Absolute pressure sensor for vacuum measurements

Go To Last Post
14 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hello, community!

 

I'm trying to find suitable sensor for my task. My everyday work (the same is true for my colleagues) involves operating with evacuated flasks and equipment (organic synthesis) and we're using old, ugly and dangerous mercury manometer for pressure measurements. I want to replace it with compact pressure sensing module based on integrated circuit connected to AVR-driven device. My plan is to buy I2C/SPI (I2C is in preference) digital absolute pressure sensor which is capable to measure air pressure with granularity of 1 mm Hg at least and to work at very low pressures (for example, our usual working pressure is about 12 mm Hg). Operating voltage doesn't matter now - it's okay to live with 3.3 and 5.0 volts as well.

 

As for now I've found two interesting ICs which seems to fit to my request:

  1. SSCMRNN015PASA5
  2. DLV-015A-E1BD-C-NI3F

But it would be nice to know community opinion and suggestions - perhaps anybody already worked on similar tasks and can offer better options or comment my choice.

 

Thank you

Viktor Drobot
A.N. Belozersky Research Institute Of Physico-Chemical Biology MSU

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I don't think those sensors will work with a vacuum.

I once worked a control board for a food dehydrator, we had to use a vacuum thermocouple sensor to get readings at the deep vacumm levels needed.

It took some doing to drive it with the needed signal and amplify the analog output, but it worked well once it was calibrated.

We used a sensor like this one. https://www.digivac.com/product/...

 

Jim

 

Click Link: Get Free Stock: Retire early! PM for strategy

share.robinhood.com/jamesc3274
get $5 free gold/silver https://www.onegold.com/join/713...

 

 

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I don't think those sensors will work with a vacuum.

Why? According to the datasheets working range starts from zero. Also they are absolute sensors and if I understand it correctly measurements are done against sealed evacuated internal container

Viktor Drobot
A.N. Belozersky Research Institute Of Physico-Chemical Biology MSU

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

dviktor wrote:
According to the datasheets working range starts from zero. Also they are absolute sensors

That is what we thought too, but it did not work for our app, maybe you will have better luck!

Jim

 

Click Link: Get Free Stock: Retire early! PM for strategy

share.robinhood.com/jamesc3274
get $5 free gold/silver https://www.onegold.com/join/713...

 

 

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

With a quick look at Sensor #1 linked to above, I would also have thought that it would work, (note that I didn't verify the accuracy requirements).

 

Viktor, I assume the temperature of the lab setup will be within the temperature spec for the sensor.

 

This is obvious, but I'll mention it anyway:

Clearly keep the Hg column in place and attach the new sensor via a "T" connection so that you can calibrate / cross check the digital system's readout against the Hg column.

 

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I had a read of the honeywell datasheet. From what i can see the sensor should work. With the port open to atmosphere you should measure near full scale.
What i didn’t look at closely is the noise and resolution specs. If the resolution you want is sitting within the noise of the sensor, then it is not for you. Be sure to read the fine print. No Sensor is perfect.

Be it spi or i2c - keep the connection length short. As in cm. If you want a sensor to be located remotely, use a packaged sensor with 4-20mA interface.
Also be sure to put error detection in your code and communicate errors to the user. A mercury manometer is very simple - mercury? Glass tube intact? There’s very little else to go wrong. With a semiconductor sensor, there’s plenty to go wrong. Having a measurement device that is erratic or lies to you is annoying at best.

Anyway, sounds like a fun project - just like a science experiment at school!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ki0bk, could you please describe in short what do you mean when you say "it didn't work"? Just want to know specific pitfalls in your case.

 

DocJC, both sensors are rated for 0.25% accuracy and it seems pretty good for our purposes. We use vacuum for distillation, rotary evaporation and substances drying processes and the most important pressure readings should be in case of vacuum distillation, because we collect different fractions/substances according to the T-P nomogram. It's usually sufficient to get accuracy about 1 mm Hg in that case. Also one of the possible future applications is calibration of analog pressure sensors for other vacuum works (with soft requirements for accuracy, e. g. drying substances in dessicator) and these sensors are much cheaper. Of course, we'll recheck data readouts against old, ugly but robust Hg column. I plan to blow special borate glass T-connector for sensor placing and tubing connections and seal data/power wires directly into molten glass with the help of gas torch.

