Pyranometer interface to Arduino

[Wassima AIT AHMED]

I found in another work where they tried to measure the radiation using a solar cell connected to a photovoltaic pannel that it may help to use th ADS1115 what do you think about it ? is it better that the AD624 or complementry or both can be connected to get better result?

 

[dknguyen]

Also, did you realize the AD624 only has a rail-to-rail output of +/-10V when running off 15V?
How does this detail affect my circuit ? as for the data i guess you are right im just concerend since the value at wich the error starts is 0.001 mV

It affects your circuit if you thought it was a rail-to-rail output because you lose 50% of your expected output range.

I found in another work where they tried to measure the radiation using a solar cell connected to a photovoltaic pannel that it may help to use th ADS1115 what do you think about it ? is it better that the AD624 or complementry or both can be connected to get better result?

The AD1115 is an ADC with a programmable gain amplifier (PGA) in it. They are two different things but you can use them together if you want. You could use the AD624 as a pre-amplifier and then adjust the PGA in the AD1115 to get a variable gain circuit with a 16-bit ADC.

I don't think the PGA in the AD1115 is good enough that it can connected directly to the pyranometer (if only because not a lot of data about the PGA's bias currents in the datasheet).

 

[Wassima AIT AHMED]

The AD1115 is an ADC with a programmable gain amplifier (PGA) in it. They are two different things but you can use them together if you want. You could use the AD624 as a pre-amplifier and then adjust the PGA in the AD1115 to get a variable gain circuit with a 16-bit ADC.

I don't think the PGA in the AD1115 is good enough that it can connected directly to the pyranometer (if only because not a lot of data about the PGA's bias currents in the datasheet).

That is what i thought of doing also finaly i will combine both amplifiers , thank you so much for your help , but i will be greedy and ask you if you could provide me with a schematic possible for this circuit im quite new to the field and i can't find if i shouf add other components or not

 

[Wassima AIT AHMED]

As for the rail to rail detail i need to get all the data at the output of my circuit and specialy if the input signal is quite low , is combining AD1115 and AD624 is the solution or there is an alternativ for that ?

 

[dknguyen]

That is what i thought of doing also finaly i will combine both amplifiers , thank you so much for your help , but i will be greedy and ask you if you could provide me with a schematic possible for this circuit im quite new to the field and i can't find if i shouf add other components or not

That's bad news on your end then since schematics as well as physical circuit layout is very important for 16-bit ADCs since it's easy to make things so noisy that you can't use the full 16-bit resolution.

If you're new to this, then I suggest you just use one AD624 with a modest amount of gain as a pre-amplifier that will drive the inputs of multiple AD624s where each one has a different gain.

Then you just always choose use the output from the the AD624 with the highest gain that does not have a saturated output signal. You will have to normalize the readings from all the amplifiers afterwards (since they have different gains) with some multiplication, but this is the simplest way to maintain resolution over a wide input range.

 

[Wassima AIT AHMED]

Im sorry but you lost me in the last comment i didn't quite get what you mean can you elaborate more !! I was thinking on combining both amplifiers realise a printed circuit but now i don't know anymore !!

 

[dknguyen]

Im sorry but you lost me in the last comment i didn't quite get what you mean can you elaborate more !! I was thinking on combining both amplifiers realise a printed circuit but now i don't know anymore !!

For example, suppose:

-Input signal range is 1uV to 1mV
-Desired output range is 1V.

I can use a preamplifier with a gain of 1000 This preamplifier feeds one (or more) amplifiers. For this example let's just pretend there is only one with a gain of 50.

If my input is around near 1mV, I use the output of the preamplifier directly so my output voltage is as high as possible without exceeding 1V.

If my input is very low (between 1uV and 20uV), I use the output from the x50 gain amplifier for a total gain of 50,000. If the input becomes higher than 20uV the output range exceeds 1V so I must switch to an amplifier with a lower gain.

To know which amplfiier to use I would start reading the amplifier with the highest gain and make sure it doesn't read near the max output of 1V. If it does, I move to an amplifier with lower gain until it doesn't read very close to 1V anymore. Then I will have the amplifier with the highest gain that doesn't saturate the output.

In this way, I can measure very low voltage signals with good resolution and high voltage signals without saturating my output. It's just like manual ranging on a multi-meter.

 

[Wassima AIT AHMED]

i start to understand what you mean actully, and that was my first idea is to use a preamplifier and an amplifier after, however i got stuck on finding an amplifier with a gain=1000 without noises , any suggestions ? also the switching to amplifiers part how can i realize it without automatation wusing simple analogic devices ?

 

[dknguyen]

i start to understand what you mean actully, and that was my first idea is to use a preamplifier and an amplifier after, however i got stuck on finding an amplifier with a gain=1000 without noises.

That's the same difficulty you get when using a 16-bit ADC. There is no easy way around it:

1. Quiet power supplies
2. Good amplifiers
3. Low noise resistors
4. Filtering
5. Non-microphonic capacitors (like C0G/NP0 ceramic capacitors)
6. Good multi-layer printed circuit board layout.

It's not an easy task. The gains in my example were not meant for you to use. They were quickly chosen to be easy to understand. You have to decide how many gains you want to get the results you want.

also the switching to amplifiers part how can i realize it without automatation wusing simple analogic devices ?

