OPT3101: Temperature deviation

Stefan, 

While I'm discussing with our design team on the issue, please provide some more details:

Did you use on-chip temperature sensor or external one? What temperature range did you calibrate?  At what distance and target reflectivity conditions did you see 15%+ distance deviation?

Thank you.

Wenjuan

Hello Wenjuan,

We are using the on-chip temperature sensor. The calibration was done for a temperature range of -40 to +60 °C.

The power-on distance deviation of 15% was observed on a White 90% target at 1000 mm distance.

Thank you.

Stefan

Hi Stefan,

How does the temperature calibration result look like? What temperature coefficient did you apply to the system? What is the signal amplitude level when you see 15% of deviation?

Thank you. 

Wenjuan

Hi Wenjuan,

following the results of the power-on drift measurement.

Distance deviation vs. Tmain shows a linear correlation. We thougth about an additional compensation in µC that way. But the linear gradient varies with target distance and for different devices.

Our temperature calibration:

We use Super-HDR-Mode with the following LED currents. Our temperature coefficients are like this:

LED currents	Xtalk temp. coefficients	Phase temp. coefficients

In most cases LED current TX1 high is used. Following some calibration results for TX1 high:

Xtalk	Xtalk	Phase	Phase

Thank you.

Stefan

Stefan,

If your distance deviation vs. Tmain chart is correlated with distance vs. time chart, the system will be stabilized at 55C, correct? 

You mentioned distance deviation vs. Tmain is linear but with different gradient. How many parts did you see the distance deviation issue?

You mentioned the calibration was done for a temperature range of -40 to +60 °C. However, the Xtalk calibration charts show 19~27°C. Did you apply the temperature coefficient from these charts (19~27°C)?  The Q Xtalk chart looks noisy, do you see same results to other parts?

Thank you.

Wenjuan

Hi Wenjuan,

The power-on drift measurement was done at room temperature. Here the IC internal temp. Tmain ends up at about 55°C. That's correct.

We measured 10 parts here.

Calibration was done in a thermal chamber, with temperature variation -40 to +60 °C. For this range, the IC internal temperature Tmain changes between 1900 and 2700 codes. This corresponds to -19 to 82 °C.

BR,
Stefan

Hi Wenjuan,

Some more general questions on the calibration procedure:

1) According to SLAU791.pdf , to determine temperature coefficients, a best line fit with I / Q / phase data on the Y-axis and temperature Tmain on the X-axis has to be done.

Y-Axis: For crosstalk, OPT3101 SDK prints out I & Q and ScaledI & Scaled Q. Which to use here for the best line fit. I/Q or ScaledI/ScaledQ?
X-Axis: Here, Tmain in codes or Tmain in °C has to be used?
(According to SBAU310.pdf and SDK User Guide, Tmain in codes should be used. Is this correct?)

2) Via the best line fit, temperature gradients as floating point numbers are determined. E.g.:

According to SBAU310.pdf, these gradients need to be scaled down to a 8-bit signed number.

In OPT3101 SDK, temperature coefficients register values are calculated as followed. Here a cast to an 8-bit unsigned number (uint8_t) is done.

--> What is the exact data format the OPT3101 IC register expects?

Could you post an example of a calculation of the temperature coefficients register values?

3) According to SBAU310.pdf, at crosstalk temperature correction, the following calculation needs to be done:

According to the SDK, the slope additionally is multiplied with the magnitude:

What is the reason for this?

Thank you.

Stefan

Stefan,

I'm looking into the questions and will get back to you tomorrow. 

Thank you. 

Wenjuan