very low pressure

Fluctuations? How fast?

I'm thinking 'microphone'.

-- Paul Hovnanian snipped-for-privacy@hovnanian.com

---------------------------------------------------------------------- Have gnu, will travel.

If this is something you are going to mass produce, there is a cute trick for making a pressure to frequency converter. If the pressure pushes a plate in towards a second plate, the capacitance increases. If this is part of a resonant system, the resonant frequency changes.

0.1" of water pressure is about 0.2 torr; that's thermocouple gauge territory at the low $ end of gauge tech. Commercial gauge tubes are in the $30-50 range, controllers $100 and up. Basically a four terminal device with a fine wire between two pins that a current is sent through to heat it up, and a thermocouple on the other two pins attached to the middle of the fine wire to monitor temperature. You can either use constant current and read variable temp, or vary the current to keep the temp constant and use the current as the output. You can also just use a fine wire and put a small current through it and monitor the voltage drop (a thermal conductivity gauge) but they are less sensitive, drift more, and don't hold calibration as well (all things you say you don't care about :-)). Big problem with all of these is that they also respond to changes in gas flow so transient response can be bad.

----- Regards, Carl Ijames

Not very dynamic of an input. I actualy thought of carbon granules like the old microphone. But seemed too messy.

Not mass produced but a few maybe. I considered a capacitance type setup as you mention. Only the incident pressure is to the outside of a cylinder that is giving. And may not be evenly distributed. So a foil cylinder on the in and out side of a soft walled cylinder would allow for instance the grip to be measured of an infant. But the difference in capacitance would be buried in noise near as I can figure.

A heated wire sensor is something I have used in vacuum measurement. The (discernable) resolution is poor. The current requirements are too high for this application anyway. I intend to power up the sensor and it's associated components only briefly and periodicaly so that would eliminate this approach also. Good idea though. They make for great anemometers on the roof.

Thanks for the replies people. I was pleased to have interested so many smart people. Let me give you more info and it may shake loose just the right gem. Picture a small soft toy. A rubber teddy bear for instance. Only a few inches long. It is hollow. Like all such toys we have all seen a lot of. When you squeeze it the pressure inside increases. As it happens the pressure swings from ~.10" to ~1" of water. Bet you thought it was more right? Well so did I. So I measured it. Anyway then I started looking at sensors. Ones that were in stock and you could buy one of. Which eliminates most of the first ones that climb off the google results page and up your arm. The only one that showed a range of 0 to 4" cost +30 $'s !! And it is made for a range 4 times what I need. Which means I get only 1 fourth the output to start with. And the output only swings ~3 volts. So I gotta piss around with reference voltages to try to squeeze all 8 bits from the adc. And on top of that the offset is near v/2 at 0" of water. So that tosses half of whats left of my resolution out. Unless I use a pic that lets you set both ends of the adc but then what do I use for the bottom reference? Jeez. So it looks like I need an op amp. In which case I might as well use an un integrated sensor and save a few bucks there. Only now I have considerably more components when you consider bypassing for power and and input leads becoming radio receivers. And I have a nagging suspicion the gain and offset will have to be programmable if I want to retain a fair number of bits resolution. And I am still not sure from the data sheets that there is a usable output at these pressures for the low cost sensors. So you can see where a sensor with a buffered output and a range of 0 to 1" would be great. Maybe I am going about it all wrong. Tony

It's an interactive toy, right? Why responds to squeezing and not shaking? Accelerometers are far cheaper and easier to implement.

Sorry but no. I am trying to measure squeezing.

You are trying to measure small pressure changes at or around atmospheric pressure?

And for a toy no precise absolute calibration is needed it is delta pressure over some fairly short time interval that matters. You mya be able to find basic crystal earpieces that will behave as a somewhat microphonic transducer with sufficient sensitivity to do what you want. There will be false triggers for loud bangs though.

The other way would be a springy bladder with a controlled leak past a suitable sensor.

When someone says very low pressure it normally means hard vacuum mass spectrometer pressures.

Regards, Martin Brown

If you'd said 1" water pressure differences at atmospheric pressure I wouldn't have wasted our time suggesting a solution suitable for .1-1" water absolute, sigh. Why not just line the toy with a kynar film aluminized on each side and use the piezo output to monitor squeeze/deflection?

----- Regards, Carl Ijames

Is the interior of the "squeeze toy" evacuated, or is it really at atmospheric pressure and you're trying to measure 0.1"-1" above atmospheric?

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
 I love to cook with wine     Sometimes I even put it in the food

Or measure the change in capacitance?

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
 I love to cook with wine     Sometimes I even put it in the food

Ok Ok I am sorry. I meant .05 Above atmospheric. I am talking about inches of water here. Which implies such I would think. But in truth I goofed in not being clear.

I built something similar into a toy for Mattel at a total cost (sensor, wiring and microcontroller) of about 20 cents, but it was just a squeezing / not squeezing 1-bit sensor. Tell me more about what you are planning to do with the amount-of-squeezing information and I might be able to devise a low cost solution.

--
Guy Macon

Try an electret microphone, some are less that a buck in cost. Desensitize it with a thin plastic film (saran wrap) over the mike. Use as many layers as necessary to get the sensitivity range you want by experiment. Amplify the output into the voltage range and offset you want and digitize it with a

10 ADC in a microprocessor. Seems straight forward enough for a toy.

Note, the microphone must work at DC, constant pressure. It cannot be capacitivly coupled nor can it be a dynamic or magnetic element type unless the squeeze is momentary giving a pulse that can be measured. The output would be speed sensitive in that case. Piezo elements may be problematic as well. An electret microphone is a capacitive element in a permanent electric field who's output is a function of displacement. An internal FET follower is the usual output device.

why not just take it literally and set up a manometer with a column of water which blocks light to a photocell (maybe put dye in the water)?

oh, i see. i was thinking one-shot lab experiment. yeah, you probably wouldn't want a manometer made of a column of water for this item.

Use an electret microphone without the DC blocking cap.

--
Paul Hovnanian     mailto:Paul@Hovnanian.com
------------------------------------------------------------------
Life would be so much easier if we could just look at the source code.