• State TI Thinks Resolved
  • +1 person also asked this people also asked this
  • Locked Locked
  • Replies 3 replies
  • Answers 2 answers
  • Subscribers 80 subscribers
  • Views 237 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.

MSP430FR6047: MSP430FR6047

RAUL ALARCON TARRAGA
Prodigy 40 points
Part Number: MSP430FR6047


Tool/software:

Hi TI team,
I’m developing an ultrasonic water meter based on the MSP430FR6047 using USSLib v02.40.00.00 and the USS Design Center (UltrasonicWaterFR604x_02_40_00_00 demo).

On my “golden” meter body I created an accurate, multi‑range VFR calibration curve (flow‑rate vs. ΔTOF) that yields excellent accuracy.
When I flash the same firmware + curve (or try the same electronic) to other meter bodies of the same mechanical design, the error rises noticeably. Could mechanical tolerances (transducer spacing, O‑ring compression, mirror position, etc.) be the cause?
After studying the documentation I believe the following two‑step fast calibration per meter body should let me keep the master curve while compensating unit‑to‑unit variation in a few seconds (We cannot afford to spend a lot of time calibrating each meter.):

1 ) Transducer Length Calibration (offset correction)

  • USS Design Center → Calibration → General → Transducer Length Calibration
  • Water at 0 l/h, stable temperature.
  • Press Start, wait ~5 s, Stop → accept the new Transducer Length which updates "Transducer Length" in USS_temperatureLUT.h.

2 ) Meter Constant Calibration (gain / slope correction)

  • Same GUI page → Meter Constant Calibration
  • Run a single, stable flow – ideally ~50 % of Q3.
  • Enter the reference flow, Start (~60 s), Stop → enter the true average flow.
  • GUI computes the new Meter Constant.

3) Store constants

  • Push "Request update" to update the meter.

4 ) Quick verification

  • Check 0 l/h and one flow point (Q1 or Q3).
  • If residual error > ±2 %, tweak Delta TOF Offset (ps) in small steps (±200 ps).

Questions for TI team:

  • Is this "Transducer length" + "Meter constant" procedure the officially recommended way to port one VFR curve to many identical meter bodies?
  • Should deltaTOFOffset always be trimmed as well?
  • Any pitfalls when performing Meter Constant Calibration in every meter after general calibration curve with VFR?
  • For production, is there a better‑practice workflow (batch XML generation, automated CLI, etc.) you would suggest?
  • 0
    TI__Mastermind 46765 points

    Hi,

    So, you are using VFR+temperature calibration method and the temperature data is calculated from absTOF result not from a temperature sensor. 

    1. Well, I recommend to get the temperature data from a temperature sensor if you have an accuracy requirement among temperature. With a temperature sensor, you do not need to calibrate transducer length. And usually Meter Constant and deltaTOFOffset should be enough to cover most meters. 

    2. Could you find relations on the raw flowrate data among such as 10 meters or more? If you can find the relationship on the raw flowrate data, then there should be a way to cover those meters with a general calibration curve. 

    3. I would not have a recommend workflow. It usually depends on the meter manufactures experience.

    Best regards,

    Cash Hao

  • 0 in reply to Cash Hao11
    Prodigy 40 points

    Hi Cash,
    Thanks for the quick feedback. Let me clarify our boundary conditions and double‑check the recommended sequence:


    1. Temperature considerations

    • Lab tests: water temperature drifts only ≈ ±2.5 °C (total 5 °C span) during a workday.

     

    • Field use: expected water temperature range is 5 °C … 45 °C (≈ 40 °C span), but the meter is not required to display temperature, only flow.

     

    • We currently derive temperature from absTOF; no physical sensor is assembled.



    Question A:

    Given the temperature ranges described above, is it necessary to calculate the temperature using absToF (transducer length), calibrate a ‘VFR Flow + Temperature’ curve or is it not necessary at all?

    2. Behaviour across bodies

    The attached plot shows the %‑error curves for twelve meter bodies using the same VFR calibration.



    All curves share an almost identical shape, isn't it?

    Question B:

    Should be enough calibrate Meter Constant plus deltaTOFOffset (zero‑flow bias) to bring every unit inside spec?

    3. Proposed per‑unit adjustment flow

    • Zero‑flow → run deltaTOFOffset calibration.

     

    • One mid‑range flow point (~0.5 × Q₃) → run Meter Constant calibration.

     

    • Verification at Q₁ and Q₃.



    Question C:

    Is that the correct order (offset first, then gain)? Or does the Meter Constant routine internally re‑establish its own offset, making the order irrelevant?

    Question D:

    Apart from these two parameters, do you foresee any other per‑unit value that must be trimmed once the common VFR table is flashed?

    Regads.

  • 0 in reply to RAUL ALARCON TARRAGA
    TI__Mastermind 46765 points

    Hi,

    1. It depends on if you have accuracy requirement at 45degC or 5degC. If you do have requirements under those temperature, then it is recommended to calibrate flow with temperature.

    2. Those curves do have identical shape from my side. 

    3. Meter Constant plus deltaTOFOffset (zero‑flow bias) are used to bring all the meters has the similar curve. You still need to calibrate the curve inside the spec.

    4. The order is important. The offset should goes first. The meter constant will not establish its own offset.

    5. Usually those two parameters should be enough for a residential meter. 

    Best regards,

    Cash Hao

**Attention** This is a public forum