For high speed signals, there's no practical way to do this without a microcontroller that automatically decides what amplifier to choose and then multiplies all the ADC readings numerically so they are on the same scale.

But this is a pyranometer, right? So very low speed? Just hook up a a multi-meter to each amplifier at the same time and record the reading from the meter that is closest to, but not exceeding your output range. Also remember to record which amplifier the reading was from so you know the gain/scale.

 

[Wassima AIT AHMED]

Even if it is an example it is exaclty the situation here, the pyranometer delivers voltage values of uV scales at times an mV at others that is why i asked wich amplifier can assure a gain of 1000

 

[Wassima AIT AHMED]

As for the output it is intended to be connected to an arduino board in order to store the data given by the pyranometer, so the schematic you suggeted would be quite hard! the problem is the link between the pyranometer and the arduino which amplifiers ? which connections? how to avoid noises and assure a goode reading of the values at the erail monitor given by the arduino software on the computer ??

 

[dknguyen]

i start to understand what you mean actully, and that was my first idea is to use a preamplifier and an amplifier after, however i got stuck on finding an amplifier with a gain=1000 without noises , any suggestions ? also the switching to amplifiers part how can i realize it without automatation wusing simple analogic devices ?

The noise critical parts of the board would mainly be between the pyranometer and the pre-amplifier where signal voltage is lowest. If you can get that worked out it should get easier along the signal path where the signal voltage is much stronger. After that it shouldn't be too much of a problem anymore since your Arduino only has a 10-bit or 12-bit ADC anyways, right? So linking the amplifed signal to the Arduino isn't an issue since the resolution is only 5mV or 1mV anyways. It would be a problem if the Arduino was a 16-bit ADC or if you were feeding the amplified signal to a 16-bit ADC.

It's mainly just the pre-amplifier so the PCB isn't going to be too large and complicated. It's just going to be picky so you could maybe do 2 or three test boards to try different things out to see what works and what doesn't. You won't be able to use a solderless breadboard something like that. You might get away with a soldered breadboard if it has a ground plane and you use point-to-point wires. Then build that into a larger board with all the post-amplifiers since it's best if they're on the same board anyways. It's certainly nowhere near as difficult as accomodating a 16-bit ADC which is both picky and has a lot of components that are all very sensitive.

Also...don't feed a 10V signal into your ARduino's ADC...that's only going to be 3.3V or 5V (I'm not sure whch but you should know).

 

[Wassima AIT AHMED]

So what can be the best suggestion ? wich amplifier should i use ? that is my question and how does the circuit look like ? And how can I fixe the noise problem between the pyranometer and the preamplifier ?

 

[dknguyen]

Is this a photodiode or a thermopile?

I'm going to assume it's a thermopile for the time being, but if it's a photodiode it's going to require different circuits and amplifiers.

I won't draw the circuit out for you but you should be able to figure it out.

1. I'd choose something like the AD8551 or AD8628 which have very low offset voltage and very low bias currents. Low bias currents introduce less error when using large resistnaces in the amplifier. Offset voltages get amplified along with the signal too so if you're after 1uV, then the offset voltage needs to be around that or less. I would use these in a non-inverting configuration for higher input impedance.

2. Use only quiet linear voltage supplies. Really quiet ones. You're trying to measure 1uV so the noise on these regulators shouldn't be much more than that.

3. Only use thin film resistors and C0G/NP0 capacitors for any resistors of capacitors in the signal chain. (less noise and no microphonics).

4. Keep components close, and circuit traces as short as possible.

5. Use ground planes and bypass capacitors on everything.

6. You probably need a PCB shield cover on the board over the pre-amplifier portion of the circuit. I'm not sure.

7. Thermopiles measure the temperature difference so shield it from wind and mount the thermopile in an isothermal block (like a big block of copper, steel, or aluminum) to keep the internal temperature constant.

8. You might need to do things like make guard rings with PCB traces to stop leakage currents from messing with things. I'm not sure.

9. Power it off a battery to test. At least that way you know there won't be noise or ripple from your power supply. Once you know what you have you can try using other things to power the board.

10. Do not power any digital electronics off the linear regulaaor powering the pre-amp.

11. Try placing RC filters on the preamp output. I dont think its a good idea to put them at the input since the signal level is so low but it might also be necessary.

I would first try to see how good you can get it using a soldered surface mount protoboard that has a groundplane using (26-30AWG) self-fluxing magnet wire to make all the connections.

Something like this https://www.busboard.com/documents/datasheets/BPS-DAT-(SMTpads)-Datasheet.pdf

 

[Wassima AIT AHMED]

The pyranometer ? thermopile i guess

 

[Wassima AIT AHMED]

and if it is a photodiode ? then ?

 

[dknguyen]

Th

and if it is a photodiode ? then ?

Then it's harder (maybe?) but you also get to use a lens which is noise free gain.

Some of this material is focused on low noise and high speed but you only need low speed so maybe it's just a different circuit but no harder.

 

[Wassima AIT AHMED]

Alright i will check the data you gave me and come back if i have more questions thanks a lot !!