Long ADC distance...

temtronic · Posted: Fri Mar 03, 2017 6:59 am

As Mr. T says R@R op-amp is a must.

Had a quick look at Microchips op-amp page, MCP6001 series would work fine. About 40 cents each, SMT of course.. add 2nd order filter on the input and RC on the output to make it stable. I'd make a PCB for it, small cable connection to sensor. That way if a sensor fails, easy to replace it.

As for the PIC, while 28 pinners would work(10 ADC,6 LCD,..) Add up what you need, then add 5 more. I still like 40 pinners as I hate to run out of pins! As well, project like this tend to 'grow'.

Jay

newguy · Joined: 24 Jun 2004 Posts: 1902

MCP6001 is quite a bit cheaper than my go-to R-R op amp family of choice, the TLV237x (x = 1, 2 or 4), but the GBW product on the MCP6001 is only 1MHz. However, the TLV's are available in DIP packages while the microchip offering isn't.

temtronic · Posted: Fri Mar 03, 2017 9:36 am

Wow someone still makes DIP Opamps ! hen again I rememeber the Teledyne-Philbrick OP-AMPs that were the about 3x4x2 INCHES.
I still prefer DIPs AND sockets as some things do go 'poof' in the real World,especially during Tstorms with lightning !

Selection of devices these days should be a simple 'drag/drop/choose/selection' process in some 'ap'. I only suggested th MCP as it was first in the list of hundreds....

It's best to fab up 2 or 3 sample units and test in the real World as work benches are always nice and warm and dry.....

Jay

Gabriel · Joined: 03 Aug 2009 Posts: 1067 Location: Panama

1-Wire Network...

I have seen the light!

G.
_________________
CCS PCM 5.078 & CCS PCH 5.093

ELCouz · Joined: 18 Jul 2007 Posts: 427 Location: Montreal,Quebec

Just to be sure we are on the same page...

You are asking for op-amps because impedance will be high?

I checked in my junk box, I have a bunch of LM358 in DIP8 pkg. See anything wrong using the LM358?

Thanks guys!
_________________
Regards,
Laurent

-----------
Here's my first visual theme for the CCS C Compiler. Enjoy!

temtronic · Posted: Fri Mar 03, 2017 7:29 pm

What you need is a 'rail-to-rail' op amp, configured as a vlotage follower (X1 gain). The LM358, if memory is still good, only swings about 1.5 volts below VDD. So if powered from +5, the output of the LM358 can't go higher than 3.5, even if the input is say 4 volts.
With a rail-to-rail (R2R) op amp, you'll get 4.230 volts out for 4.230 volts in

If the input can never exceed say 3.5, then yes the LM358 should be OK.

however, just to add confusion, be sure the input can go R2R as well...some op amps might not like getting close to +5 ( supply voltage).

bench test, always bench test...use a heat gun and cold spray to 'mimick' the temperatures Mother Nature wil be for your project.

Jay

Ttelmah · Joined: 11 Mar 2010 Posts: 19221

and (of course), you don't want voltage gain. So no 'R1' in the circuit above...

I still have one circuit in production that uses a TO99 cased amplifier. However the manufacturer still makes the part, and it is only available in this package. It is a bit special a 'hyper' high input impedance instrumentation amplifier. It's being fed of a source that is over 10^9 ohm impedance. PTFE substrate board...

ELCouz · Joined: 18 Jul 2007 Posts: 427 Location: Montreal,Quebec

Alright got it! Smile

Thanks for the help.

Have a nice week-end guys!
_________________
Regards,
Laurent

-----------
Here's my first visual theme for the CCS C Compiler. Enjoy!

starfire151 · Joined: 01 Apr 2007 Posts: 195

Hi -

I just saw this thread and thought I'd chime in with my $0.02 worth...

Have you looked at the Tindie I2C soil moisture sensor (www.tindie.com)? It is also a capacitive soil moisture sensor and is fairly well water-resistant, although I added some protection to the edges of the PCB. It is a bit more expensive than your analog output sensor, though, at about $13 each. The advantage is you could daisy-chain the sensors. I've used this sensor in a couple of projects and it works pretty well.

Maxim makes the MAX3816A 2-wire extender for extension up to 60 meters at 100kbps, also.

Just FYI...
Smile

Ttelmah · Joined: 11 Mar 2010 Posts: 19221

The other big advantage this has, is the timings used to actually measure the capacitance are crystal based instead of analog. Long term should be a lot more stable.

ELCouz · Joined: 18 Jul 2007 Posts: 427 Location: Montreal,Quebec

temtronic · Posted: Sun Mar 05, 2017 5:44 pm

Methinks the KISS methodology will work just fine for you outdoors as distance really isn't an issue. XLR cables ARE robust, application is good.
The second you go 'digital network' you open a huge 'can of worms'. All of them, except 1, have serious failure modes and it seems today's engineers just keep adding chips and code to try to get around problems.
As long as you're not in a hurry and know the difference between voltage and current, you can easily go 15 miles on a single conductor.

Jay

Ttelmah · Joined: 11 Mar 2010 Posts: 19221

and (of course), you can reject a lot of noise with careful selection of filtering on the ADC reading. Spikes in particular can be removed with 'Olympic' averaging.

Remember also though to add protection on the ADC inputs. Don't rely on the chip's own protection. Have something like an MOV on the input, followed by a filter capacitor, and then diode clamps to the supply rails. For your low frequency application, you don't have to worry about the diode recovery time. Remember also that the energy has to go somewhere, so there has to be sufficient smoothing on the supply to cope.

temtronic · Posted: Mon Mar 06, 2017 8:27 am

Mr. T is right about protection, MOVs and caps and such. I even used NE-2s ( neon lamps) as well. Gee, do they even MAKE NE-2s anymore? I must be a dinosaur....feeling old today.

Jay

newguy · Joined: 24 Jun 2004 Posts: 1902

My "standard" protection scheme is this:
- series R, usually ~100R.
- cap to GND. 0.1uF makes this LPF cutoff freq ~ 16kHz. Tailor this C and the series R to hit whatever frequency you want/need.
- Bidirectional TVS (transient voltage suppressor) to GND (in parallel with the cap).
- "sacrificial" op-amp, usually configured to a be simple voltage follower. My line of reasoning is that if something fries, I'd like it to be an op-amp and not my processor.
- any additional analog filtering you require can follow (if necessary)
- finally the signal is input to the A/D.

This protection network depends very strongly on the ability of whatever you're reading to drive this network. As such, the system as a whole has to be considered. Sometimes an extra series PTC resettable fuse needs to be added to the input if there is the possibility of a cable fault shorting a "high" voltage supply line to your analog input. The nice thing about a PTC resettable fuse is that they'll automatically open when there's a fault. The only thing that needs to be done to reset it is just cycle the power.