Page 1 of 1

Tech on Pull-up/pull-down resisters

Posted: Tue Jan 08, 2013 3:54 am
by barkerdm
I'd like to get a little bit more clarification about when these resistors are needed.

What I am looking to do is to take my existing aftermarket gauges and connect them to RaceCapture for logging. I understand that pull-up/down resistors are meant to limit current draws to prevent shorts, but are they need if I pull my signal from my gauges?

These are simple Stewart Warner gauges with resistor style sensors. Feed the gauge power, one wire to the sensor, which is grounded via the threads. Basically we have a voltage divider, so shouldn't the internal resistance of the gauge limit the current?

I'd assume any sensor I would add, say a potentiometer, I could power via the 5Vref with a pull-up resistor.

Thanks.

Posted: Tue Jan 08, 2013 4:19 am
by brentp
Hi Barkerdm,

pullup resistors are needed when dealing with passive resistive sensors such as temperature sensors. The pullup resistor makes for a voltage divider, where the pullup resistor is fixed, and the sensor's resistance is variable.

In the diagram below, the vout is what get's fed into the analog input of race capture; Vin connect to the voltage reference - the 5v output of race capture, for instance.

In your case with the gauges, the gauges are supplying their own voltage reference supplied by battery voltage. The downside of this is that this voltage reference may not be very well regulated compared to the 5v coming out of RaceCapture's 5V VRef; any instability will affect the readings.

Could you do a simple test? could you disconnect the wire going from the gauge to the sensor, and measure the voltage between this sensor wire and ground? That will tell us what the voltage reference is for the gauge.

Posted: Tue Jan 08, 2013 3:41 pm
by barkerdm
Brent

I have. Voltage out of the gauge is 10V. Gauge is fed battery voltage, so 12.5-13 volts.

I initial thought was to regulate the input voltage feeding the gauges. I haven't tested it yet, but any reason why the gauges wouldn't work if power them with 5V versus 12V?

Thanks,

Denny

Posted: Thu Jan 10, 2013 1:54 am
by brentp
I think the voltage would be off. The gauges might have internal regulation to 10v, however!

Sensor linearization calculation

Posted: Sat Feb 23, 2013 7:24 pm
by GTIspirit
In case some of you were wondering how to put this information to practical use, let's work through an example:

VW KE-Jetronic ECU with Volkswagen coolant temperature sensor

Vsupply=5V (Is a pretty standard voltage supply value for Bosch ECU's)
Rpullup=5000 Ohms (Is a pretty standard internal pullup resistor for Bosch ECU's)
Temperature sensor linearization curve is pretty standard too, use this chart for example:
http://www.bosch-motorsport.de/en-US/li ... 569739.pdf
Let's pick a few points off the chart to make some example calculations:
30degC - 1707 Ohms
100degC - 187 Ohms

The voltage at the sensor can be calculated as follows (skipping the circuit math and getting straight to the point):
Vsens=Rsens/(Rsens+Rpullup)*Vsupply

Vsens@30degC=1707/(1707+5000)*5=1.27V
Vsens@100degC=187/(187+5000)*5=0.18V

These calculations can be repeated for the entire sensor range according to the sensor specification of resistance versus temperature. Then this sensor linearization curve can be used to transfer the voltage reading to temperature, something like this, degC=[20,30,60,80,100,120],Voltage=[1.67,1.27,0.53,0.30,0.18,0.11].

If you made a plot of these values you would notice that the linearization curve is not linear. Assuming a somewhat limited number of values to program the linearization curve into the Race Capture Pro, the points should be judiciously selected to yield good resolution in the are of interest. Meaning if you are not interested in cold starts and only have five or six values in the transfer curve array, pick say 40, 60, 80, 90, 100, 110degC.