Interfacing A Pressure Sensor

"Michael" schreef in bericht news:ZfENg.31335$ snipped-for-privacy@fe1.news.blueyonder.co.uk...

Michael,

There's nothing about electrical characteristics. As far as I can see from the schematic you need to connect some power supply to the pins 1 and 3 and some differential amplifier to the pins 2 and 4. But required voltage and current of the power supply and expected output signal are al in the dark.

petrus bitbyter

Hi, Michael. You need a differential amplifier that has essentially infinite input impedance at both inputs. This differential amplifier gives as an output the difference between the two voltages. You can then read that voltage with the dreaded PICAXE. The data sheet doesn't say anything about the type of output you'd expect, but I'll bet it will be less than 1V, so you might want to amplify that output voltage.

You can do this with 3/4 of a quad op amp and a handful of precision-matched resistors, like on page 14 of National Semiconductor's AN-31, "Differential Input Instrumentation Amplifier" at the top of the page.

Or you can just buy the thing in one IC package called an instrumentation amp. Analog Devices' AD622 isn't too expensive, and works great.

If this is a bit too complicated, please post again, and indicate what you don't understand.

Cheers Chris

You need a differential amplifier. Depending on the precision you need, something like the AD620 or AD623 should suffice.

Watch out for temperature, pressure sensors can be notoriously bad over the temperature range. There are temperature compensated ones available, so if it's a problem use one of them if possible.

Dave :)

Thanks for the replies.

Could I use a LM13700 amp? Datasheet:

Michael

Probably, but it's not the right part for the job. Much better to use a proper precision differential amplifier who's gain can be set with a single resistor. The AD623 is a perfect example:

The AD620 is higher spec equivalent if you need it.

Dave :)

Hi, Michael. Last call for the LM13700 was about 8 years ago -- it's obsolete, isn't it?

I'm not sure why you want to use a transconductance amplifier when a standard op amp will do.

Look. This is for a PICAXE, and I'd guess this is for a class in the U.K. Also, I would suppose you're looking for the lowest cost, easiest solution here. Let's start from the beginning, and see what you need to get this done.

First, you didn't mention which model of PICAXE you're using. It's one with analog inputs, I'm assuming more than one. Your PICAXE gives a result from 0 to 255 for an analog conversion with the "readadc" command. That's an 8-bit A to D converter.

See p.39. Now, since you neglected to mention the model of your sensor, I've got to make some assumptions here. Let's assume you've got an output span (that's the difference in voltage between terminal 2 and terminal 4) of 200mV to 400mV. You could take your Wheatstone bridge sensor and just attach the two terminals (2 and 4) to 2 inputs. That would initially seem like the easiest, and it is. The problem is, you won't get much of a result. Here's why.

Your Wheatstone bridge gives a differential output. Some of the types referred to in the data sheet you link have a full scale span of 200mV. That means that terminal 2 will be, say, 2.4V, and terminal 4 will be

2.6V. When you readadc these with the PIC you will get results of 122 and 133. If you subtract, the difference will be 10. That means your sensor, which is made to measure pressure of, say, 0-15psi with 5% accuracy, will only give you a number 10 for 15psi and 0 for 0psi. That's *very* poor resolution, and is basically wasting all the precision of the sensor -- 1.5psi per count.

Now, if you can find a way to electronically subtract the difference between the two outputs, and then amplify it, you will have a much more precise reading. If you subtracted and then multiplied by 10, you would have a 0 to 2V reading, which would correspond to a number 0 to

102 or so. That will give you a reading of about 0.15psi per count, much more information.

Op amps are made to do this type of thing. You could easily do it with just one op amp, except that a standard difference amplifier configuration would load down the Wheatstone bridge and mess up the voltage divider. It would be better to use the type of op amp setup I mentioned above.

If you've got an LM324 in the class (the world's most popular op amp), an ohmmeter, and a handful of 10K and 100K resistors, you can put together a three op amp difference amplifier (instrumentation amplifier) that will get you an "A". Looking at the National Semiconductor reference shown above, you'll see that the first two op amps are basically voltage followers. The third is a standard difference amp (view in fixed font or M$ Notepad):

| | VCC | + | | | o----. | | | | .-. .-. | | | | | .---------. | | | | | | | | '-' '-' | | | | | | |\\| | | | | '--|-\\ | ___ ___ | | | | >---o-|___|--o--|___|---. | o----)--------|+/ 10K | 100K | | | | |/ | | | | | 1/4 LM324 | | | | | | |\\ | | | | '--|-\\ | Vout | Sensor | >----o---o | | | .--|+/ | | | | |/ | | | 1/4 LM324 | 1/4 LM324 | | | |\\ | | | o--------|+\\ ___ | ___ | | | | >---o-|___|--o--|___|---. | | | .--|-/ | 10K 100K | | .-. .-. | |/| | === | | | | | | | GND | | | | | | | | '-' '-' '---------' | | | | o----' | | | === | GND (created by AACircuit v1.28.6 beta 04/19/05

For classroom purpose, just find the two 10K resistors that are closest to each other, and the two 100K resistors that are closest to each other. As it says in the appnote, if you match the resistors like this, the gain will be 100K / 10K, or 10. So your 200mV differential input signal will be converted to a ground-referenced 0-2V signal you can read with one PICAXE input.

I hope this has been of help. Please feel free to post again if you need more help, and let us know how it goes.

Cheers Chris

Thanks Chris,

I did say which PICAXE I was using in the original post - 08M.

The reason I asked about the other one is that I'm in the UK and that was the one my usual supplier has.

It's not for a classroom project, although my knowledge is about that level.

Having looked at the costs for a LM124 (£26GBP

) I'll have a play with the LM324

Cheers,

Michael

f -

-

ed,

ver

el.

)

Hi, Michael. You could try RS -- they've got the LM324N listed as their p/n 268-0002 at =A30.30 ea. Try to match your two pairs of resistors to better than 1% of each other for best results. You might want to buy 10 of each of the 10k and 100K to get a best match.

But even so, you'll get much better results with an op amp than just using the PICAXE by itself.

Good luck on your efforts -- you seem to be doing a good job of teaching yourself. Feel free to post again.

Cheers Chris

Hi, Michael. You could try RS -- they've got the LM324N listed as their p/n 268-0002 at £0.30 ea. Try to match your two pairs of resistors to better than 1% of each other for best results. You might want to buy 10 of each of the 10k and 100K to get a best match.

But even so, you'll get much better results with an op amp than just using the PICAXE by itself.

Good luck on your efforts -- you seem to be doing a good job of teaching yourself. Feel free to post again.

Cheers Chris

Thanks Chris,

The parts are on order - let you know how it goes...

Michael