

Andy, Since I'm playing in this area right now, I would be happy to help with the C code. Please send me (offlist) the exact error message you got with Barom.c and I will see what I can do. I will be meeting with Peter next week at his workshop and if there is a general problem, it would be good to know for all. Dan Simple CCS Code at: http://204.233.101.40/robots/pic1.html >Piclist Gurus, >I am trying to interface an MPX5100 absolute pressure sensor to a PIC via an {Quote hidden}>LTC1298 12 bit A/D, where I require an accuracy of +/-2ft between ref. >ground level and 1000 ft.. >Using the CCS compiler (can't cope with asm!) and the example prog. and >header written for the LTC1298, I can't get a stable reading off the A/D. >I have seen various examples of this on the 'net, especially 'Barom.c' >(http://www.phanderson.com)which reads pressure and wind speed, but can't get this >to compile. >As an aside, I have tried voltage to frequency conversion, then measure >pulses as well, which also proved to be unstable. >I have no doubt that the problem is my maths/code, and not the hardware, as >a DVM on the measured output on all the schemes I have tried is solid. >Any help/pointers greatly appreciated - I need to sleep! >Andy Holdaway

Do what the DVM does and integrate the signal. Maybe you need a low-pass filter, check the pressure sensor data sheet and take a look out the output on a scope. On Thu, 8 Jul 1999, Andy Holdaway wrote: {Quote hidden}> Piclist Gurus, > I am trying to interface an MPX5100 absolute pressure sensor to a PIC via an > LTC1298 12 bit A/D, where I require an accuracy of +/-2ft between ref. > ground level and 1000 ft.. > Using the CCS compiler (can't cope with asm!) and the example prog. and > header written for the LTC1298, I can't get a stable reading off the A/D. > I have seen various examples of this on the 'net, especially 'Barom.c' > (http://www.phanderson.com)which reads pressure and wind speed, but can't get this > to compile. > As an aside, I have tried voltage to frequency conversion, then measure > pulses as well, which also proved to be unstable. > I have no doubt that the problem is my maths/code, and not the hardware, as > a DVM on the measured output on all the schemes I have tried is solid. > Any help/pointers greatly appreciated - I need to sleep! > Andy Holdaway >

Definitely low-pass (1/4-second to 1/2-second Tau, RC filter at least) the signal. Or active filter it. There can be a surprisingly large AC component in wind (Some geophysical researchers are thinking that causes some of the under-snow temperature rises they're trying to figure out) - I did a sensor package for this that was actively filtering the output differential put out by an Atmospheric differential sensor, inputs were atmospheric pressure and the output of a mechanical low-pass filter (large chamber with tiny orifice to outside air), to try to FFT the power at various frequencies. Some think Avalanches are enhanced by air being pumped in & out of the pores in snow, causing friction heating. I know that I sure saw some interesting temperature changes in buried thermistors when weather fronts hit in the Arctic, while up there installing that beast... Amazed the heck out of me, I'm "somewhat of a believer" due to the effects I've seen. Mark Jamil J. Weatherbee wrote: {Quote hidden}> > Do what the DVM does and integrate the signal. > Maybe you need a low-pass filter, check the pressure sensor data sheet and > take a look out the output on a scope. > > On Thu, 8 Jul 1999, Andy Holdaway wrote: > > > Piclist Gurus, > > I am trying to interface an MPX5100 absolute pressure sensor to a PIC via an > > LTC1298 12 bit A/D, where I require an accuracy of +/-2ft between ref. > > ground level and 1000 ft.. > > Using the CCS compiler (can't cope with asm!) and the example prog. and > > header written for the LTC1298, I can't get a stable reading off the A/D. > > I have seen various examples of this on the 'net, especially 'Barom.c' > > (http://www.phanderson.com)which reads pressure and wind speed, but can't get this > > to compile. > > As an aside, I have tried voltage to frequency conversion, then measure > > pulses as well, which also proved to be unstable. > > I have no doubt that the problem is my maths/code, and not the hardware, as > > a DVM on the measured output on all the schemes I have tried is solid. > > Any help/pointers greatly appreciated - I need to sleep! > > Andy Holdaway > >