Bedfordman

On the S4 :: How close do the readings from each pins to earth on Temperature Sensor II need to be?

I have 4.90 k-ohm and 5.05 k-ohm at 5 degrees C ( about 40 F)
and
260 ohm and 270 ohm at about 80 degrees C ( about 170 F)

Peter

FredR

Peter,

The difference in your readings is irrelevant I would suggest. The WSM specifies 4.4 to 6.8 kohms at 0 degrees centigrade and more importantly, 250 to 290 at 80C so you are spot on I would say.

worf928

worf928

FredR

Dave,

I never managed to fathom out exactly what the EZ channel was doing with the temperature measurement but all I noticed in ST2 was that the values displayed on the LH and EZ pages agreed to within 1 degree centigrade. I never tried to calibrate the resistance to actual temperature.after I acquired ST2 - did not see any point given ST2 was telling me what the temperature was and showed reasonable alignment with that on the dash and more importantly, what I expected to see.. As I remember the difference in LH mapping between 80C and 85C is something like 1 unit so such correction is more or less irrelevant. Logic suggested that if one channel says temperature is 85C and the other 84C in itself is interesting but probably irrelevant given the values for 80C are acceptable providing they fall within the range 250 to 290ohms. Thus if 270 ohms represents 80C then resistance values can vary by about plus or minus 6% and still be considered acceptable performance wise- thus it would be interesting to know how that value range impacts actual temperature- I suspect it amounts to plus or minus two or three degrees. .

worf928

Fred, the precision of the TEMP-II is effectively infinite since it is analog. Based upon my memory of observations the precision of the digital conversion of the NTC resistance is ~2°C. Thus there is significant loss of precision due to A/D quantization. (Note the use of 'precision' and not 'accuracy.') Furthermore, I don't know if the A/D process on the two ECUs is identical. I think not though. While at one point I was very familiar with what the EZK does with NTC-II conversions, that knowledge has, due to disuse, leaked out of my brain. Furthermore, since one cannot run the ST software without the hardware connected to a 928, it isn't convenient for me to quickly construct a new association for retrieval of the disused memory.

What I do recall was the operational conclusion I drew from my heavy use of the ST over a period of two years: Minor resistance variations across the two NTC circuits was to be expected and not a concern. However, if the variation was higher than 5-ish percent, and there were un-diagnosed engine operating issues, that replacing the TEMP-II was a step in the right direction if for no other reason than variable control for fault analysis.

FredR

Dave,

I would expect the analogue to digital conversion to be the same in both the LH and EZ units given they are both made by Bosch [unless they were farmed out]. I suspect it is more likely that there is some small difference in the analogue [resistance] values. Such a device probably has an accuracy of no better than 2% but that should be more than adequate thus the range of values Porsche specify in the WSM.

I struggle to remember the cells in the LH warm up map but they go from zero C to around 85C in something like 5 steps if my memory serves me correctly. I seem to remember that the cell width reduces as temperature increases and the last adjustment cell covers something like a 5 degree range from 80C to 85C and adds 1 unit of correction [virtually nothing] so the warm up correction is the same for any value in that array and anything above 85C gets no correction thus why accuracy is not that critical. Apologies if my memory of the numbers is a little out but the principal holds.

Although they are ancient systems by today's standards, I was actually quite surprised at just what these systems can do, clearly state of the art when new. Digitisation hit the process industries big time in the late 70's with systems like Honeywell's TDC2000 and needless to say the need for digitisation goes with such systems. That NASA were using digitisation in the Apollo missions in the 60's should be no surprise but at enormous unit technical cost. Needless to say a modern smart phone dwarfs our control units in terms of computing power. They probably give away things in a packet of corn flakes that are more powerful than what our computers run.

Bedfordman

Wow, all the detail you get from our Rennlist Experts. Thanks.
I mapped the Resistance Ranges against Temperatures given on a graph -- yes, on a sheet of good old graph paper.
I measured the resistances at garage temp and operating temp. My readings came out on the boundary lines...
So, probably OK. Given Rennlist Discussion: I have ordered new Temperature Sensor II.
I believe some owners consider this sensor a service item, it's not expensive in 928 terms.
(I wonder if the new sensor will be as 'accurate' as my old one).

Peter

FredR

Peter,

Temp 2 is one of the many items on the car that degrade and ultimately fail with use and age. Pre-emptive replacement if the original item is still fitted is just common sense really. Just keep the one you remove as back up - new kit has been known to fail.