PH Probe A/D Coverter ?

[shawn]

I am trying to interface a PH probe to a pic 18F458. The amplifier and buffer circuit for the probe puts out 0 - 14 volts correlating to the ph scale. I tried using the on board AD converter but the resolution is not good enough. I ran the signal threw a voltage divider so that I could scale down my 0 - 14 Volt ph reading. And calibrating the probe could be a problem because Vref+ is 5v. Calibration is at 4, 7, and 10. 10v / 2 = 5 which would be at the top of my vref with no room for error. I've looked all over the internet for an ad converter to fit my purpose but can't find one. This is what I am looking for and why. I want a 14 bit AD chip that is I2C bus capable, its Vref+ needs to be able to goto at least 16.384 volts and vref- needs to be 0v.The 16.384 v is so that I would have a resolution of 1mV. The I2C bus is needed because it is easy. I would try SPI but I am not very good at the protocol. Is there an AD chip thats Vref+ goes that hi. Does anyone know where to look for a chip like this. I looked at MC and Maxim but did not see anything close. They both had chips but nothing that has a Vref+ that hi.
Thanks
Shawn

[shawn]

Another question to help me solve this problem. When I first started playing with the VREF on the pics I thought you could raise vRef+ above vdd, well looking at the data sheets I learned I was wrong. My next question is can I raise Vref- above ground, say 1.33v or does vref have to be ground. The reason I ask, is because if I divide the output of ph probe amplifier by 3 then I should be able to get the range I want out of it in a 2V swing. If I can set vref+ to 2.354V and Vref- to 1.036V I will get a 2mV/Bit of resolution which I think will work for what I want to do.
It would still be nice to find an AD chip so I do not have to use voltage dividers.

Shawn

[skimask]

My next question is can I raise Vref- above ground, say 1.33v or does vref have to be ground.

Yep, you can put Vref- to anything you want (>Vdd not included )
If your crafty enough, you can even use a couple of the PICs pins to make an R-2R ladder so a guy can even vary Vref+ and Vref- to suit the conditions. (i.e. read 0 on the A/D, lower Vref-, read 1023 on the A/D, raise Vref+, and so on...)

[brid0030]

Not a real answer to your hardware questions, but could you rig up a simple manual switch that lets you select from among 3 or 4 voltage-divider arrangements? Whenever I've used a ph meter I have always known whether I had a strong acid, something more or less neutral, or a strong base before getting an exact reading. If you have a clue as to what your are testing beforehand and can anticipate a reasonable range of output, you might be able to tailor the voltage divider to the output and get a lot more out of the 10-bit ADC.

[skimask]

but could you rig up a simple manual switch that lets you select from among 3 or 4 voltage-divider arrangements?

That's what I was saying with using the extra pins.
If you take the extra pins, bring them out, connect a 1K (for example) in between each pin, put +5v and the top, ground at the bottom of the string, and change the pins from input to output (or back), and high to low (or whatever), you can dynamically change the Vref (+ and -) points without having to slide a switch...let the software do the work. Start wide and dial the reading down to a tight reading.
For that matter, you could even use a couple of digital pot's as a variable voltage divider under software control.

[shawn]

Ok guys this is what I did. When I think of a voltage divider I think of 2 resistors thus divide by 2 which does not get it done. So what I did was use a 3 resistor voltage divider to divide by 3 which puts me under 5v max output. The highest ph reading I need is a little over 10 for calibration purposes. Dividing the 0 - 10V ph reading by 3 gives me .00333 volt change for every hundredth in ph degree. So then to figure my Vref+ I used this (vref+ / 1024 = .00333)
Now my variable that is read from the ad port is my exact ph +/- a hundredth, which I can live with since measuring ph is pretty much a joke anyways.

My next question is, what is a good way to make a non drifting precision regulator. Accurate to the mV and will hold voltage in rm temp. I hate using pots and caps on LM317 regulators it always seems like they drift allot, is there a better way.