Op amp virtual earth conundrum

Thread Starter

kutalinelucas

Joined Nov 20, 2007
98
Hey guys

I'm having some difficulty with my sensor circuit...I posted a question a couple of days ago to no avail, so I thought I'd try and simplify things somewhat

The sensor part of the circuit (http://s23.postimg.org/ej7btm6ij/bre...19_06_2013.png) consists of two op-amps each with a separate matched cold-junction compensation IC (http://www.linear.com/product/LTK001) to read temps from two K-type thermocouples, powered by a regulated split 12v (-6v to +6v) The result is fed into the ADC of a PIC18F4550.

And here's the problem...when I heat a block of aluminium with the thermocouples fixed inside, the voltage measured across the thermocouple terminals increases with temperature as expected, but the output, measured against common ground sits at about 0.3v until the thermocouple output measures about 2.5mA at 70 degrees. Then the voltage rapidly increases to ~5v, settles and continues to rise fairly linearly with temperature.

The only thing I notice is the splitter circuit is not earthed in anyway...would this be an issue, and how would I get around it?

Thanks for looking
 

ErnieM

Joined Apr 24, 2011
8,377
The only thing I notice is the splitter circuit is not earthed in anyway...would this be an issue, and how would I get around it?
You have a ground symbol in your schematic, but it is not clear how this is connected.

You do not have a split ±6 V supply. You have a single +12V supply.

You would need to make a zero volt reference to use as ground. Sometimes this is as simple as two equal resistors in series across the 12V: the common center serves as a ground.

Here I am using "ground" to mean a simple zero volt reference point; it has nothing to do with the planet upon which I sit.
 

Thread Starter

kutalinelucas

Joined Nov 20, 2007
98
Hey ErnieM

Thanks for getting back to me...to what "ground" are you referring to? All grounds in the above schematic are common to each other, two peripheral circuits, 4 power supplies and my PC. The only ground which is isolated is that of the 12v supply which is split to power the op amps, and the virtual ground of the uA741, which is where I think I may be going wrong. Putting a large resistance between the splitter output and common ground doesn't solve the problem.

I do have a single regulated supply which is split into 2 separate split ±6v rails using the 741 as a buffer. The rail has been split twice to power each pair of op amp & cold junc. compensator in case noise because an issue, but I assure you I have 2 separate ±6v rails powering all relevant pins, when measured against the 741s output (virtual earth).

The system works, the issue I have is it doesn't amplify until 2.5mA is measured across the thermocouple.
 
Last edited:

Thread Starter

kutalinelucas

Joined Nov 20, 2007
98
Quite possibly, but it provides a stable ±6v so I don't think thats the problem...As I said it works, but only kicks into action once a PD of 2.5mA is presented across the thermocouple
 

John P

Joined Oct 14, 2008
2,025
It seems to me that your U7 and U8 are serving no function. They will generate half the power supply at their respective output pins, but there's no connection from there to Gnd as used elsewhere in the circuit. And why are they both there at all? Surely if you're generating a pseudo-ground, you only need to do it once.
 

Thread Starter

kutalinelucas

Joined Nov 20, 2007
98
Hi John

This is my first attempt of using op amps so I'm not too savvy quite yet. I lifted the voltage splitter circuit from his web page...http://tangentsoft.net/elec/vgrounds.html

I duplicated the voltage splitter circuit from the common 12v supply simply because I have two sets of op amps with individual thermocouples and didn't know if noise would be an issue, but I have removed one set of ICs anyway, until I can figure out the problem.

But I think that is what I'm stuck on...how do I ground the 741? I currently have a ground connection from the output to common ground through a 0.1uF capacitor, but this doesn't affect the behaviour of the circuit, ie, there is still no response until the thermocouple hits 2.5mA
 

Thread Starter

kutalinelucas

Joined Nov 20, 2007
98
Hi Dave

pins 4&5 of the 1025 are currently connected to common ground, but are you saying the gnd pin of the 1025 should be connected to vgnd? (pin 6 of u7)? what about -R?
 

Thread Starter

kutalinelucas

Joined Nov 20, 2007
98
Could you please tell me how you would connect vgnd to common? I have tried more ways than I care to remember in the past few days, and although the output behaviour changes, the fact that the op amp does not amplify until the thermocouple registers 2.5mA remains
 

John P

Joined Oct 14, 2008
2,025
Just with a wire. But before you do, be certain that you haven't assumed that either of your two power supply levels (+6 or -6) are actually Gnd. You need to have Gnd, +6 and -6 all entirely separate.

By the way, I agree with what others have said--the 741 is a poor choice for this project. The LM358 or LM324 are nothing great, but they'd be better.
 

Thread Starter

kutalinelucas

Joined Nov 20, 2007
98
Its the same story...if I connect them just with a wire the ltkAo op amp doesn't respond until the thermocouple hits the 2.5mA threshold, and the split rail voltages are deffinatly isolated (along with the power supply) from all other circuitry
 

t06afre

Joined May 11, 2009
5,934
Why do you NOT look at the tips in post 3:confused: The person behind that hint has a lot experience, and those solutions are also tested and working. But no you just keep on insisting doing it your way
 

Thread Starter

kutalinelucas

Joined Nov 20, 2007
98
But no you just keep on insisting doing it your way
Trust me I'm not stubborn, nor do I appreciate condescension. I do however have the PCBs manufactured, populated and all other systems calibrated and programmed. Its not as simple as just throwing in the towel and starting again. I'll just have to keep trucking on this, I'll figure it out eventually and hopefully learn some stuff along the way.

Thanks for all your time guys, I really do appreciate it

Martyn
 

t06afre

Joined May 11, 2009
5,934
A circuit board is very seldom made right the first time. Your problem is the power part. Use some stripe board. And make some circuits that you glue on top on your PCB board. Then keep the good part of your board as is. That is what any true engineer would have done in the real world. Then you have something that is working make a new board. Now boss in the world would ever hang you out to dry for this. Nor would any proper school/uni teacher
 

ErnieM

Joined Apr 24, 2011
8,377
You made a PCB before you tried the circuit? I knew you were trouble when you walked in.

Oh, oh... Trouble, trouble, trouble
 
Top