SG8 / SH8 Temperature Sensor Specification Errors

[nguzzard]

My application requires the calculation of temperature from the internal temperature sensor of the S08.  We plan on using the SG8 since it is aimed at automotive apps, but we are starting with the SH8 since it is nearly equivalent, and the chips are readily available for prototypes.  The processor supply will be 5V.

Due to the significant similarities between the SG8 and SH8 I will only refer to the SH8 from here on out unless there is a specific difference.

The following excerpt is from section "9.1.4 Temperature Sensor" of the SH8 datasheet Rev.1 11/2007:

"Equation 9-1 provides an approximate transfer function of the on-chip temperature sensor for VDD = 5.0V, Temp = 25°C, using the ADC1 at fADCK = 1.0MHz and configured for long sample.

TempC = 25 - ( (VTEMP - 0.7013) / 0.0017 ) Eqn. 9-1

0.0017 is the uncalibrated voltage versus temperature slope in V/°C. Uncalibrated accuracy of the temperature sensor is approximately ± 12°C, using Equation 9-1."

The following is my understanding of the preceding excerpt so please reply if I am off in any way...

1. The constant 0.0017 (1.7mV/C) is supposed to be the temperature sensors slope with VDD = 5V.
2. The constant 0.7013V is supposed to be the voltage of the temp sensor at 25C with VDD = 5V.

Contrary to these 2 points is the data at the bottom of "Table A-12. ADC Characteristics".a

1. The temp sensor slopes are 3.266 and 3.638mV/C for -40 to 25C and 25 to 125C respectively with VDD = 5V.
2. The temp sensor voltage is supposed to be 1.396V at 25C with VDD = 5V.

Both of these data pairs cannot be correct.  As it turns out, the numbers used in Equation 9-1 of the SH8/SG8 datasheet are incorrect for VDD = 5V, but would be correct for 3V.  The data in Table A-12 is correct (with the exception that the 25C/5V temp senor voltage is shown as 1.396mV instead of 1.396V).

My local FAE has confirmed that the numbers in Table A-12 are correct, and the numbers in Equation 9-1 are incorrect for VDD = 5V; however my FAE also thought that the SH8 was a 3V part when it is actually a 5V part.

So to double check, can anyone confirm that for an S08 part (preferably an SH8 or SG8) that with VDD = 5V, that the 25C temp-sensor voltage is ~1.396V and that the slope(s) are near 3.4mV/C?

[bigmac]

Hello,

The formula given as Eqn 9-1 is similar to that given for many other S08 devices that simply use the voltage drop across a single diode as the sensor output.  However, if Table A-12 is at least partly deemed correct, this would indicate that a different temperature sensor is used in the SH8 device.  I agree that the specified output voltage at 25C would be incorrect, since this would involve negative values at one of the temperature extremes.  In all probability the units should be in volts, rather than mV.  You might easily check this with an actual measurement of the temperature sensor.

If a new temperature sensor is involved, it cannot be taken for granted that the temperature coefficient will be negative.  However, it would appear that the new sensor might consist of two diodes in series, but this would need to be confirmed.  So assuming a negative coefficient, I suspect that the previous formula should be as follows, for the high temperature range. 

TempC = 25 - ((VTEMP - 1.396)/0.003638)

Normally temperature sensors will produce an absolue ouput voltage, and would not be linearly dependent on Vdd, but this needs to be clarified for the SH8 case.  It would only be the ADC reading that would be Vdd dependent.  Hence the need to also measure the reference voltage, as stated in the descriptive information.

This issue has been discussed in recent threads on this forum, that can be found using the search engine at the bottom of this page.

Regards,

Mac