Maker Pro
Maker Pro

Op Amps

Xenobius

May 15, 2012
125
Joined
May 15, 2012
Messages
125
Hi,

I have an oxygen sensor which has just 2 wires and outputs a minimum of 7mV and a maximum of 11mV. I calculated that if I amplify this voltage by 450x, the output will be suitable for a micro controller to read from a 10bit ADC.

My question is this. I have never used Op Amps and I have read tutorials about them but I'm still not able to design that part of the circuit.

Would anyone kindly guide me to build the Op Amp portion of the circuit such that I input the sensor and output directly to uC amplifying the voltage by 450?

http://oxycheq.com/index.php?main_page=product_info&cPath=1_6&products_id=22

Thanks a lot
X
 

Xenobius

May 15, 2012
125
Joined
May 15, 2012
Messages
125
Wow I think that's perfect. I have 2 last questions.

1. Do you think that I should be using less gain? What would be the least amount of voltage for a 10bit ADC pin to read without difficultu? Atmel uC

2. Do you know of a different way of doing it without using 2 op amps? Mainly asking this for knowledge purposes not because your solution isn't satisfactory :)

Thanks
X
 

Six_Shooter

Nov 16, 2012
98
Joined
Nov 16, 2012
Messages
98
What O2 sensor is this that only works from 7 to 11 mV?

Any (narrow band) oxygen sensor I have dealt with is a 1V range.
 

Laplace

Apr 4, 2010
1,252
Joined
Apr 4, 2010
Messages
1,252
I believe the 7 mV - 11 mV is the calibration voltage for the partial pressure of oxygen in a standard atmosphere. The sensor output voltage will be higher or lower depending on the concentration of oxygen.

Also, as an example, the Atmel AT8xEB5114 version of the 80C51 8-bit microcontroller has an internal voltage reference of 2.4 V, so the maximum input to the ADC analog input would be 2.4 V, unless you substitute a different external reference.
 

Attachments

  • schem_b72149.pdf
    86 KB · Views: 90

OLIVE2222

Oct 2, 2011
690
Joined
Oct 2, 2011
Messages
690
Don't know about the oxygen sensor output range, the following link state an output current of 330-590µA and being compatible with Teledyne. According to the Teledyne spec you have posted a 10K load must be used.

http://www.yorkinstrument.com/york/html/2010/9/964.htm

Maybe be the OP have already tested the sensor.

By the way to have a better mapping with the µC ADC you can have a look at the SINGLE-ENDED ANALOG STAGE paragraph of the following document to build the first stage


http://people.ece.cornell.edu/land/courses/ece4760/analogDigital/PracticalAnalogDesign.pdf


Q1/
The gain is define by the ADC voltage reference, can be 1.024V or 2.5V or other. The total gain must almost allow you a full scale reading, if not you will lost resolution.

Q2/
Yes you can use an instrument amplifier with gain up to 1000. The need 2 or 3 op amp but modern one are integrated in a single package. By the way dual opamp are also available in single package.

http://www.analog.com/en/specialty-...roducts/index.html#Instrumentation_Amplifiers

Olivier
 

Xenobius

May 15, 2012
125
Joined
May 15, 2012
Messages
125
Hi thanks for yoru replies.

The sensor is this

http://oxycheq.com/index.php?main_page=product_info&cPath=1_6&products_id=22

and it states the following

This 7 to 11mV +/- oxygen sensor has a 6 second response time and a 36 month life expectancy in air. Range 0-100% O2, +/- 1%.

So its mV not current and 11mV = 100% oxygen and 7mV = 0% oxygen I assume but there is something I can cater for in the software.

As for loosing resolution, you're right that is why with my assumption to use 5v on ADC, a 450gain would take the 11mV upto 4.95V thus within range and not loosing precision.


Actually I am also looking for a helium sensor too which might have different readings but once I know my way around Op Amps, I can do the rest

:D
 

OLIVE2222

Oct 2, 2011
690
Joined
Oct 2, 2011
Messages
690
For the resolution you also need to normalize the amplifier (see the Circuit cellar article) if not you will have 3.15V to 4.95volts values which correspond to around 2/5 of your resolution only!

better to limit the amplification as needed, a gain of 225 with a reference of 2.5V is better so. Look at your µC chip for internal reference(s) value(s). Also sometimes the adc neg ref can be user setted above GND, this is an asset as you can set it to 3.15V for instance.

Olivier
 

Xenobius

May 15, 2012
125
Joined
May 15, 2012
Messages
125
Ok so here is the sensor part that I've designed based on the PDF that you guys gave me. Because I have absolutely NO CLUE what this is doing, would someone check it out for me?

Basically all I did was COPY a circuit without the 4 channel ADC IC as I will use the uC's ADC pins. But the circuit was designed for thermocouples which although they output mV just like my oxygen and helium sensors, might be different...

Any confirmations would be greatly appreciated.

