Limting peak voltage

[Skynet]

Hello everyone,

I'm looking for a way to suppress peak voltage on my A/F LAMBDA SENSOR. The point is to run my car richer during racing and not have the PCM emissions kick in and suppress the extra fuel and oxygen. The way the PCM works is it actively monitors spikes and drops in voltage and if the average is above or bellow a certain percent value over a time window, then an error is thrown with the assumption that something broke, which then puts the vehicle into limp mode. No good. I'm trying to figure out a way to limit peak voltage, very similar to an inductor, but would work on a DC circuit. I cant use simple resistors because resistors also effect the floor of the voltage range, which at idle or cruise speed would cause the lambda value to be to low and an error would be thrown. I need to suppress just peak voltage, not a hard stop, just apply more resistance the stronger the DC voltage gets. Is there a mechanism that does not involve computers that I can employ to give me anything close to the effect?

 

[rjenkinsgb]

It very much depends on the output circuit of the lambda transducer.

If it is would not be damaged by a partial short circuit, a simple shunt voltage regulator or even just a diode or diode + resistor across it may do the trick.
But, that may damage some types of device.

Do you have an exact part number or datasheet for the specific sensor?

 

[Skynet]

The part number is LFJD188G1A. I do not have a data sheet, I don't know where to get one.

Correct me if im wrong, but wont a resistor affect the whole voltage range? This may not work because it would drive the output voltage to low at idol and would throw an error. After doing research I was thinking of using a transistor resistor, as this would allow me to dynamically control how much voltage to let through by controlling the base voltage. But, as you pointed out, doing resistance type shunting or regulating may damage devices.

 

[rjenkinsgb]

I cannot find any data on that device..
From one of the bits of generic data I found, the sensor output typically varies from 100mV to 800mV or more.

A normal silicon diode conducts at around 0.6 - 0.7V, so would limit the voltage to that range.
A diode in series with a lowish value resistor (100 Ohms?) would progressively load the sensor output as the voltage increased above the 0.6 - 0.7V level. It would have no effect at lower voltages.


Adding eg. a 100 Ohm resistor in series with the sensor output to the ECU, then the diode + resistor combination across the ECU side, should half the output above the 0.7V point without too much excess load on the sensor.

Or the 100 ohms series and something like a 470 ohm pot or preset resistor from the ECU side of that to the diode would allow you so "tune" the amount of reduction, above the diode threshold voltage, from a hard limit to only a slight reduction.

For a lower voltage limit threshold you could use an adjustable shunt regulator.

 

[Skynet]

I guess knowing how much resistance the sensor can take before it fries would be a really useful thing before throwing in resistors into the mix.

 

[Skynet]

Would applying resistors before the sensor have similar effect of suppressing peak voltage without subjecting the sensor to excessive resistance?

 

[picbits]

The output of a lambda sensor is slow enough to sample it with a microcontroller and replicate the signal in real time after applying some kind of simple digital filter.

 

[ClydeCrashKop]

Would a zener diode work here?

 

[upand_at_them]

I like the zener diode idea, since it would run in parallel and you could switch it in and out easily. But the lambda sensor outputs less than a volt...forcing a lean conditition would only be about 0.2V, I think.

I suggest unplugging the O2 and plug in a [possibly adjustable] voltage circuit.