Electronic Otterstat

[Bitsyncmaster]

I'm guessing the accuracy problem is the senders. Although another possibility is oxidation on those nuts where the calibration resistor is mounted. If anyone has there gauge out of the car, can they please remove the resistor and measure it. My guess would be it's in the order of 500 to 1000 ohms.

If anyone has an old OEM sender, I will pay you $20 for it so I can run it through my testing setup.

Here is my test setup. I cover the aluminum plate with water so it should represent what the sensor sees in the car.

[Bitsyncmaster]

I think I thought of the math needed to have my otterstat compensate for car voltage changes. The sensor voltage will be a set percentage of the car voltage at one temperature. At 260 degree it will be 50% of the car voltage.

So I now I have to devise a way to write that code in assembler language using my simple 8 bit micro.

[Bitsyncmaster]

I thought of an easy way to read percentage of the sensor voltage. I will feed my AD (analog to digital) reference voltage with the cars voltage (divided by 3). So I don't need any software to convert the sensor reading (adjusting for car voltage changes).

I also thought of another user warning check. Give the user a "beep" if the car voltage is to great (above 15 volts). I did have a friends car have the alternator go bad and over voltage (above 16 volts).

[Bitsyncmaster]

I've got the bench calibration done for my software. I plan to check the 1/8 gauge reading and the 1/4 calibration values in my car. I won't check the higher values in the car since I don't want to purposely over heat my engine.

I'm leaving for the beach for a week so won't be doing any DeLorean work till I get back.

[Bitsyncmaster]

I got a used gauge sensor from an owner here. It looked exactly like the Hervey sensor I tested before. It even has the same marking of "AC 12V".

So I ran the same test but this time I started at 212 degree F and turned off the burner so the cool down was much slower and would give me more accurate values. The 212 reading was exactly the same and the other readings very close. So I may need to change my testing to drive the same current into the sensor that the car uses. Remember I measured 54 ohms drives the sensor via the cars 12 volts. I'm guessing that self heating will change the values to what I expect. I did not think that heating would change the resistance since the sensor is running in moving fluid. But I will have to retest the sensor with that same drive current. The multimeter uses almost no current to read resistance.

Anyway use this table to test your sensor with a multimeter.

212 deg. F = 168 ohms
200 deg. F = 210 ohms
180 deg. F = 297 ohms
160 deg. F = 428 ohms
140 deg. F = 633 ohms
120 deg. F = 965 ohms
100 deg. F = 1489 ohms

[Bitsyncmaster]

I ran a test driving the sensor through a 54.3 ohm resistor powered to 14.00 volts. My calculated ohm values now appear to be what I expect from my measurements in my car. So it looks like these two sensors I have are the correct calibration. I was not able to finish the test because the door bell rang and someone driving by my house wanted to buy my DeLorean and I missed recording the values as it was cooling down.

213.2 deg. F = 99.9 ohms. Note in my car the 220 mark needed 94 ohms.
200 deg. F = 119.3 ohms.
180 deg. F = 175.4 ohms.
160 deg. F = 288.1 ohms.
140 deg. F = 452.5 ohms.

No I am not selling my DeLorean. The garage door was open and he saw it from the road.

[Bitsyncmaster]

I plugged my "electronic otterstat" into my car. My DIP switch only has 8 selected values right now. Starts at 1/8 scale on the dash temp gauge and each 1/8 increase after that. I wanted to check the accuracy of those selected values. The first one at 1/8 came out perfect (I set my software to beep above the setting). The 1/4 scale came on a little before the engine got to that gauge mark. By the way that is when my thermostat opens since it started beeping and then the beeping stopped which means the engine cooled down when the thermostat opened.

Now I want to check the rest of the calibration but my engine only heats a little past that 1/4 mark. So I plan to plug one of my spare sensors (that I verified are correct calibration) replacing the sensor in the engine. So I can put that sensor into a pan of hot liquid and finish the testing. That will even give me temperature values for the marks on the temp gauge.

So I need to find a liquid that can be heated past 260 deg. F at atmospheric pressure. Regular antifreeze only increases the boiling temp a little bit past water. The spec of 265 deg. F is at 15 PSI.

I want to calibrate those 1/8 marks so I can use those values to add the finer user selection. I think I will need 32 selected trip points. I need the large range to compensate for each cars tolerance of gauge and sensor accuracy.

[DMC5180]

Do you have a small deep fryer? Cooking oil boils at about 300 C.


Sent from my iPhone using Tapatalk

[Bitsyncmaster]

Do you have a small deep fryer? Cooking oil boils at about 300 C.


Sent from my iPhone using Tapatalk

Thanks. Good idea.

[Bitsyncmaster]

I did some more in the car testing. I wanted to calibrate the unit to coincide with the OEM otterstat. To my surprise the OEM otterstat turns the fans on a little below the 3/8 gauge mark and does not turn them off until just a little above the 1/4 mark. I never see this because most of the time my AC is on and the fans are running all the time. Or the car is moving and that provides air flow. So that OEM otterstat has a lot of hysteresis (large gap between turn on and turn off). I don't think I will put that much hysteresis in my solid state otterstat.

I bought a hot plate and some canola cooking oil which says it can heat to 500 deg. F. I'm going to put the gauge sensor into a pan of that oil so I can check calibration values for gauge readings up to 260 deg. (full scale).