How is CLT simulated?

spandit · Post by **spandit** » Thu Dec 06, 2012 11:59 am

Since I am converting an air cooled engine I don't have a coolant sensor so am currently using a single CHT thermistor.

However, my plan is to fit 4 sensors (1 per cylinder) and run them to an Arduino. This will run a CHT gauge on the dash and average/normalise the 4 readings. I'd like to then send this to the MS but not sure how to do it. It must be possible as the JimStim etc. manage so what kind of signal should I send up the CLT/sensor GND wires?

Post by **jsmcortina** » Thu Dec 06, 2012 12:46 pm

The stim uses a variable resistor, just like the thermistor it is simulating.

James

spandit · Post by **spandit** » Thu Dec 06, 2012 3:17 pm

Oh, is that manually turned (varied) then or is it something clever?

Could I disable the CLT input and use a PWM input instead?

Post by **jsmcortina** » Fri Dec 07, 2012 1:48 am

The schematic for the basic stim and for the coolant input circuit are both available online - please look at them.

James

spandit · Post by **spandit** » Fri Dec 07, 2012 2:59 am

Thanks, James. I had a look and it showed that the CLT is used as a voltage divider but I'm new to electronics & am still confused. I'll try to work through the JimStim schematic and see if I can understand it

spandit · Post by **spandit** » Fri Dec 07, 2012 3:03 am

Ah, OK, I see that it just uses manual potentiometers, might not be possible then

SymTech Laboratories · Fri Dec 07, 2012 3:34 am

spandit wrote:Could I disable the CLT input and use a PWM input instead?

You could do that, but you would have to send the PWM signal through a low-pass filter and remove the bias resistor inside your MegaSquirt ECU (R7).

spandit · Post by **spandit** » Fri Dec 07, 2012 8:52 am

I've got a MicroSquirt so can't really do that. I think what I meant to say was, can I set a PWM signal, generated from the Arduino, up as another sensor and use this for the warm up enrichment etc. instead of the CLT pin (to which I could just wire a fixed resistor)?

racingmini_mtl · Post by **racingmini_mtl** » Fri Dec 07, 2012 10:16 am

spandit wrote:I've got a MicroSquirt so can't really do that. I think what I meant to say was, can I set a PWM signal, generated from the Arduino, up as another sensor and use this for the warm up enrichment etc. instead of the CLT pin (to which I could just wire a fixed resistor)?

No.

elaw · Post by **elaw** » Fri Dec 07, 2012 10:48 am

But couldn't he just put a voltage into the MS pin that normally would go to the sensor? If the voltage source were fairly low impedance (buffered with an opamp), the bias resistor would not have any effect. After all, the A/D is really just reading the voltage.

The toughest problem might be getting a "calibration curve" loaded, as it would have to be linear. So none of the standard thermistor curves would work, and neither would the curve generator in TS.

SymTech Laboratories · Fri Dec 07, 2012 12:25 pm

elaw wrote:But couldn't he just put a voltage into the MS pin that normally would go to the sensor? If the voltage source were fairly low impedance (buffered with an opamp), the bias resistor would not have any effect. After all, the A/D is really just reading the voltage.

That could work too. If the bias resistor had a much lower resistance (<100 ohms), it could complicate things. With an op-amp voltage follower, though, the standard resistor will have a negligible effect.

Post by **jsmcortina** » Fri Dec 07, 2012 1:01 pm

Or how about a simple and reliable answer. Add a 5th temp sensor and use that for Megasquirt. Use the other four for your dash gauge project.

James

spandit · Post by **spandit** » Sat Dec 08, 2012 2:48 am

jsmcortina wrote:Or how about a simple and reliable answer. Add a 5th temp sensor and use that for Megasquirt. Use the other four for your dash gauge project.

James

Simple I like

, but I was hoping to use an average of the four sensors as my CLT reading. I've had a play with some numbers and wiring the 4 thermistors up in two parallel pairs gives fairly good readings over the normal range expected.

Some good info above, though. Need to work out what the H most of it means. I know that MS uses a potential divider to measure the varying voltage obtained by having the thermistor in the circuit so thinking a filtered PWM output from the Arduino might have the same effect. I know the Arduino measures resistance in the same way (and has the same reference voltage).

Producing a curve isn't necessary as it will just be using data from the 4 sensors which all follow the same curve anyway.

Plenty of playing to do which keeps me off the streets!

spandit · Post by **spandit** » Sat Dec 08, 2012 2:51 am

elaw wrote:But couldn't he just put a voltage into the MS pin that normally would go to the sensor? If the voltage source were fairly low impedance (buffered with an opamp), the bias resistor would not have any effect. After all, the A/D is really just reading the voltage.

The toughest problem might be getting a "calibration curve" loaded, as it would have to be linear. So none of the standard thermistor curves would work, and neither would the curve generator in TS.

This is pretty much what I'm planning, although as per my note above, the curve isn't an issue (and even if it were, I could work it out algebraically)

Not sure what you mean about the op-amp stuff (new to electronics) but will read up on it

spandit · Post by **spandit** » Sat Dec 08, 2012 2:52 am

racingmini_mtl wrote: No.

About the only answer I've truly understood so far!

spandit · Post by **spandit** » Sat Dec 08, 2012 3:24 am

OK, using the voltage divider equation of Vout=Vin.(Z2/(Z1+Z2)) and taking the in-built bias resistor as Z1 (2490Ohms), I've constructed the following table, the Z2 values being drawn from the thermistor data table:

It would be fairly trivial to output a matching voltage from the Arduino (I think) from a PWM pin with a suitable filter. I think I'll just try it and see what happens

SymTech Laboratories · Sat Dec 08, 2012 12:21 pm

spandit wrote:It would be fairly trivial to output a matching voltage from the Arduino (I think) from a PWM pin with a suitable filter. I think I'll just try it and see what happens

Give it a shot, you don't have anything to lose. The op-amp buffer is a must, though. Most electronics stores carry them. Here's the datasheet for the LM358, a relatively common op-amp, and a link to some information on using an op-amp as a voltage follower (buffer).

http://www.ti.com/lit/ds/symlink/lm158-n.pdf
http://en.wikipedia.org/wiki/Operationa ... plifier.29

spandit · Post by **spandit** » Sat Dec 08, 2012 6:20 pm

Thanks. Have ordered a load of components and think there is an op-amp in there. Will try it without first and see what happens...

seishuku · Post by **seishuku** » Sun Dec 09, 2012 10:25 am

Could use a digital potentiometer
http://datasheets.maximintegrated.com/en/ds/DS1267.pdf

SymTech Laboratories · Sun Dec 09, 2012 11:31 am

seishuku wrote:Could use a digital potentiometer
http://datasheets.maximintegrated.com/en/ds/DS1267.pdf

That might work if he calibrates the ECU with a linear and relatively narrow thermistor table, but he would have to use a 10-bit pot. In general, digital potentiometers don't have enough resolution to accurately emulate a thermistor.

Megasquirt Support Forum (MSEXTRA)