Altimeter accuracy
r.daneel
Posts: 96
The overview for the Altimeter Module MS5607 on the Parallax web page indicates:
"The Altimeter Module is a high resolution (20 cm) altimeter sensor. It will give extremely accurate, altitude, pressure, and temperature readings."
Is it really accurate, or is it just precise? How accurate can the (pressure) altitude be when weather conditions change - if the is a high pressure system over my location one day, and a low pressure the next, surely this altimeter will be affected by those weather conditions? Other than change in altitude, and perhaps rate of change, I'm struggling to see the usefulness of these altitude sensors - unless I suppose I can calibrate them for local conditions at the start of every use.
"The Altimeter Module is a high resolution (20 cm) altimeter sensor. It will give extremely accurate, altitude, pressure, and temperature readings."
Is it really accurate, or is it just precise? How accurate can the (pressure) altitude be when weather conditions change - if the is a high pressure system over my location one day, and a low pressure the next, surely this altimeter will be affected by those weather conditions? Other than change in altitude, and perhaps rate of change, I'm struggling to see the usefulness of these altitude sensors - unless I suppose I can calibrate them for local conditions at the start of every use.
Comments
If you don't want to calibrate then you're stuck with pressure altitude which won't give you accurate ASL but will give you precise rate of change over time.
I would think it would be useful on a quad, given the limited area, and flight time. The initial calibration would not change that much from start to finish of flight.
I think it would be useful in "Return To Home" scenario. The altitude may not be accurate, but it would be the same when landing in the same spot.
You may be better off with a high quality GPS like a uBlox, but even GPS doesn't always do well with altitude at times.
Suppose to have 15 minutes of autonomy.
During that time local pressure won't change so much.
This is what usually happens with multirotors.
I've got the new uBlox from Parallax, and the readings will vary by 5 meters or so. GPS historically has not know to be good at altitude. (unless you in the military)
New-Altimeter-Barometer-Module
how-to-set-reference-point
Altimeter-temperature-changing-the-altitude
Altimeter-Barometer-Temperature-Anomaly
You're correct about the usefulness of the pressure sensor as an altimeter when conditions are changing. Even when conditions are stable there can be temperature inversions that depart significantly from the standard lapse rate. Also there can be local effects, for example on a mountainside that makes its own weather. The description really should not say "accurate" due to the nature of the beast. If you find yourself camping on a mountainside next to a known lake, you can recalibrate on the spot, but if you later find yourself lost next to a nondescript outcropping, the altimeter calibrated that morning at the lake might add more confusion than help. That said, I agree about using it for a quadcopter. The weather will not change much during a flight of 10 minutes, and nonliinearity and lapse rate will not be important factors. Unless maybe the quad is probing a tornado. Better find a different altimeter for that purpose!
You said it right there. "Pressure altitude" is basically your altitude above sea level on a standard day. It rarely coincides with your true altitude. The sensor (if accurate) will always give you the true pressure altitude. It is up to you to translate that into the true altitude.
Typically you calibrate on the ground before each flight. For a short flight the weather is not likely to change much.
Depends upon what you are relying on it for. If you intend to use it to know your precise distance from the ground, such as in judging a landing, then no. But if you are using it to clear an obstacle, reach a certain vantage point or to hold an altitude then yes.
-Phil
That would be pretty neat to try. The one on the ground would have a static mount. The difference between where it thinks it is and where it really is could be calculated as the altimeter setting, then beamed up to another craft. You would have to have a known good altitude to program into the ground altimeter, if that is accurate the rest should be.
You have to compensate the altimeter for local air pressure at sea-level. With a pressure-based altimeter, there's no such thing as absolute altitude.
-Phil
At the time you want to measure your altitude. Calibrating on a day with a high pressure overhead will not be useful on the following day if there is a low pressure overhead.
You should be able to make a routine where you input the known altitude then it applies the necessary correction each time it is powered up. All you would need is two buttons to increment/decrement the altitude.
That is pretty much implied at this point. So what did you want to do with this sensor?
Did you want to be able to pull up your altitude anywhere you are, like a portable project?
Making an altimeter is totally doable, if I can get it working anyone can
Measuring above ground level, or measuring your elevation change it's going to work just fine for you. You might be a foot or so off after an hour if the weather is changing, but a foot off isn't that bad for a DIY project. You also do not need a local setting for this to work.
Since you have a GPS, you should have a 3D position from that and can use it to seed your ASL, which is subsequently maintained by your altimeter.
AltimetrySnippet.spin
CON VS = 5.0 VS_RECIPROCAL = 1.0/VS VAR long SLinHg ' altimeter setting, sea level, inches of mercury DAT MSLval word -5000, 0, 5000, 10000, 15000, 20000, 25000 inHgval long 35.74440921,29.92479422,24.89931844,20.58245917,16.89480251,13.76304329,11.11998511 OBJ F : "FloatMath" { MPX4115AP circuit: ┌─────────────┐ │ │ +5V ───────────┤ ├──────┐ │ MPX4115AP │  750Ω │ │ ├────── VOUT ──── to ADC │ │  0.33µ └──────┬──────┘   } PUB Main | V, altitude { presuming you have a 10 bit ADC available somewhere, poll it for a count value, convert that to a voltage (V) in the range 0..VS and preset the sea level pressure in inches of mercury in SLinHg } V := ADC.Get(MXP4115channel) altitude := kPatoMSL(MPX4115AP(V),SLinHg) PRI MPX4115AP(V) ' static pressure (see: http://en.wikipedia.org/wiki/Atmospheric_pressure) ' Vout = Vs * (0.009*P-0.095) ± (Pressure Error x Temp. Factor x 0.009 x VS) ' Vs = 5.0 ' TA = 0 to 85°C ' 15..115kPa (2.18...16.7 psi) [4.43000591..33.9633786 inHg] ' Voff = 0.204±0.069 ' V/P = 46mV/kPa ' Vspan = 4.59V { solving for P: Vout/VS = 0.009*P-0.095 Vout/VS+0.095 = 0.009*P (Vout/VS+0.095)/0.009 = P } return F.FMul( F.FAdd( F.FMul( V, VS_RECIPROCAL ), 0.095 ), 111.1111111 ) 'kPa (kilo Pascals) PRI kPatoMSL(kPa,SLP) | inHg, i, MSL inHg := F.FMul( kPa, 0.295333727 ) repeat i from 0 to 6 if F.FCmp(inHg,inHgval[i]) => 0 quit MSL := MSLval[i-1] + F.FTrunc( F.FMul( F.FSub( 1.0, F.FNormalize(inHg,inHgval[i],inHgval[i-1]) ), 5000.0 ) ) MSL += F.FTrunc( F.FMul( F.FSub( SLP, inHgVal[1] ), 1000.0 ) ) return MSL ' mean sea level referenced altitude (linear interpolated) PRI kPatoInHg(kPa) return F.FMul( kPa, 0.295333727 )