LM3914N IC Problems (A/F Ratio Meter)

[motown07]

I'm a newbie to electrical projects, but I saw a neat project for building an Air Fuel Ratio Meter. I created the meter benched tested the meter with a 9V battery. Excited that I had completed my project I wired my meter to my cars 12V battery and not all the LEDs (10) light up. At most I get 4 of the 10 and it takes about 10 minutes for the 4 to light up.

The diagram can be seen at
http://www.scirocco.org/tech/misc/afgauge/diagram.jpg

Instead of the R1 = 2.3KΩ resistor and R2 = 3.3KΩ resistor that was called for in the diagram I used 2.2KΩ resistors.

Can anyone help my to try to figure out why my meter works with a 9V battery, but not with the 12V battery in my car?

[SgtWookie]

Why did you expect to see something other than a set number of LEDs to come on?

In a vehicle with a normally operating narrowband lambda sensor, the sensor's output will continually change from nearly 0v to nearly 1v. It will look like your LEDs are flickering. The ECU will constantly adjust your injector pulsewidth so that close to stochiometric air/fuel ratio will be maintained (about 14.7:1). You really should see any variance under normal condtions. However, operating with wide open throttle may cause your system to go open-loop (no feedback used from the lambda sensor) - and then you may see the LEDs constantly lit or constantly off.

Such a meter would only be useful if you had wideband lambda sensors.

[motown07]

This meter was built to adjust the carburetor on my 1970 Chevelle. The meter uses a narrow band sensor. The weird thing is that if I disconnect the 12v source and hook up a 9v battery to the meter inside the car works how it should work and reads at ~ 15.1:1 or 5 to 6 leds. I confirmed this by removing the meter and connecting a digital meter to the O2 sensor and the ground, meter reads about ~ .59V.

When the 12V supply is hooked up to the meter I get at most 3-4 leds.

Could it be the ground? Can I regulate the 12V source?

[SgtWookie]

Ahh, I see.
I don't have my LM3914 calculator with me; I'm not at home. The LM3914 has a built-in regulator. I don't know why they used an external regulator offhand.

It'll take a couple of days before I'm back at my home computer.

[motown07]

I got this from the data sheet Voltage = 1.25 (1 + (R2 / R1). Is this what I have to use to calculate the voltage?

[SgtWookie]

It's more complicated than that.

The resistors also control the LED current and the reference voltage. You can't just slap in any value resistors; they have to be calculated rather carefully - and it's late, and I'm tired, and I'll probably screw it up terribly if I try at this hour.

Oh, I had a '69 Chevelle years ago - with the "powerslide" tranny lol.

[SgtWookie]

OK, I just have a couple of minutes.

Realized later that they used the LM340 MP 5v regulator to dissipate most of the power that the LM3914 would otherwise have to drop.

The lambda sensor can only indicate the presence or absence of oxygen. One way it outputs nearly 0v, the other way nearly 1v. There is no "in between". So, you would adjust your carburetor so that it went from the too rich condition to just into the too lean condition, and leave it there.

If you really want to know what your A/F ratio is, you will need a wideband lambda sensor.

[motown07]

69 chevelle w/ a 327? Very nice. Thanks for all your help by the way.

I understand how the wb o2 will give me an exact reading, but why am I getting an accurate reading with a 9v source and a lean reading with the 12v .

I could deal with the resistor issue but now i'm in the dark again.

[SgtWookie]

OK, like I mentioned before - I'm not at home, have no reference materials available to me, nor the LM3914 calculator I put together. Unfortunately, this means I won't be able to give you a better answer until I get back home; in the next day or so.

In the meantime, if you could find a wideband lambda sensor, you could get a MUCH better idea of what your actual A/F ratio is. Those wideband sensors have a range of around 11:1 to 20:1 or so - and you could tweak to your hearts' content.

With the narrowband, you'll only know if you're too rich or too lean.

[SgtWookie]

The LM340MP-0.5 is a fixed 5v regulator.
You say that the circuit works better with a 9v battery as a supply.
You can try this test:
1) Disconnect power from the circuit.
2) disconnect the GND terminal of the LM340MP from ground.
3) Connect a multimeter set to the 20mA scale from the GND terminal of the LM340MP to ground.
4) Apply power to the circuit, record the mA reading. I suspect you'll read approximately 5mA current.
5) Disconnect power from the circuit.

Then, take how many volts you want to increase the output of the regulator by, and divide it by the current you measured. The result will be how many Ohms resistance between the GND terminal and GND.

For example, if you measured 5mA current from the GND terminal to ground, and you wanted to increase the output of the regulator by 4v (from 5v to 9v)
4v/0.005a = 800 Ohms. In this case, you could use a 750 Ohm or 820 Ohm resistor to get close - or just use a 1k Ohm pot, and you'll be able to adjust the output of the regulator from about 5v to about 10v.