This site has proved to be an incredible resource for an ME trying to do EE things. I have been able to search and find most of my answers I have needed thus far but now I am a bit stuck and looking for some help if possible. I have found a few threads similar to what I am looking for but I don't think they are exactly the same based on the results I have been getting.

The engine is a 4BT Cummins and I built the engine wiring harness I am using around a 5v sensor/ground bus with my assumption that the datalogger I am using, OneGauge (Arduino based) would require a 5v square wave signal from the crank position sensor but I was wrong. It needs a 12v square wave signal (minimum of 7v or so to drive the isolator from talking with them). I do not have spare pins in my firewall plug, nor do I want to tear apart my harness to send 12v and the separate ground to the sensor.

Initially I was using the factory dodge sensor (Cummins 392442) but I couldn't get it to function properly with the hub. This could be because I had it wired wrong or a fried something in the sensor. It seemed to do what I wanted to when I bench tested it, but then when I installed it, it didn't work like it was on the bench. The balancer I am using is shown below and you can see that it is essentially the opposite of most toothed wheels in terms of geometry. 5000rpm is max so that should be 167hz at max.


Using a resistor to pullup the signal to 5v seemed to work, so I thought maybe I could pull output signal up to 12v and wired it like below. It seemed to work fine on the bench for me but didn't work when installed.


I gave up with the factory sensor and picked up a GS100701 hall effect gear tooth sensor datasheet as that is one that OneGauge has used with good success in the past. My initial thought for converting the 5v to a 12v was just to use a transistor like a switch and then have the output of the sensor connected to the base acting as the trigger. I did some research for using a transistor as a switch and found this.


What I ended up with looked like this and again seemed to work on the bench. The pullup resistor to get the sensor to output 5v was external from this little module this time. Rb was a 1kohm, Rc was 10k IIRC. In the picture below, black is ground, green is the signal coming from the sensor (Vin), red is 12v+ (Vcc) and white is the "signal" going out to the data logger (Vo). I don't have the exact part number of the transistor I used offhand and I forgot to grab the little module today but I can check tonight if having the exact parts I used would be beneficial.


I assumed I had something messed up with the design with a transistor being current controlled so I switched to a Mosfet since they are voltage controlled, and put this together:

This seems to work, but it gets considerably hotter than I was expecting. Thinking more about it, my current limiting resistor might be a 400ohm one (I just used components I had available at work) which would be about a 1/2watt of power and I am pretty sure that is a 1/4w resistor I used. I am not 100% sure if the data logger is reading the signal properly because with this setup it essentially gets 2 ground pulses per revolution as it outputs 12v when near metal and 0v (ish) volts when at the gap. I bought a laser tachometer to check what the actual idle speed is so I can compare it to what my gauge is reading as well.

Am I just way overcomplicating this or is there a simpler way to use the square wave output of my hall effect sensor to send a 12v square wave to my datalogger? Any and all help is greatly appreciated because I know there are a lot of people on here that are WAY smarter than I am. Happy to give more information and whatnot too if needbe.

 

Welcome to AAC!

nor do I want to tear apart my harness to send 12v and the separate ground to the sensor.

The linked sensor will require 3 wires going to it, unless the ground connection is made to the vehicle bodywork near the sensor (but that might introduce noise due to currents in the bodywork).
If you already have a ground wire and +5V wire going to the sensor you don't need an extra wire going to it for +12V.

 

I gave up with the factory sensor and picked up a GS100701 hall effect gear tooth sensor

There is no need for an additional transistor as the sensor has a built in open collector design.
Just run the sensor off 12 volts like below. Output directly to the datalogger

 

Welcome to AAC!

The linked sensor will require 3 wires going to it, unless the ground connection is made to the vehicle bodywork near the sensor (but that might introduce noise due to currents in the bodywork).
If you already have a ground wire and +5V wire going to the sensor you don't need an extra wire going to it for +12V.

The sensor does have 3 wires going to it, apologies if I wasn't clear. It currently has +5v (from my sensor power bus), ground (from my sensor ground bus), and then the output signal wire. All the sensor grounds tie to a DT bus and the sensor +5v to another DT bus in the engine bay before passing through the firewall connector. All 3 wires come back to the gauge hub since the OneGauge hub is the one providing the +5v and ground to the sensor busses. I wanted to keep dedicated paths for the sensors in order to minimize noise. I don't want to change anything after the firewall bulkhead connector so that means the sensor is stuck receiving +5v and ground from the common sensor busses. The only option I have is to modify the signal coming out of the sensor.

