Terminal boards for WebControl

[mminehan]

Hi Ross and CAI Support,
 
Thanks for the advice regarding this.  Sorry I wasn't clear on the pH probe matter.  The confusion may have been the phidget boards I was using, but want to eliminate and so that I can connect the probes directly to the webamp.
 
I'm not sure I would be skilled enough to replace an SO package chip.  And given that you said "clean them really well and keep them dry" it seems like a tricky task.  I do have a fine tipped soldering iron and would be willing to give it a go.  Maybe I could send one of the boards back and get you send one with the LMP7704 or MCP6004?  Of course I would pay for postage (to and from NZ).
 
Also, were you aware that with the regulator installed on the webamp board it does not fit in the standard CAI enclosure?  
 
I don't want to thread hijack as this subtopic seems to be occupying a few posts now.  Should we start a new thread in this section?
 
Marty

 

[rossw]

Thanks for the advice regarding this.  Sorry I wasn't clear on the pH probe matter.  The confusion may have been the phidget boards I was using, but want to eliminate and so that I can connect the probes directly to the webamp.

 
Figured that now (too late to give you better advice at the time, sorry!)
 

I'm not sure I would be skilled enough to replace an SO package chip.  And given that you said "clean them really well and keep them dry" it seems like a tricky task.

 
My suggestion of cleaning them really well, and keeping them dry, was more about trying to maintain the highest input impedance possible.
Flux, moisture and other contamination of the FR4 board substrate will have a negative effect.
 
Normally when designing for high impedances, there are techniques to keep wide spaces and "artifical ground" guard traces. This was unnecessary on the original boards and as such, wasn't done. Simply using very-low input opamps will help, but I wish I was more confident with the outcome.
 

I do have a fine tipped soldering iron and would be willing to give it a go.  Maybe I could send one of the boards back and get you send one with the LMP7704 or MCP6004?  Of course I would pay for postage (to and from NZ).

 
Unsoldering surface-mount components with a conventional iron, without causing damage to the board, is something of an art.
I'm more than happy to swap the chip for you on one (or more) boards. I don't YET have the LMP7704 devices. I DO have MCP6004 on hand though, but please remember the caveats earlier (particularly as relates to the maximum supply of 6V and therefore not getting the resolution you might want)
 
 

Also, were you aware that with the regulator installed on the webamp board it does not fit in the standard CAI enclosure?  

 
I wasn't, but then I've never had one of the enclosures! 
Is it too HIGH or too WIDE?  (The last lot of heatsinks I got were not the ones I ordered. They're thicker and hang out the side of the board (the originals didn't).
Do you have a photo? If it's too high, one option might be to lay the regulator down, parallel to the PCB. Would that then fit? (Or, bolt the regulator to the enclosure and use 3 short flying leads, and/or a plug and socket to attach?)
 

I don't want to thread hijack as this subtopic seems to be occupying a few posts now.  Should we start a new thread in this section?

 
We're still talking about the terminal boards, so it's still relevant, but by all means start another thread if you'd prefer.

 

[rossw]

Marty, I have a heap of the 4-channel amplifier boards bare, would you like me to solder one or two up with MCP6004 chips?
You probably don't need the screw connectors, which may be one less point to introduce input leakage.
 
It might be easier to mess about with and test this way?
 

 

[mminehan]

Hi Ross,
One of the smaller boards would be great.
I would only need one of them.
I can pay via PayPal.
Also I have attached a picture of the regulator in the enclosure. I measured it and it protrudes about 8mm above the case.
As you can see in the picture I trimmed the bottom of the heat sink a little. This allows me to carefully bend the regulator to about 45 degrees allowing it to fit in the case, althoug in the picture it is upright. Not sure how much this will compromise the cooling capacity though.
Marty.

 

[rossw]

Marty, I've been thinking on your pH problem and come to the unavoidable conclusion that the circuit I presented earlier (the summing amplifier) can never be made to work satisfactorily in your application without a buffer amplifier between it and the pH probe.
 
The MCP6004 amplifier input current is less than 1/5000th that of the LM2902. At around 1pA for he MCP6004 vs 5nA (or more) for the LM2902, this chip is probably just about good enough. Pity about the 6V maximum supply.
 
The answer, I think, will come from using the 4th (spare) amplifier with a trimpot, to generate a low-impedance 600mV offset (+0.6V) as the "ground" lead to your pH probe, in effect level-shifting it from -0.6 to +0.6V output to 0 to +1.2V output.
 
