• State TI Thinks Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 5 answers
  • Subscribers 46 subscribers
  • Views 959 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Connecting solar reference cell to ADS1248

kosma
Prodigy 40 points
Other Parts Discussed in Thread: ADS1248, LM7705, ADS1220, ADS1220EVM

Hi all,


This is my first post here so please be understanding :-)

I need to measure precisely voltage across solar reference cell rated 32mV for 1000 Wm-2.  I choose ADS1248, because I need second input (apart from voltage) - temperature - to be measured in order to calculate solar irradiance.

Reference cell has PT100 sensor inside it's package and I built breadboard prototype with PT100 connected to ADS1248 which is actually working. Now I would like to know how to connect solar reference cell ( two 2 meters long leads: + and  - ) to ADS1248 inputs?

I think of single-ended configuration (AIN positive to cell + and AIN negative to cell - and to analog GND) but then I would have to seek for bipolar analog supply (using for e.g. LM7705 low noise negative bias generator) rather than single ended. Is there another solution which could make use of fully differential input configuration and/or possibly single supply?

Thanks in advance for your help,

Kosma.

  • 0
    TI__Guru* 97835 points
    Kosma,


    The LM7705 may help in certain applications, but it may depend on what input range you expect to use. Is the input range only 32mV? or will there more solar cells to make a larger input signal?

    If you are using the ADC in a low gain, I would use the ADS1220 instead. It is similar in function, has the second input channel and the input range would go to AVSS-0.1V to AVDD+0.1V. Because gains 1 to 4 are implemented without the PGA, you'd be able to use this without the LM7705 and be able to get a gain=4 with the actual PGA disabled. This might be enough for you depending on the required resolution.

    If you are using the ADC in higher gain, then the LM7705 may still not be able to help. Going back to the ADS1248, since the PGA is implemented as an instrumentation amplifier, you still need to worry about the gain violating the output common-mode voltage of the PGA. Since the input signal and the output gained signal measured by the ADC have the same common-mode voltage, you might violate the output range with a large gain. There's a discussion about this on the following post.

    e2e.ti.com/.../1383715

    If you read through the entire thing, there's a description through a link in one of the posts about the input voltage range requirement. There's also a ppt file describing the input range with a few examples.

    I'm not sure if there are any other simple solutions. It might depend on the solar reference cell and if it could be referenced to mid-scale. If you need higher gain, I think the only solution is to measure the signal at the ADC's mid supply.


    Joseph Wu
  • 0 in reply to Joseph Wu
    Prodigy 40 points

    Thank you Joseph for your answer.

    Input range is only 32 mV since there will be only one cell connected to ADC. I would like to get best possible resolution so I think PGA with high gain would be recommended. Satisfying resolution would be 0.032 mV (32 uV) - can it be achieved using ADS1220 with PGA disabled?

    I made some input range calculations for single-ended measurement using ADS1248 + LM7705 and result shows that in this configuration even for gain 1 common-mode voltage requirement is violated. As ppt presentation mentioned above state "Single-ended measurements using a unipolar analog supply are in general not possible with devices utilizing a true PGA!".

    VINP VINN VIN VCM GAIN VCM_MIN
    0 0 0 0 1 0,025
    0,02 0 0,02 0,01 2 0,42
    0,04 0 0,04 0,02 4 0,74
    0,06 0 0,06 0,03 8 1,38
    0,08 0 0,08 0,04 16 2,66
    0,1 0 0,1 0,05 32 5,22
    0,12 0 0,12 0,06 64 10,34
    0,14 0 0,14 0,07 128 20,58
    0,16 0 0,16 0,08
    0,18 0 0,18 0,09
    0,2 0 0,2 0,1
    0,22 0 0,22 0,11
    0,24 0 0,24 0,12
    0,26 0 0,26 0,13
    0,28 0 0,28 0,14
    0,3 0 0,3 0,15
    0,32 0 0,32 0,16
    AVDD 3,3
    AVSS -0,235

    If it doesn't work that way, then I should choose ADS1220 which is intended for single-ended measurements. I have a doubt if ADS1220 can work with unipolar supply (such as AVDD=3.3V, AVSS = 0V) in single ended configuration (with negative input connected to ground), because these ICs look so similar. Will ADS1220 work in this fashion?