I put the transistor/mosfet modules between the signal wire and the OneGauge hub so in the mosfet diagram, the only wire coming back from the sensor is the one labeled "CPS". The other connections all come from the hub. GND ties back to the common hub ground, +5v is used to pull the output signal "CPS" up to 5v, "To RPM1" is the new signal wire going to the hub, and +12v is the 12v source I was switching with the mosfet that comes from the 12v terminal on the hub. I had the transistor setup wired pretty similar to the mosfet setup.

I can make a better diagram of what I have if it will make more sense and I can post a picture of the hub terminals as well.

In the mosfet circuit you can use a much larger load resistor say 10k.

Ok I wasn't completely sure on the sizing as I wanted to make sure there was enough current. This will probably change either way if I wanted to invert the signal to send a +12v pulse instead of the ground pulse it is currently sending, correct?

 

There is no need for an additional transistor as the sensor has a built in open collector design.
Just run the sensor off 12 volts like below. Output directly to the datalogger
View attachment 355709

I would do that if I could run 12v to the sensor. See my previous reply to Alec describing the wiring layout.

 

Another suggestion using an inverter circuit to maintain the polarity of the output signal.

 

Since the output of the sensor is open-collector I don't see why you can't simply use a pull-up resistor to +12V, even though the sensor supply is +5V?

 

Another suggestion using an inverter circuit to maintain the polarity of the output signal.
View attachment 355711
This seems very simple and straightforward so I will give that a shot. Thank you very much.

Since the output of the sensor is open-collector I don't see why you can't simply use a pull-up resistor to +12V, even though the sensor supply is +5V?

I thought the same thing and tried wiring it that way and it would not work. I wired it like the second picture in my original post with a 10k resistor coming from the 12v terminal to the hub connecting to the output wire of the sensor. From my understanding, the 5v supply of the sensor doesn't have enough to pull it back down after pulling up to 12v.

 

Another suggestion using an inverter circuit to maintain the polarity of the output signal.
View attachment 355711

Do I need to use the 2N3904 transistors or can I use an equivalent? I need to look around through the rest of the stashes here but so far the closest I have found is the 2N3906 which is the PNP version.

 

Equivalent is fine.
The NPN transistor you were using before should be fine as well.

 

In the picture below, black is ground, green is the signal coming from the sensor (Vin), red is 12v+ (Vcc) and white is the "signal" going out to the data logger (Vo).

Are you sure you don't have the 1k and 10k swapped? Looks to me like you have the 1k (brown/black/red) going to +12V. If so, the sensor probably can't sink 12mA.

 

Are you sure you don't have the 1k and 10k swapped?

I see Brn-Blk-Org, 10K.
The other resistor is definitely 1K.

 

So I built the transistor circuit sghioto drew up and bench tested it. Found some 2n2222a transistors. Had an issue at first but then realized my mistake after building the circuit on falstad (love that site) and realized I needed a limit resistor since I was not using the actual sensor itself. If I put 5v directly to the base of the bottom transistor I image it is shorting itself out since even 1v applied shows 3200amps


I threw a 1k in there and everything worked fine. It then got me thinking about the circuit and saw that I only needed about .6-.7v to operate the lower transistor.


Would I be able to just use the direct output from the sensor to drive the lower transistor instead of having to use a pullup resistor? I assume I would have to make some adjustments to the resistors since the sensor says it has a max output voltage of 400mV and a max output current of 25mA.

I am happy that it works using a pullup resistor so if I can't directly drive it, that is completely fine. I was more curious than anything to see if I could simplify it at all. I might try and see if I can use the factory dodge sensor in this manner as well.

I really appreciate the quick responses and whatnot from everyone.

 

Definitely need the 1K pull up.
That is the bias resistor for the base.
Install the 1K pull up at the sensor output where it is located and then feed the output line to the base of Q1.
As shown in post #6.
An additional 1K resistor can be installed between the sensor output and the base of Q1 for safety purposes. Install this resistor on the same board as the transistor circuit.

 

Sorry for the lack of updates on here but have been dealing with some other electrical gremlins on the project. I have the sample board made from the schematic that sghioto provided and it seemed to function correctly on the bench. I installed it in the car but have not had a chance to test it and verify the output.

I will trim the board and whatnot once I confirm everything works like I expect it to.

On another note, turns out the current limiting resistor that I had on the mosfet setup that I thought was an 800ohm was actually 360ohms, which would explain why it was so hot. Oops.

I have to travel for work next week so I probably won't have an update here for a bit unless I have time tonight. I greatly appreciate all the help from everyone.