 

Kartman, the Honeywell sensor is of type RN which means single port - I suppose it's because this sensor is of absolute type, not differential and have highly evacuated internal chamber so second port is connected to it. Both sensors have 14-bit ADC so it should be fine to use them. Because of mounting design described above I will have to use quite long wires (perhaps up to 30 cm length); I hope that I2C interface should work fine in Normal Mode for that case. About error detection: for the test purposes I want to make very simple setup with 1 sensor, 1 7-segment display and 1 MCU; I suppose that this is enough for simple pressure measurement. In case of I2C bus unstability/sensor failure it is enough to print simple Err on display. In my opinion it's much easier to get things horribly wrong with Hg manometer - our equipment is being used by a lot of people including students and other unqualified personnel and we can't control every their action. It's enough to forget to close manometer valve and open system to the atmosphere and then you will see nice fountain of mercury drops and the rest of the day (in the happiest case) will be about disposal of toxic metal. It's very common for newbies to make liquid metal rain in lab, but they are never participating in fixing aftermaths :) So I think IC pressure sensor will be the safest option for our type of works.

 

What is confusing me: for Honeywell sensors I can't find digital protocol specifications. May be it's just too late and I should go to bed but I can't find description of chip registers and command specs...

Viktor Drobot
A.N. Belozersky Research Institute Of Physico-Chemical Biology MSU

Last Edited: Tue. Nov 19, 2019 - 12:56 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In our case we needed to measure a very low vacuum, 600-800 mTorr, IIRC, the pressure sensor would not read vacuum, we had to use a vacuum sensor.

So it was our mistake, by not understanding that pressure and vacuum are not the same (yes, I know a vacuum is a very low pressure, or is it the lack of pressure), anyway a similar pressure sensor as you listed above would not work in our project as it would not read vacuum(pressure?) in the range we needed.

 

Jim

 

Click Link: Get Free Stock: Retire early! PM for strategy

share.robinhood.com/jamesc3274
get $5 free gold/silver https://www.onegold.com/join/713...

 

 

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

dviktor wrote:
Both sensors have 14-bit ADC so it should be fine to use them
My very point! It might have 14 bit Resolution, but some of those bits may be buried in noise. That is why I suggested to read the datasheet carefully.

 

30cm should be ok, again, refer to the datasheet! You may require power supply bypassing. These are not simple devices!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 I plan to blow special borate glass T-connector for sensor placing and tubing connections and seal data/power wires directly into molten glass with the help of gas torch.

How cool!

 

I'd love to see a photo of the setup once you have it constructed!

 

JC 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hey Doc! Have you seen this?

 

https://www.youtube.com/watch?v=...

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

0.25% accuracy is of full scale, so if you're using a 0-1bar device, the accuracy is ~2.5mbar which may not be close enough for you. It may have a 14-bit ADC but the overall accuracy is only about nine bits (though differential accuracy may be much higher).

 

I've used barometer devices to read altitude at better than 100,000 ft - so in the millibar range - but we had no idea of how accurate they really were except by comparison with a mechanical (and itself uncalibrated) gauge in a chamber. They were factory calibrated and had a large bundle of correction factors on chip which had to be read - both pressure and temperature - to compensate for what the pressure really was. Ten years ago, no external port, and probably discontinued by now anyway, so not a lot of use to you.

 

You might get better accuracy with a differential probe and a number of calibrated pressure references, so the final stage is differential between (say) your target and 10mb.

 

Neil

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ki0bk, thank you for clarification. Your process really requires very low pressure while our lower limit is usually about 12 mm Hg (we use water jet pump for evacuation) so I think we should be ok with that.

 

Kartman, even if I have 4 least significant bits screwed up then I will lost info about 0.7 mm Hg for 1 atmosphere scale sensor: 24/214 * 760 = 0.742 mm Hg. For my purposes it should be pretty enough.

 

DocJC, I think having wires fused into glass will be the safest and most robust way of using vacuum equipment. Of course, when all will be set up I'll post here my results.

 

barnacle, in that case for 1 atmosphere scaled device absolute error will be about 2 mm Hg. Still very good for my requirements. Anyway, I can't find more accurate sensors at the first glance so I'm asking here for real life usecases :) I hope that ADC will give me 10 significant digits at least. I also have no pressure references and can't rely on them so my choice is absolute sensors against sealed vacuum chamber (I suppose high price for them is due to difficulties during production of this type of sensors)

Viktor Drobot
A.N. Belozersky Research Institute Of Physico-Chemical Biology MSU

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Off topic for just a second:

 

Kartman, Wow!  That video on making Nixie tubes is incredible!

What a process.

Thanks for the link!

 

Now, back on topic.

 

JC