    Could you please expand the subject of measuring solar cell referenced to mid-supply? Does it mean that one cell lead (-) is connected to AINN configured to be at mid-supply potential while other cell lead (+) is connected to positive AINP? How does it change the measurement performance?

    Regards,

    Kosma.

  • 0 in reply to kosma
    TI__Guru* 97835 points
    Kosma,


    I think that the ADS1220 could work for you. I'll go through some of the characteristics of the device and show what features are important.

    First the input range of the device when the PGA is bypassed is from AVSS-0.1 to AVDD+0.1. In this case you could use AINN as ground and still make measurements without having to violate the input common-mode range. Also, with the PGA bypassed, you still can get a gain of 4 and still not have to worry about the PGA range. This is discussed on page 23 of the datasheet.

    Now look at Table 3. THe noise is 3.91uVpp in Gain of 4 with the PGA disabled. This is for a data rate of 20SPS, you could get a faster data rate but you would sacrifice noise performance.

    If you are in Gain=4, with a 2.048V reference, that means the LSB size is 2.048V/2^23 or 0.24uV.

    Of course you'd still need to have the low noise reference (or the internal reference should be low noise enough) to give you resolution down below your desired 32uV. You'd also need to make sure the supplies were free of noise so that any power supply rejection could limit the noise contribution. Any design should also consider EMI effects in the layout.

    The ADS1220 is a bit better suited to your application because in the ADS1248, the PGA cannot be bypassed, that limits the ADC with the input and output common-mode range of the PGA. There's little way to get the ADS1248 to work outside a gain of 1. (I think that the in PGA=1 it still works - Since the common-mode limitation on the negative side is AVSS+0.1=-0.135V)

    In a previous post, I had suggested that you use a mid-supply measurement with the ADC, so the PGA could be implemented. However that would require you to generate a bipolar supply (maybe +/-2.5V) so that the solar cell negative output could be referenced to ground. However this might be difficult to do without a switched supply and an LDO to clean up the power supply noise.

    Based on the features and performance, you should be able to make this work. I'd recommend getting an ADS1220EVM and you should be able to test this right away. You should at least download the user guide. There's information and recommendations for it's use and you can look at a potential circuit and layout.


    Joseph Wu
  • 0 in reply to Joseph Wu
    TI__Guru* 97835 points
    Kosma,


    One of my colleagues pointed out a couple of other things. If the cell has a low output impedance, you need to consider the input current (which may be several nA). The PGA would lower the input current if that could be used.

    Another thing is that if the solar cell is a floating output, then you could use the ADS1248 in gain, and set the negative input with the VBIAS setting (which is described in the ADS1248 datasheet).


    Joseph Wu
  • 0 in reply to Joseph Wu
    Prodigy 40 points
    Thank you Joseph. Your answers are very informative.

    Your previous posts convinced me to use ADS1220 with PGA disabled, but now we know that there are more things that should be taken into account.

    Solar reference cell is internally loaded by (probably 1 ohm) resistor so that short-circuit current can be measured with a voltmeter. Then no doubt that cell has low output impedance and I should consider input current. I guess that you mean that delta-sigma modulator shouldn't be loaded with current from signal source and PGA could lower this loading effect by acting as buffer. What is an effect on the measurement with input current being neglected?

    If I had to use PGA then I need low noise bipolar supply (such as +/-2.5V). Are there any ICs that can provide such bipolar supply? And what about LDO for negative voltage supply?

    I don't know if I understand what you mean by floating output. If any cell lead (negative in particular) is connected to some potential it is not floating anymore. Now it is referenced to this particular potential.
    What do you mean by using ADS1248 in gain?
    Setting VBIAS connects selected input to midsupply (AVDD+AVSS)/2. What does it change with floating cell output? Is it used with bipolar supply?

