Return of the Hobo

In a previous installment, you read about my glorious triumph installing a flex fuel sensor. I had taken a flex fuel sensor from an old Ford Taurus in the junk yard and fitted it to Grace, and everything was awesome. Turns out it was all a lie. 

Granted, I didn't know it was a lie at that point. I installed the sensor, built a circuit on the ECU board and it was reading the ethanol content of the fuel, or so I thought. I soon found that while out driving, the engine would suddenly run really rich, almost to the point of dying. And then with no indication, it would go back to normal. I took a couple data logs, and found that the ECU was showing a fault on the ethanol circuit and would go into a sort of limp mode, and when ethanol percentage read 1% without any fluctuation, that meant the sensor was not reading correctly.

Lies. ITS ALL LIES!

Data logs, AKA my lifeblood. I spend most of my days at work looking at data logs not unlike this.

 I felt fairly confident that the old junkyard sensor was bad, so I set out to find a new one. After wasting half a Saturday to find one in a junkyard (a new replacement sensor costs a minimum of $300) I gave in and bought an sensor I found on eBay. I had high hopes for this new sensor, but was grateful that it came with a 30 day warranty.

New old junk: another 1st generation metal body ford flex fuel sensor

New old junk and old old junk side by side.

I installed the new old sensor and viola!, same result. By now I was suspicious that maybe I had done something wrong. I checked my ECU several times to make sure that my circuit was right, I checked my wiring, I asked on every forum that I thought might have some knowledge on these sensors. Of all the advice I got, the best piece was this: make a sound card oscilloscope. I needed to look at the signal coming from the sensor to determine if it was toast or if I had screwed something up on the ECU side. 

A sound card oscilloscope uses a bit of free software on your PC to read electrical signals coming through your sound input jack. I had to build a very simple circuit to reduce the voltage of the signal in order to protect the sound card itself, though that was a pretty quick job that cost essentially nothing. Normally a good o-scope will cost hundreds of dollars, but if you just need to check a low voltage signal occasionally, a sound card o-scope can be the perfect tool.

A couple wires, a couple diodes, a resistor and a potentiometer. Behold: the sound card oscilloscope.

I probed the circuit right at the ECU so I could make sure that I was seeing exactly what the ECU was seeing. It was a very clean signal, coming in at 39.9Hz. I had found an old Ford spec sheet for this specific sensor.  Its nominal range is 50 Hz at 0% ethanol and 115 Hz at 85% ethanol (or 125 Hz at 100%). The chart also showed a tolerance band of +/- 10 Hz, which meant that my sensor could theoretically read as low as 40 Hz and still be within range. This seemed to indicate that the sensor was dead.  I still wanted to know for sure, so I ran a small amount of denatured alcohol through the sensor. There was no response in frequency of the signal, so at that point I gave in an ordered a new GM style flex fuel sensor.

1st Generation Ford (Siemens) flex fuel sensor range

39.9 Hz signal = dead sensor

The whole reason I had avoided the GM sensor from the beginning was because of its quick release style fittings. I wanted to have -AN fittings throughout the entire system. It was kind of a silly requirement, but sometimes I get it in my head that something needs to be a certain way and then I'm unreasonably stubborn about making it that way. My wife wouldn't know anything about that... (unrelated side note: planning a house remodel is strangely like planning a wedding. You are required to have an opinion on many things, and once you finally come around to having that opinion, it's wrong ¯\_(ツ)_/¯ ) Anyway, I ended up finding a fitting that made me comfortable with using quick release style fittings in my fuel system. Russell Performance makes a fitting that is a very solid piece for adapting -AN to quick release, and it's actually pretty reasonably priced. I did what I should have from the start and picked up two adapters, a GM (made by Continental) flex fuel sensor, and a new wiring connector.

This two piece adapter fitting threads together to prevent accidental disconnection.

Ford flex fuel sensor connector on the left, GM connector on the right

After way too much fiddling around with sensor location and hose routing, I got the mounting sorted out. When I fired up the ECU and enabled the flex fuel circuit, it instantly started reading right around 14% ethanol. I then realized the Ford and GM sensors have different ranges (GM is 0% ethanol = 50Hz, 100% ethanol = 150Hz), so I  corrected that and it read 11% ethanol, which is a very reasonable value for the E10 pump gas here in California.

About a week ago, I called someone on the phone that I had been meaning to call for some time. I set an appointment with Church Auto Testing to have Grace tuned. The moment of truth is approaching. This Friday we will know how much power she's got. I'm actually far more interested in the shape of the torque curve than I am the maximum power figure but since it's a fun game to play, let me know in the comments how much power and torque you think she'll put down. I'll be doing a 91 octane pump gas tune and hopefully also an e85 tune, so be sure to specify what fuel when you add your guesses!