With a little over 4X amplifier gain we can boost that to 0-5V, with 2.5V=0pH
By shifting the pH reference in this manner, we don't need a summing input, and can make use of the highest possible input impedance configuration.
 
I'm still not entirely convinced that the leakage on the PCB itself won't be a problem. Lifting the pin from the IC and connecting it via a flying lead to your pH probe output still makes more sense to me. Can draw you up a schematic of what I propose if this isn't clear.

 

[Tschmidt]

I've been following this thread with some interest since I just built an WC based aquarium controller for my wife and was thinking about adding PH measurement to the system.
 
Rather then running the probe barefoot  have you looked at using an analog or digital PH front end module to simplify connecting to the WC?
 
/tom

 

[mminehan]

Hi Ross,
I'm keen to give the MCP op amp a try. But would need help with the schematic as my electronic skills are moderate (enough to turn a circuit diagram into a working circuit, but not skilled enough to design a circuit).
I looked up the MCP6004 and ph electrodes and didn't find much. But there are a few circuits using the MCP6xxx series as simple gain amps for pH probes.
I presume the same amp circuit could be used for the ORP probe (without the need for an offset of course).

I did come across this which may be a simple way to solve the problem.
http://www.jameco.com/1/1/58265-sen0161-analog-ph-meter-kit-water-quality-testing.html
The schematic is on the web page. It has an output of 0-5V, which I think is what using the MCP6004 will do.

Tschmit,
I did look at using a pH front end, but they are all quite expensive (~$400) and then only have an analouge output in mA. That would require another circuit to covert the 4mA - 20mA output to 0-10V.

Marty.

 

[rossw]

Hi Ross,
I'm keen to give the MCP op amp a try. But would need help with the schematic as my electronic skills are moderate (enough to turn a circuit diagram into a working circuit, but not skilled enough to design a circuit).
I looked up the MCP6004 and ph electrodes and didn't find much. But there are a few circuits using the MCP6xxx series as simple gain amps for pH probes.
I presume the same amp circuit could be used for the ORP probe (without the need for an offset of course).

 
OK, basically this is what I'm suggesting. An existing (or new) board, fitted with MCP6004 instead of LM2902.
Option 1: use an existing board, swap the chip (or I'll send you another with the MCP6004). Use the "spare" chip on-board, and the prototyping area to add the extra 25-turn pot and fixed resistor shown here:
 

 
The 33K resistor and 10K multiturn pot simply let us create an "artificial ground" between 0V and 1.2V approximately.
That is presented on pin 7 of the MCP6004 as a low-impedance, buffered version of this reference and drives the pH probe on what would normally be its "ground" pin.
 
The return from the pH probe goes in to pin 3. With no other summing resistors or anything else, the opamp should have minimum possible input bias.
R2, the 50K pot typically already fitted on the board is used to set the gain. You will need to remove the 7805 regulator, replace it with a wire link from input to output, and power the board NOT via the +V connector on the analog input, but through the WebControl DC in, so its 5V reg will power the circuit.
This means maximum input will be 5V. Calibration will be easier if we ignore the pH probe initially.
Assume you want full range 0pH to 14pH.  7pH will be nominally at 0V, so adjust R1 for 600mV (about half-way). Put a wire link instead of the pH probe.
Adjust the 50K trimpot for 2.5V output on pin 1.
To check the gain, adjust R1 so pin 7 is 300mV. Check the voltage on pin1 is 1.25V
Re-adjust R1 so pin 7 is 900mV and check pin1 is 3.75V
Return R1 to original position (pin7 = 600mV). Remove the link, install pH probe and check.
 
The advantage of this system is that you can choose to have an "expanded scale" and read for example only from pH 3 to 10 (or any range you like) by adjusting the offset and gain.
 
Option 2.
Use one of my 4-channel boards, again with MCP6004. Use exactly the same diagram and procedure as above, except this time you can use the output of the pH amplifier (or ORP cell) from the MCP6004, into the existing webamp board. Set the amplifier gains to 2.00. Now, the 0-5V input from the MCP6004 will be amplified to 0-10V to the webcontrol board so you get maximum analog input resolution AND you can still use the external regulator if you wish.

 

[mminehan]

Ross,
Thanks again for you help. I'll go with option 2 seeing as I already have the WebAmp boards with the LM2902 installed. And I am sing one of the channels with a pressure sensor. Plus that way I can leave the regulator on the board too.
Can you send me a PayPal invoice for the 4 channel board?
Thanks, Marty.