I will then decide if I should use 3.3v or 5v and on the gain.

Thanks
X
 

Attachments

  • sensor-part.jpg
    sensor-part.jpg
    86.2 KB · Views: 120

Xenobius

May 15, 2012
125
Joined
May 15, 2012
Messages
125
For the resolution you also need to normalize the amplifier (see the Circuit cellar article) if not you will have 3.15V to 4.95volts values which correspond to around 2/5 of your resolution only!

better to limit the amplification as needed, a gain of 225 with a reference of 2.5V is better so. Look at your µC chip for internal reference(s) value(s). Also sometimes the adc neg ref can be user setted above GND, this is an asset as you can set it to 3.15V for instance.

Olivier

Ok so I checked the datasheet and if I use an ATmega328P I have an internal reference of 1.1V or I can supply my own.

I don't understand how I can loose precision with the following understanding.

The sensor reads 0 to 100% oxygen +/- 1% (datasheet)
The sensor outputs 7 to 11mV to indicate the percentage (datasheet)

If I amplify the output by 450x I get a range from 3.15V to 4.95V, than the 10bit decimal will be from 645 to 1012 (a maximum of 1023)

Is that loosing precision? I always thought that if my amplified value is greater than 1023 will be loosing precision... interesting to learn this :)

Thanks
 
Last edited:

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
OK, I can't find a datasheet for these. Can you point me at one. I'm sure the 7 to 11 mV means something, but I can't imagine it's the difference between zero and full scale reading on the device.
 

Xenobius

May 15, 2012
125
Joined
May 15, 2012
Messages
125
You're right... it doesn't make sense.

This is another specification of a different model from Maxtec (its the most alaborate thing I found so far. I've emailed them but still waiting an answer.

Specification http://www.maxtec.com/docs/tech_specs/MAX-11i-specs.pdf

This is the manufacturer (there is no direct link. you need to find Max 11I to match the same specification) http://www.maxtec.com/fio2/fio2.php#

Here in this model it says:

SPECS
Output17.0 ± 3.0 mV
Response Time≤ 12 seconds
Linearity< 3.0 % of full scale
Warranty12 month

I dont understand why for example here it just gives out 17mV +/- 3mV
I am getting a feeling its actually 17mV per 1% of Oxygen...
 

OLIVE2222

Oct 2, 2011
690
Joined
Oct 2, 2011
Messages
690
Here an AN for an oxygen sensor. The output is specify as a current, when stated as voltage a load is specified, intrinsically the are current source to me. In the AN depending on the model we have 30 or 50mV full scale on 10K.

http://www.alphasense.com/industrial-sensors/pdf/AAN_003.pdf

Olivier

edit: from the spec above should be 17.0 ± 3.0 mV full scale
 
Last edited:

Xenobius

May 15, 2012
125
Joined
May 15, 2012
Messages
125
... damn... its getting complicated all of a sudden...

I didn't quit understand what you just said. I think I'm headed in the wrong direction...:rolleyes:
 

Xenobius

May 15, 2012
125
Joined
May 15, 2012
Messages
125
Hmmph now that makes a lot of sense :) So are all sensors like that?
So if the circuit is sensing current, how does the op amp amplify that and feed it into a voltage-reading-adc-pin?

Thanks for your time :)
X
 

OLIVE2222

Oct 2, 2011
690
Joined
Oct 2, 2011
Messages
690
All sensor based on the same technology should work like that.
The current is converted to voltage by the (internal or external) load resistor.

You can maybe go for a mockup to validate that; By the way how do you intend to calibrate the sensor conditioner ?
Olivier
 

Xenobius

May 15, 2012
125
Joined
May 15, 2012
Messages
125
I can take a reference of 0% , 50% and 100% oxygen and calculate the rest. I was going to tackle the calibration part of the project later on. I have sources for any oxygen percentage and I can use them to calibrate (if that's what you meant)...

The final product will have 2 sensors (one for helium and one for oxygen) both of which will be connected with the same pipe. I will first input 100% oxygen and use the results from both sensors to calibrate them. Helium should be 0% and oxygen should be 100%.

I will then input 100% helium and do the same. Oxygen should be 0% and helium should be 100%.

Any other ideas? or perhaps you where talking about something different?
Calibration of this sort would have to be done every time a sensor expires and is replaced.

X
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Assuming you use this for the intended purpose, I would recommend that you test it with a mix of 18 to 20% O2 as that is the critical point.

0% and 100% are irrelevant as far as this device's purpose goes.

I would be concerned with the devices performance over temperature too. Is this designed to be operated at depth? If so, the temperature is likely to be far lower than surface air ambient temperature and with very low input voltages you will be concerned with how things like input offset and gain change with temperature as they are likely to affect the output voltage.

(Of course, you're unlikely to measure how these things change, just assume this is the cause based on the changing output. A design which minimises the effect of these variables (perhaps a chopper stabilised zero offset op-amp?) might be in order).
 
Top