SparkFun Forums

[integ3r]

The HSCDANN001BA2A3 is a very accurate absolute pressure sensor from Honeywell with an I2C output and an attachment for an optional hose (the HSCDNNN001BA2A3 has no hose attachment). It has applications in almost everything, especially high altitude balloon projects because of its very accurate range of measurement over a wide range of pressures.

These are great little sensors, but unfortunately you can't find them ANYWHERE unless you order over 25. Maybe SparkFun could order some of these? They don't even really need a breakout board since they're DIP (although Honeywell DOES make SMT equivalents).

I am posting because I have tried acquiring one of these from a parts supplier and they are asking me to order at least 10. -.-

All of the pressure sensors SparkFun sells aren't rated up to very high altitudes (the MEMS ones are rated to 30 kPa, making them unsuitable for a high altitude balloon that needs accurate pressures for various reasons). These ones can go from a 100 kPa to a vacuum, giving them a pretty wide range and some great accuracy.

What do you think? These would be great for SparkFun to have as an addition to their existing pressure sensors. Plus, they're absolute pressure sensors, which makes them that much more suited for high altitude readings.

EDIT: Honeywell Silicon Pressure Sensor Guide: http://sensing.honeywell.com/index.cfm?ci_id=148494
HSC Datasheet: http://sensing.honeywell.com/index.cfm?ci_id=157750
I2C Communication Docs: http://sensing.honeywell.com/index.cfm?ci_id=156988

[trialex]

+1 vote from me

[Daniel Wee]

The sensor is only rated down to 0-degrees Celsius so that's not going to be enough to proper high-altitude work. On top of that, it only goes down to 1 psi. You really want something lower than that, with a whole lot of resolution (at least 16-bits) from 0 to 1 psi. I don't think any of the available barometric sensors will work for real high altitude applications.

Daniel

[nickersonm]

Those sensors all go down to vacuum; the "1 psi to 250 psi" in the title of the datasheet are the available ranges, i.e. 0-15psi, 0-1 bar, etc. Pages 5 and 6 of the datasheet have details on that. Output is 12 bits over the range; you could use multiple sensors for better accuracy at low pressures if desired.

The sensors are also rated to operate down to -20C, although they aren't compensated below 0C. Presumably you could keep the pressure sensor heated sufficiently for operation in colder conditions; I imagine the rest of your electronics would not be happy at high-altitude ambient temperature, either.

Overall, it looks like a nice sensor line - I might buy a few if SparkFun listed them.

[integ3r]

Yeah, I forgot to mention that. They all go down to a vacuum. The ranges are just 1-250 PSI for the maximum end. These are REALLY top-notch sensors.

Regarding the low temperatures... my team has thought of ways to keep the balloon at a reasonable temperature through the flight. We've obviously been consulting other projects too, including SparkFun's balloon.

[Daniel Wee]

You guys may want to look at the actual specifications for the sensor series. I don't think it goes down to zero actually, last I checked.

Secondly, even if it did go down to zero, the resolution just isn't sufficient for it to be useful.

Thirdly, I'm not completely sure how the temperature is compensated but typically if you artificially warm the sensor up, that might mess up the compensation.

Daniel

[nickersonm]

Perhaps you should check the specifications again. Right on pages 5 and 6 of the datasheet that integ3r linked, it lists the pressure ranges, as I mentioned before. Pmin for all the absolute and gauge sensors is 0 psi. Specifically, the "absolute" series measures down to vacuum without any external references:

... reading pressure over the specified full scale pressure span ... The absolute versions have an internal vacuum reference and an output value proportional to absolute pressure.

The output is 14 bits, with a 12 bit guaranteed resolution. Unfortunately, the "absolute" series with an internal reference does not have a smaller span than 0-1 bar. I don't know what you consider useful, but that would provide 0.244 mbar resolution or better, which works out to about 350 ft accuracy way up at 100000 ft altitude. It could be up to 4x better than that, depending on individual sensor accuracy; the calculations were for the 12 bit guaranteed resolution, while the sensors have a 14 bit output. Obviously the altitude resolution would be better at lower altitudes (around 6' at sea level, or 7' in Denver). Of course, localized pressure variations due to weather would probably be the main source of error at lower altitudes. Edit: these calculations were done without accounting for the transfer function. Actual numbers should be about 20% worse, since the output range starts at 10% (=vacuum) and ends at 90% (=max pressure).

There's no reason temperature compensation would be messed up; presumably, the pressure sensor would be inside an insulated chamber with all your other electronics, not on top of a heating element. As long as that chamber wasn't airtight, it would correspond to the exterior pressure. It's not as though one would put the pressure sensor inside the balloon or dangling on a long wire outside your electronics enclosure.

Again, I think this would be a good sensor line for SparkFun to carry. I would likely purchase several.

- nickersonm

[integ3r]

The accuracy of this sensor is pretty good. What defines how many bits the sensor is willing to output though? I know that 12 bits of accuracy are guaranteed, but what would enable 14-bit operation?

There are many other things that would increase altitude accuracy as well, such as using the models for the US Standard Atmosphere and just plugging in the converted number of counts to pressure.

Also, where did you get the 0.244 millibar resolution from?

These would be great sensors for SparkFun to have, no question. They also have an internal temperature readout, depending on how many bytes you request from the TWI bus.

[nickersonm]

It looks like they output 14 bit data directly, but the specified minimum resolution is 12 bits. I imagine the 14 bit count may vary a bit, or perhaps just be slightly incorrect. I didn't see details in the data sheet.

I got 0.244 mbar from the 0-1 bar range, spread over 12 bits (2^12 counts). Since I forgot the 80% full range output (0 bar = 10% output, 1 bar = 90% output), it should really be: 1 bar/(2^12 * 0.8) = 0.31 mbar, for the minimum resolution. For 14 bits, that would be 0.076 mbar (1/4th the minimum), but I don't know how reliable that reading would be. If it is simply inaccuracy and not signal noise, a proper calibration per sensor should be able to get an accurate 14 bit reading.

- nickersonm