 

[rossw]

Ross,
Thanks again for you help. I'll go with option 2 seeing as I already have the WebAmp boards with the LM2902 installed. And I am sing one of the channels with a pressure sensor. Plus that way I can leave the regulator on the board too.
Can you send me a PayPal invoice for the 4 channel board?
Thanks, Marty.

 
Marty, I've just built you a 4-channel amplifier board with MCP6004.
I've modified the circuit slightly. Rather than using 50K pots on the gain, and a 10K for offset, I've user all 50K pots and a 150K resistor instead of 33K. Provides the same range. I also added another 100nF cap across the divider just to help reduce any noise.
 
Adjusted the offset and gain as described above, and was perfect at 0.3V, 0.6V and 0.9V. (Well within 1mV on the output - 0.3V gave 1.2508V, 0.5V gave 2.5000V, 0.9V gave  3.7492V), when running from a 5.03V input supply.
 
I don't have any 500M resistors, but I then returned the offset to 0.600V and used a 10M resistor to the input. I still saw 2.500V on the output so the opamp wasn't loading down with 10M input. I'll be fun to try it on your pH probe.
 
I would have loved to put the board in the ultrasonic cleaner for a while, it's only been "quickly" cleaned with white spirit to remove most of the flux etc. You can try it as it is, or you can try cleaning it better when you get it.
 
I'm about to run it down to the postoffice, I think they're open today, not sure if parcels will leave today or just wait until monday.
 
Here's all I had to do differently when assembling the 4ch amp board:

 
And here is your board completed and tested, so you know what to connect where!

 

[mminehan]

Ross,
You are certainly going beyond what would be considered reasonable assistance. Thank you so much for your time and effort. It's nice to know that there are genuinely good people out there willing to help. Please allow me to provide you with some sort of remuneration for at least the hardware and postage if not the time involved.

I'm looking forward to trying it out. It will be the icing on the cake of the pool automation. And no doubt other people will be interested both from a pool perspective and a reef aquarium setting.

I have some circuit board cleaner (in a spray can) that I can use to give it a clean.

Regards, Marty.

 

[mminehan]

Hi Ross,
The small board arrived last week and I have been doing some experimenting with mixed results.
The board worked perfectly with the pH and ORP probes in the test solutions (isolated from the pool).  It only required some minor adjustment.  And I used the WebAmp to double the gain to get the 0-10v range.
BUT......I ran into some problems when I put the probes into pool system.
I have had major issues with noise.  It seems there are significant stray currents in the pool water.  I connected a scope to the raw probe output and discovered a lot of high frequency noise (+/- 200 to 400mV), which the board simply amplified meaning the end readings ranged from 1 to 7v.
I then tried adding various combinations of low pass filters (pF to uF capcitors and resistors) to the input which helped a little, but not enough for a reliable end result.
Unfortunately the best result I have had is with the phidgets ORP/pH boards (http://www.phidgets.com/products.php?product_id=1130).  They seem to have better noise filters than I can produce, and work with the webamp board to get an output in the 0-10v range.
Researching the topic it seems noise in pool systems is a significant problem.  I am waiting on a section of 60mm stainless earthing pipe to install near the filtration system to try and ground the system and reduce some of the noise.   But this may not work as desired.
I'll keep working on it and let you know how it goes once the earthing pipe has arrived.
Marty.

 

[rossw]

I'd wondered about that, Marty.
The water in the pool will be "reasonably" conductive - salt and other things in solution - and there's a lot of it, especially if you include such things as piping, pumps, heaters etc.
 
With such a high-impedance source, just about ANY noise will be coupled straight in.
 
Filtering will so substantially lower the input impedance as to render the benefits of the amplifier null and void.
 
My only thought would be something like a 1uF cap from the opamp out to its inverting input to reduce the AC gain to unity, while keeping the non-inverting input impedance high. You could then add a lowpass filter on the output of the first opamp. It's still not ideal though.
 
The last time I did this, we ended up with a tricky system that had our monitoring in an "electrically isolated" sample tank. A water sample came in via a valve and (importantly) an airgap. We'd dump the current sample, take a new sample then stop the flow, measure while it was sitting, isolated, then repeat. Messy, but was the only way in the application.

 

[rossw]

Marty, this might be interesting to you.... I have seen SOME commercially, but not sure of their price or how readily available they are.
http://cdn.intechopen.com/pdfs-wm/29116.pdf

 

[LarrylLix]

Just some thoughts.

Would installing the probe inside a grounded perforated metal shield help eliminate some stray induction like a Faraday cage?

There is perforated metal sheeting that could be rolled into a cylinder but the saline solution may be a problem eating certain metals.

Galvanic action could become a problem with dissimilar metals also.