Several years ago I found out my '78 had something commonly referred to as a "carb fan" and associated thermistor and controller. I didn't notice it was there because it had never worked. A PO had replaced the fan controller's inline 5A fuse with a 15A fuse. That's likely why the controller board was scorched, diodes fried, and inoperative. I replaced a couple of burnt diodes but gave up attempting to fix the controller. I replaced the controller with a salvaged one and the system worked (and continues to work) fine. However, I thought it'd be a fun challenge to build a modern drop-in replacement. This summer I finally devoted some time to designing a circuit, programming a controller, breadboarding, testing, and prototyping.
The default firmware profile is identical to the '78 Federal OEM controller. When the ignition is switched off and the thermistor indicates over 100C, the fan relay closes and the fan runs until the temp is below 85C or 30 minutes elapses, whichever occurs first. Although I have a CA-spec engine, I do not use the CA-profile (fan on if over 120C and off below 105C) because I don't have the "thermal reactor" exhaust manifold installed.
Although it only indicates about 20mA at idle (fan off), I designed it to completely cut power to itself after the fan cycle is completed, so there is no power drain when the ignition and fan are off. It powers on when the ignition is switched on and powers off when the ignition is off and the fan profile is complete.
First pic is the circuit on a breadboard. I was operating the circuit for extended periods (24 hours) at 15V and measuring max current with the relay closed (fan on).
Here's the PCB prototype. I laid out the components functionally so you can see what's what, so to speak. Ignition determines if the ignition is on or not, Power Supply provides a stable 5V and shutoff logic for the board, Relay is just that, and MCU is the micro controller unit and comm header. The shop did a nice job rounding the corners and notching it just like the OEM board. That big hole in the center is for the screw that holds the connector to the board.
Here's the board undergoing some further testing using a trimpot rather than the thermistor so I could test it under varying conditions without getting the engine hot. The board is connected to the laptop via a USB-UART interface to record the data. For the mcu temp curve I recorded the thermistor resistance at various temps and computed a profile with Steinhart-Hart coefficients.
And finally, on the left is my old burnt board (minus the connector) and on the right my updated controller (with the cover removed). The relay may look petite but it's a 12V automotive relay capable of 30A for over an hour, well exceeding a max 5A/30min requirement. I added a protection diode and snubber for the relay to protect it from any inductive transients. The 5V voltage regulator was also designed for automotive use and because the current requirement is so small, I was able to go with a smaller TO-92 package. The two-pin header is for data logging.
Controller works great. I want to add a Bluetooth module to read the temp data wirelessly and replace the relay with a MOSFET circuit to run the fan at different speeds using PWM. Maybe a future version. It's unfortunate that it uses a connector that's no longer produced. It could be easily reprogrammed to use a different thermistor profile, a different brand thermistor, or even a single-wire temp probe.
There's lots more great info on fan controllers in this thread.
The default firmware profile is identical to the '78 Federal OEM controller. When the ignition is switched off and the thermistor indicates over 100C, the fan relay closes and the fan runs until the temp is below 85C or 30 minutes elapses, whichever occurs first. Although I have a CA-spec engine, I do not use the CA-profile (fan on if over 120C and off below 105C) because I don't have the "thermal reactor" exhaust manifold installed.
Although it only indicates about 20mA at idle (fan off), I designed it to completely cut power to itself after the fan cycle is completed, so there is no power drain when the ignition and fan are off. It powers on when the ignition is switched on and powers off when the ignition is off and the fan profile is complete.
First pic is the circuit on a breadboard. I was operating the circuit for extended periods (24 hours) at 15V and measuring max current with the relay closed (fan on).
Here's the PCB prototype. I laid out the components functionally so you can see what's what, so to speak. Ignition determines if the ignition is on or not, Power Supply provides a stable 5V and shutoff logic for the board, Relay is just that, and MCU is the micro controller unit and comm header. The shop did a nice job rounding the corners and notching it just like the OEM board. That big hole in the center is for the screw that holds the connector to the board.
Here's the board undergoing some further testing using a trimpot rather than the thermistor so I could test it under varying conditions without getting the engine hot. The board is connected to the laptop via a USB-UART interface to record the data. For the mcu temp curve I recorded the thermistor resistance at various temps and computed a profile with Steinhart-Hart coefficients.
And finally, on the left is my old burnt board (minus the connector) and on the right my updated controller (with the cover removed). The relay may look petite but it's a 12V automotive relay capable of 30A for over an hour, well exceeding a max 5A/30min requirement. I added a protection diode and snubber for the relay to protect it from any inductive transients. The 5V voltage regulator was also designed for automotive use and because the current requirement is so small, I was able to go with a smaller TO-92 package. The two-pin header is for data logging.
Controller works great. I want to add a Bluetooth module to read the temp data wirelessly and replace the relay with a MOSFET circuit to run the fan at different speeds using PWM. Maybe a future version. It's unfortunate that it uses a connector that's no longer produced. It could be easily reprogrammed to use a different thermistor profile, a different brand thermistor, or even a single-wire temp probe.
There's lots more great info on fan controllers in this thread.