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.

Analyzing Wirewound Coil without an Impedance Analyzer?

Other Parts Discussed in Thread: LDC1000EVM

Alfred Gomes' video about adapting external coils was cool except I do not have an impedance analyzer...

Are there less expensive ways to characterize external coils for the Rp parameters?  I have access to digital scopes (Lecroy 1GHZ+), impedance meters (handheld and Sencore LC102), and function generator to maybe 2MHz.

I would like to experiment with adapting to existing or homemade metal detecting and pinpointing coils.  These existing coils are designed for a variety of detector architectures - pulse induction, and a variety of frequencies (the pinpointers usually have a ferrite core).

  • There is this: http://www.sillanumsoft.org/ZRLC.htm

    Does it look like it will give us the parameters we seek?

  • Hello Barry,

     

    If you don’t have a network analyzer or LCR meter, you can use the LDC1000 EVM to characterize your coils for the RP parameters. Simply break off the coil portion of your EVM, and attach your target LC tank.

     

    If you are using a value other than 100pF for the LC tank capacitor, please set the sensor capacitor field on the GUI to the appropriate value.

     

    Then, follow the process from our forum FAQ to optimize your RPmin and RPmax settings:

     

    1) Set Rpmin to 3F, RpMAX to 0.
    2) Expose the coil to the maximum metal coverage for your application (closest target position, thickest part etc)
    3) Start reducing RpMIN setting 1 code at a time, and take Rp Measurements (they will go up with each change).To speed up tuning, it is also possible to change the value by more than 1 code at a time, for example by using binary search.
    4) When Rp gets in the range of 20,000 - 30,000 codes, that is your optimal RpMIN setting.
    5) Move your target to a position where it's exposed the least (farthest position, thinnest part) etc.
    6) Start increasing RpMAX setting 1 code at a time, and take Rp Measurements (they will go down with each change).
    7) When Rp gets in the range of 2,000 - 3,000 codes, OR the difference between RpMIN and RpMAX values reaches x25 - x26, that is your optimal RpMAX setting. It may get too noisy, then back off 1-2 RpMAX codes.

     

    After this process has been completed, you can take the Proximity Data from the GUI and calculate RP as described in the datasheet and in the FAQ:

     

    R­p = (RpMAX * RpMIN) / [RpMIN*(1-Y) + RpMAX * Y]

     

    Where Y = Proximity Data / 215, and RP will be in Ohms.

     

    Then you will have the RP of your LC tank.  

     

    Thanks,

    Dan Lanzone

  • I am trying to get a grasp on how to start at this.  My goal is first to match performance of a metal detecting probe I have that runs at 300KHz and has an adjustable inductance probe of ~140-300uH.  There has been a lot of work done on which frequencies excite eddy currents best on which metals - but this has been done for metal detectors and not necessarily close proximity probes.  I need to understand how to select a coil for the purpose I have in mind while at the same time have it be compatable with the LDC1000.  So for me, like I suggested in another thread, I'd like to have pc "knobs" for the various register settings as then I could manipulate them for different ground conditions, coils, and targets.  I would want maximum sensitivity at the expense of noise - testing the brain's ability to identify the target by sound.  I have a ton of different detecting coils - made for VLF, PI, and other types of detectors.  Most of these have TX and RX coils - some of the PI coils are mono though.  I can also easily make my own coils - just wonder if there is an inductance sweetspot to design to.  Perhaps this varies per frequency. 

    Your method for determining Max/Min of Rp depends on close and distant placement of the target - but my target will be very small and as distant as I can get away with....  I would think that there is a optimum inductance to design to and then a capacitor is chosen to resonate it.  If I change the LDC1000 frequency wouldn't I then be best to change the capacitor value?

  • Hello Barry,

    There is no real inductance sweetspot to design to, it is about working within the limits of the coil you have. For instance, I just drove a 6mH coil with a 420pF capacitor, and that operated around 80kHz. The self resonant frequency of the coil was 160kHz, so that was the limiting factor when selecting the capacitor.

    When designing our coils, we're more concerned about what exactly we are trying to do, and this is very application specific. You can find an explanation of this in our faq at: http://e2e.ti.com/support/data_converters/inductive-sensing/f/938/t/295036.aspx#Q1. The gist of it is that for a high sensing resolution, with a target size comparable to the size of the coil, the target can be sensed at a distance of up to half of the coil diameter. Presence detection can be sensed at up to a coil diameter away.