    Regards,

    Kosma.
  • 0 in reply to kosma
    TI__Guru* 97835 points
    Kosma,


    Sorry I haven't replied, but I've been out of the office.

    For your questions, a bipolar supply would require some sort of negative charge pump or DC-DC converter and then an LDO to clean it up. I'm not familiar with any good devices to easily generate this.

    For a floating output, I mean something that has no DC output bias. Sometimes the negative input is connected to some ground or earth ground. In this case it is not floating. If the output is not connected in this way, it is floating.

    By using VBIAS, you might be able to connect the negative output of the solar cell reference to VBIAS to give it some DC voltage. However, if the negative output is connected to the ground then the VBIAS cannot be used to drive the input to mid-supply.


    Joseph Wu
  • 0 in reply to Joseph Wu
    Prodigy 40 points

    Joseph,

    Maybe it is a good idea to ask question about bipolar supply in e2e power managment forum.

    Solar reference cell I am using is indeed floating output (any lead can be connected to any potential). With VBIAS enabled on negative output of reference cell I could connect AVSS to ground without having common-mode voltage range exceeded. I made some calculations introducing this configuration:

    VCELL VINP VINN VIN VCM   GAIN VCM_MIN VCM_MAX
    0 1,65 1,65 0 1,65   1 0,116 3,184
    0,002 1,652 1,65 0,002 1,651   2 0,132 3,168
    0,004 1,654 1,65 0,004 1,652   4 0,164 3,136
    0,006 1,656 1,65 0,006 1,653   8 0,228 3,072
    0,008 1,658 1,65 0,008 1,654   16 0,356 2,944
    0,01 1,66 1,65 0,01 1,655   32 0,612 2,688
    0,012 1,662 1,65 0,012 1,656   64 1,124 2,176
    0,014 1,664 1,65 0,014 1,657   128 2,148 1,152
    0,016 1,666 1,65 0,016 1,658      
    0,018 1,668 1,65 0,018 1,659   GAIN VIN VPGA_OUT
    0,02 1,67 1,65 0,02 1,66   1 0,032 0,032
    0,022 1,672 1,65 0,022 1,661   2 0,032 0,064
    0,024 1,674 1,65 0,024 1,662   4 0,032 0,128
    0,026 1,676 1,65 0,026 1,663   8 0,032 0,256
    0,028 1,678 1,65 0,028 1,664   16 0,032 0,512
    0,03 1,68 1,65 0,03 1,665   32 0,032 1,024
    0,032 1,682 1,65 0,032 1,666   64 0,032 2,048
              128 0,032 4,096
    AVDD 3,3        
    AVSS 0     GAIN VIN VFULL_SCALE
    VREF 2,048 1 0,032 2,048
    2 0,032 1,024
    4 0,032 0,512
    8 0,032 0,256
    16 0,032 0,128
    32 0,032 0,064
    64 0,032 0,032
    128 0,032 0,016

    As you can see with Gain = 128 VCM_MIN is higher than VCM_MAX, Full Scale Voltage is higher than expected maximum input voltage (32mV) and voltage at PGA output exceeds supply rails (0-3.3V). So the highest gain I could use in my application is 64.

    I could use also ADS1220 with gain of 4, having PGA disabled, solar reference cell connected to inputs in single-ended fashion. But what about input current considerations? If ADC has fairly high input impedance compared to reference cell output impedance is it still a matter?

  • 0 in reply to kosma
    TI__Guru* 97835 points

    Kosma,

    If the gain of 64 is the highest you'll be able to use, you'll still be able to drop the noise floor significantly. If the solar reference is really a floating output the VBIAS solution might work best.

    As for the input current, the biggest factor would be the output impedance of the solar reference. If this is moderately low, then the input current shouldn't be a problem. I'm not familiar with what you are using, but it's something you can calculate out if you can reference a datasheet to the cell. I would note that with the PGA disabled, the input current is still very low, less than 50nA even up to 85C. There are graphs showing the input current in the Typical Characteristics curves in the datasheet for the ADS1220.

    Joseph Wu