Analog max impedance

andrewg · Posted: Wed Jan 28, 2009 8:23 am

I'm looking to connect a few soil moisture sensors that have a 100K impedance to an 18F2580. The PIC data sheet says the maximum recommended impedance is 2.5K. The only downside I can see is the extra length of time taken to acquire the correct voltage before conversion. I don't mind spending many milliseconds waiting for each one.

Is that all and it should work fine with enough acquisition time, or is there something I'm missing? If I actually had a sensor I'd just try it out, but I thought I'd ask before buying them.

Thanks for any advice.
_________________
Andrew

SherpaDoug · Joined: 07 Sep 2003 Posts: 1640 Location: Cape Cod Mass USA

If your sensor can charge a cap over time, that cap can provide the low pulse impedance the A/D needs. As the A/D takes its measurement internal impedances and capacitances will change. The signal source needs to maintain voltage over these changes. Normally a low impedance source does this, but a cap can provide the low impedance for the short duration of the measurement.
_________________
The search for better is endless. Instead simply find very good and get the job done.

RLScott · Joined: 10 Jul 2007 Posts: 465

It is more than acquisition time. The datasheet says the input leakage current on an input pin can be as high as 1 uA. If your source impedance is 100K, that 1 uA will drop 100 mV, which is 20 counts on your 10-bit A/D converter (running at 5 volts). So if you don't mind a possible 100 mV error on your readings, then I guess 100K source impedance would be OK too.

Edit: SherpaDoug, a capacitor will not address this constant leakage current problem.
_________________
Robert Scott
Real-Time Specialties
Embedded Systems Consulting

andrewg · Posted: Thu Jan 29, 2009 6:39 am

Ah, I see now. Thanks. The 2.5K is because beyond that the leakage current will start affecting the read values.

My sensor's minimum output is 500mV, so the drop could be 500mV and I'd still be able to read the full range of the instrument.

I'm just using them in a "wet" vs "dry" capacity, so while I might still work out how to compensate for the leakage current in software, at least I can get away with just experimentally finding the desired threshold ADC reading.

Thanks for all your advice!
_________________
Andrew

crystal_lattice · Joined: 13 Jun 2006 Posts: 164

Alternatively you could buffer them with an OPAMP - and also increase your output voltage swing.

SherpaDoug · Joined: 07 Sep 2003 Posts: 1640 Location: Cape Cod Mass USA

RLScott has a very valid point. Worse yet the 1uA could be positive or negative and may vary between chips and vary with voltage and with temperature. On the other hand 1uA is the worst case value and you are not likely to see leakage that high, especially at room temp or cooler. 1uA is the value above which Microchip will scrap the PIC rather than sell it. Since they don't want to scrap many you can bet they keep the average leakage lower. They have been using that 1uA spec since at least 1990. I'll also bet they have gotten better since then.

I recommend you test any PICs you have on hand to see what their leakage is and get a feel for what to expect. If you are designing satellites or pacemakers the worst case specs are more important that if it is a home project.
_________________
The search for better is endless. Instead simply find very good and get the job done.

RLScott · Joined: 10 Jul 2007 Posts: 465

daraos · Joined: 16 Oct 2008 Posts: 18 Location: Chile

the safest bet is using an OPAMP as voltage follower such as the classic lm741, or a lm324 (four 741 in one IC)

andrewg · Posted: Fri Jan 30, 2009 9:27 am

Hmmmm. This looks like it is more complicated than I thought. At least, more complicated than the PIC only solution I was hoping for. The LM324 is sounding like the best option by far.

Thanks for everyone's input!
_________________
Andrew

asmallri · Joined: 12 Aug 2004 Posts: 1634 Location: Perth, Australia