Any body use X9C102?

[Stephen Moss]

when engine temp 120C I need to send 40 ohm (gauge pointer right side)

m just check it (33mA)

If 33mA is for full scale defection of the meter, have you tried measuring the resistance of the meter? As it may be possible to drive it from a voltage which is easier, either smoothing out a PWM or a DAC.
Alternatively you could try placing a 40R resistor in series with the gauge to limit maximum current and power it from a couple of AA batteries, or even a single one if that is easier (assuming the other end of the gauge is at Gnd and not 12V) to see what effect that has on the gauge and from there work out the voltage/ current requirements to run it off the 5V out your would likely get from a PIC or DAC rather than from 12V.

[Maxi]

this morning Im measuring again

when input pin empty, input-gnd voltage 11,10 volt
connect 102 ohm to gnd > input voltage is 2,64v
connect 989 ohm to gnd > input voltage is 8,35v

so, input pullup resistor is ~326 ohm

gauge minimum side input resistor 382 ohm
gauge max side input resistor 40 ohm

when I connect input 390 ohm resistor(to gnd), its take 15.4mA
there is extremly high I think

when I connect input to gnd, its take 33mA (as I measure yesterday)

[Stephen Moss]

I am struggling to make sense of those readings.
When you say "pullup resistor" do you mean the resistance of the gauge or is the pull-up resistor a separate component and how do you know that is is 326 Ohms?

To clarify, when you say "when input pin empty, input-gnd voltage 11,10 volt" and "its installed engine water hose and send resistive data to cluster gauge it has two pin, other pin connect engine block (like gnd)" Do you mean that one end of the gauge is connected to the battery and the other, "input" end  as you are referring to is connected through the NTC thermistor to GND?

Looking back through the thread to try an make sense of the circuit connection I only just realised that, if correct...

minimum resistor 380 ohm (60C)
mid point 57 ohm (90C)
max 40 ohm (120C)

appears to be a logarithmic progression with a change of 323 for the first 30 degrees, but only a change of 17 for the second 30 degrees.
That may in part explain why the values you proved were not making sense to me. I hope you are good at mathematics to calculate the same logarithmic progression for your output.

Also, I think you are focusing too much on the fact that the sensor is resistive

The problem is "how can I send or generate this resistive data to cluster?"

rather than what function the thermistor is performing, which is likely to be either changing the current through the gauge or the voltage across it depending on how the two are connected which can be achieved through means other than a variable resistor.
Which brings me back to my original reply (subsequently echoed by others) that you should be looking to control either the voltage across or the current through the gauge rather and calculating those requirements rather than a resistance.

[Maxi]

ok, cluster have a circuit module
sensor first come this module (and other all signals)
than goto temp gauge , fuel gauge etc
I dont know how work this module, its in the cluster.

so ,I know ntc sensors how can be measure (also you know ofcourse)
if is not very very special circuit
only a pull up resistor location input.

trying add simple picture here


VOUT goto this hidden circuit
if I measure VOUT, I can measure +Vs ? (no any NTC sensor connected)
its 11.10volt

after this,
if I connect 100 ohm or 1k ohm to gnd
I can calculate RS resistor?
is it a simple resistor divider?

why gauges resistors not linear?
I think this is ntc resistor scale
ntc sensors not linear and resis value very different
I trying add a picture again, if I understand wrong you


tell me pls my mistake or measuring
maybe I dont understand well
not very good my english sorry

[Stephen Moss]

Would it not be simpler to just recreate that circuit using an NCT thermistor with a similar logarithmic gradient to that of the original, feed the output voltage into an op-amp connected as a voltage follower all running off 12V rather than using a PIC?
Then you could probably leave the original Rs in when connecting the amplifier output to the instrument cluster, although I would remove it.

If you really want to use a PIC then you could...
1) Recreate the provided circuit but run it off 5V
2) Read the input voltage with an ADC, although given the non liner change I would be inclined to use a 10 or 12 bit DAC to more accurately track the voltage changes.
3) Use the ADC result value to generate a suitable output voltage using a smoothed PWM or DAC.
However, unless using a 12V DAC your output voltage will be the same as the input as you are limited to the 5V supply of the PIC/DAC, requiring you to amplify the output to fit the original 12V scale and use that to drive the instrument cluster. Which is why an amplifier would be the better option, as you will likely have to use one anyway. 

[Maxi]

final update

Im search and found water temp part number
buy one this sensor
install it and testing with hot water (89C)

gauge pointer stay a bit right side (I want middle point)
after this, install a 33 ohm resistor serial to sensor.

now, no any problem, well done.
but I really curious cluester`s circuit module.

thanks your supports and @Stephen Moss

[Yves]

You can improve the linearity to an NTC by judicious choice of resistors. RP must be in the middle of the range of interest. You may loose a bit of resolution but if you use an 16bit adc converter it will compensate for the loss.