Page 5 of 5
Re: NXC: Mindsensors PPS35 - ADC values to bar?
Posted: 17 Nov 2011, 20:09
by HaWe
thx, I'll try it tomorrow (my daughter now wants me to join her watching "High School Musical 3")
SensorMSPressure is almost properly scaled, but lacks necessary offset.
what means: it's faulty...?
But in the end only a calibration with a real barometer will be able to resolve the uncertainty. I would really like to know why the hardware manufacturer has't got that executed properly yet (providing a correctly working calibration formula).
Re: NXC: Mindsensors PPS35 - ADC values to bar?
Posted: 17 Nov 2011, 20:13
by mattallen37
Well, I emailed Deepak (from mindsensors), and I'll see what he says. Indeed, a "real" pressure sensor would be very good to compare to.
Re: NXC: Mindsensors PPS35 - ADC values to bar?
Posted: 18 Nov 2011, 15:17
by HaWe
IMO the formula in the MS pdf provided with the graph is completely wrong.
If I take 2 points and calculate the formula y=mx+n then I get (roughly, as a first approach):
(note that ordinate and abscissa in the graph are reversed!)
test:
RAW=896 => psi=9.7 (10)
RAW=384 => psi=31.6 (32)
that's not prefect but more or less plausibel...
but of course I still don't know if those psi values will really match to
my RAWs (what ever which "RAW" will be, will be...)
Re: NXC: Mindsensors PPS35 - ADC values to bar?
Posted: 18 Nov 2011, 18:22
by mattallen37
Deepak from mindsensors wrote:There were several models of PPS35, with different characteristics. Unfortunately, we did not use version numbering in PPS35 models, but whenever new model was introduced, the doc was updated with appropriate formula and graph.
The problem happens now, that people are still using all those sensor models, but don't have the corresponding doc.
I wish I could answer your question accurately, but we also don't have all the sensors and their corresponding docs.
So there you have it. I guess your best bet it to just settle for "good enough" or else use RAW values. I personally would probably go with RAW values.
Re: NXC: Mindsensors PPS35 - ADC values to bar?
Posted: 18 Nov 2011, 18:53
by HaWe
I need to have a "real calibration" actually by hPa ([800;900] hPa for one certain application, [1000;1300;1500] hPa for another).
And the statement of Mindsensors is just one thing: embarrassing.
Re: NXC: Mindsensors PPS35 - ADC values to bar?
Posted: 18 Nov 2011, 22:07
by mightor
Helmut,
Do you not have the Dexter Industries pressure sensor as well?
- Xander
Re: NXC: Mindsensors PPS35 - ADC values to bar?
Posted: 19 Nov 2011, 08:34
by HaWe
no, I got 1 Dexter sensor which was defective (leaky) after 3 weeks, and then I got another free Dexter pressure as a refund, this was also leaky after a couple of weeks, and then I got the MS sensor as a free refund (by my distributor in the Switzerland).
Re: NXC: Mindsensors PPS35 - ADC values to bar?
Posted: 30 Nov 2011, 13:34
by HaWe
I now purchased a 60645 Lego® Pneumatic Technic Manometer
took a few readings (fractionals estimated):
Code: Select all
996 0
997 0
971 1
966 1.2
955 2
929 3.5
903 5
855 7.7
859 7.3
862 7.2
790 11.2
770 12
676 17.8
559 25
488 29
442 31.5
and got this linear regression formula:
Code: Select all
PSI= -0.0574 * SensorRaw(port) + 56.8
Correlation coeff = (r²) = 0.99939
about 1 psi (≈ 70hPa =0.07bar) is the lowest visible value so I don't know the reliability of values below this
about 30 psi (≈ 2000hPa =2bar) is the max pressure with my Lego pump.
Hope that could help also someone else.