    We are also constrained by the minimum and maximum equivalent parallel resistance that the LDC1000 can measure, which is 800Ω and 4MΩ, respectively.

    In short, design your coils with the guidelines listed in the FAQ in mind. At smaller inductances, the parasitics will be more of a factor, but at high inductances you will be constrained by the self resonant frequency.

    Thanks,

    Dan Lanzone

     

  • You guys keep responding to my inquiries like i know my target and i have a coil already.  My target can be big or minute and i want to discern whether it is ferrous or not.  Your sensor outputs frequency and proximity data and that's it - right?  So how do these outputs vary in my test case?  Your eval gives proximity and inductance - inductance is a calculation.  I want an audio response that gives me indication of material type and proximity/size.  If i only needed to reliably detect a door closure i'd have no issue using this part - but i am asking about using it to gain as much information about the target as i can get.  Broadly the goals should be:  probe coil diameter perhaps 1/2 - 1" or so, discern gold from iron/black sand.  I don't know the inductance but likely 100-200uH range?  Tank capacitance choice depends on set frequency right?  Should i strive for lower frequency so if i use a probe cable its capacitance has less effect?

    I am also wondering what possible uses the cfb pin may provide.  I haven't looked at it yet - do you see uses for this signal?

    Maybe discuss the setting of the Response Time register and how this affects results as well.

  • Hi Barry,

    1. The sensing distance,

    2. expected resolution (in the case of gold / iron sand, what is the size you want to detect?),

    3. material (e.g. to sense ferrous materials you would have to work at a lower frequency, but that is also dependent on the size of the material due to different penetration depths for different frequencies)

    4. sampling rate

    are all critical to determining the L and C parameters of your sensor.  The LDC1000 data is read digitally through a SPI bus, so you would need to have a microcontroller at the minimum to interface and drive a speaker for audio.  We do not test or support any other way to retrieve inductance information from the chip.

    The Cfb/Cfa pins are for the filter capacitor only, the selection of which is described in the datasheet.

    The response time register affects the output data rate; again, this is all documented in the datasheet.

    Keep in mind that the more specific you can be on these 4 sensing constraints the more specific we can be on the support.

    To get an idea of what you need, you can try out the WEBENCH coil design tool on the product page:

    http://www.ti.com/product/ldc1000

    Thanks,

  • Jeez, been almost a year and I'm back.  No I'm no smarter - dumber because I've been away from the eval kit too long.  Anyway, Dan suggested a process where RpMIN is set "1 code at a time" and to "take Rp Measurements".  Also mentions "binary search".  How is a" "Rp Measurement" done"?  I don't see such a data field in the GUI unless you are talking about the Proximity data values...  The title of the thread is "...without an Impedance Analyzer" so what are Rp Measurements?

    I'm also confused by the datasheet (mine is Dec 13).  What is with Figure 12?  It doesn't seem to be discussed.  Its RpMIN/MAX values are not the same as discussed in the next page.

  • Hi Barry,

    I'm sorry you had these issues with the LDC1000EVM and GUI.

    To clarify - yes, you are correct, the RP measurement is labeled as Proximity in the GUI.

    As for the datasheet, we will have an updated datasheet in the next few months. You can refer to the LDC1000-Q grade datasheet (http://www.ti.com/lit/ds/symlink/ldc1000-q1.pdf). The package is different, but the functionality is not changed.

    As for Figure 12 - it shows an example output code for a given sensor RP, with the RPMIN=16.16kohm and RPMAX=48.481kohm. As an example, if you had those settings and the sensor RP=37kohm, you would have an output code from the LDC1000 of ~5000.

    Regards,

    ChrisO

    Regards,

    ChrisO 

  • Ok, I can change register data as you suggest.  Any chance you guys could improve the GUI by having a start/stop button in the Config/Register pages to save clicks?

    Is it correct to refer to the Proximity value as Rp?  I'd think you'd want to call it something else to avoid confusion.

    I'm still confused about Figure 12 or even Table 10 in the new spec - same deal.  The RpMIN/MAX numbers referenced in the table title differ from those discussed in the spec text.  RpMIN = 16.160K vs 2.394K etc. even though like you say RP=37K = 5000 count.  

  • Hi Barry,

    Yes, it is correct to refer to the "proximity" as RP; the GUI is a bit confusing in that manner.

    Yes, you are correct, the figure settings do not align with the text; but it is provided as an example response.

    Regards,

    ChrisO