Oxygen Sensor Clamp - Car project.

[GaryMachin]

First, Hello and second I'm new to electronics so please forgive my newbie questions.

I'm looking for a circuit to switch between two voltages.

What I have:
1) A 0.5vdc signal
2) Source Voltage A (0.38vdc) - simulates an oxygen sensor and fed to an engine computer to prevent it reducing fuel to injectors.
3) Source Voltage B (0 to 1vdc) - variable output from an oxygen sensor

What I need:
When 0.5vdc signal is on, source voltage A is passed (to a car engine computer)
When 0.5vdc signal is off, source voltage B is passed.

If the signal voltage were higher (say 5vdc) I would have used a simple mechanical relay, does a solid state relay exist in these ranges? Or, does anyone know of a simple circuit that could do this?

 

[MikeMl]

How about using a transistor to switch a 12V relay ON/OFF? What does the 0.5V signal come from? How much current can it drive?

 

[GaryMachin]

Thanks Mike. Like I said I'm a complete newbie.
The 0.5v comes from a Knock sensing device called a J&S Safeguard. It has boost sensing on board and has two outputs. A fixed 0.5v and a manually variable 0 to 0.9v. Both outputs turn on when the device senses boost pressure from a supercharger. The J&S already has the switching I need built in but it is broken. The circuit inside is potted and non repairable. Replacement of the J&S is around $500US. (All the other functions of the J&S are working).

I guess another option would be. When Source voltage A is on, pass it. When Source voltage A is off, pass Source voltage B. (Note Source voltage A will be set at 0.38v and Source voltage B will vary with oxygen sensor output and it may be above or below Source voltage A)

So, I assume that a transistor switching a relay on/off is fairly common but I have no clue where to start (see, told you I was a newb) and I don't know how much current the 0.5v can support. If we assume that it is robust enough for what you are thinking would you be interested in sketching a circuit or pointing where I can find one?

 

[MikeMl]

The 0.9V signal is more useful for what I have in mind. Can you do a simple test on the 0.9 signal?

Do you have a 1K resistor? Get the knock sensor in the state where it puts out the 0.9V signal that we want to use to pull-in the relay. Measure the no-load voltage with a DMM in DC Volts mode. While watching the DMM, connect the 1K resistor across the DMM leads to see if the 1K resistor loads down the signal. Report back and I will cobble something together.

 

[GaryMachin]

Ah yes, unfortunately the 0.9v signal is also the 0.38v signal that I need for Source voltage A. Sorry, I should have been clearer in explaining that. Will your 1K resistor test apply to the 0.5v signal?

 

[MikeMl]

... Will your 1K resistor test apply to the 0.5v signal?

Yes, I need to know if it is capable of driving ~1K. If it isn't, there is another way to do it, but it will more complicated.

 

[GaryMachin]

Ok, I'll make with the 1K soonest and revert. It will likely be a couple of days. Thanks Mike.

 

[Oznog]

FYI, messing with the O2 sensor signal isn't really a great idea. There's no telling how the computer program in the ECU will respond unless you've got some detailed inside info on how it works. It may look at a sudden change in O2 sensor reading as impossible and simply go into "limp home" error mode for awhile until it sees a consistently reasonable signal again.
The ECU does have memory. An alteration of the O2 sensor can affect the parameter table for some time after the time when you altered the reading, even if you shut off the engine and restart several times.

 

[GaryMachin]

Hi Oznog,
Thanks for your input. In this particular case there is much documented on signal modifying the O2 sensor in a 1990-2005 Mazda Miata (MX5). There is no limp home mode and, as the boost signal is short lived, i.e. only under hard acceleration, the ECU does not register a check light for unvarying O2 signal.

 

[adamey]

The output voltage from a standard O2 sensor is very low current and will not work if expected to drive any sort of load. Exactly which J&S do you have? All the Safeguards I've seen use the knock sensor to modify ignition and have no inputs/outputs related to an oxygen sensor signal. Is it an older system no longer made?

The main problem I see is this:

O2 sensors are useless for measuring air/fuel ratios. They have a very narrow range in which they can measure (14.7 stoich) and when the air/fuel ratio is outside of this range then they bascially read full rich or full lean. They can't differentiate between being slightly rich or very rich - they'll output the same voltage.

The way manufacturers use such a limited range sensor to accurately control air/fuel ratios is in how they interpret the data. The ECU could request a fuel injector pulse width of 5.2ms and notice the O2 sensor reads rich. Then it could request a fuel injector pulse width of 4.8ms and find the O2 sensor showing lean. So the ECU will determine that 5.0ms is the best pulse width to use. By watching the O2 sensor rapidly switch from lean to rich, while making very small adjustments to the amount of fuel added, the ECU can determine to a high degree of accuracy exactly how much fuel needs to be added.

This is why you can't reliably force an O2 sensor output voltage to some intermediate value and expect the ECU to make a direct adjustment to the air/fuel ratio. All it knows is rich or lean.

A wide band O2 sensor will output a voltage that directly represent air/fuel ratio. Of course, if you have a wide band then there's not need to try and "fool" the ECU by sending a modified voltage since it knows the air fuel ratio already.

 

[GaryMachin]

Hi Adamey and thanks for your input,
The J&S I'm using was specifically made for cars modified with forced induction and includes two clamping voltage outputs for the O2 sensor. A fixed 0.5v and an adjustable 0-0.9v. Much experimentation with my particular installation has determined that the variable output should be set to 0.38v. These voltages turn on when the J&S senses boost. When there is no boost the J&S passes the O2 sensor signal unmodified to the ECU. My J&S is broken (it used to work) and when there is no boost it no longer passes the O2 Sensor signal. Under boost it still sends the clamping voltages. I want to use the fact that it still sends the clamping voltages as the trigger to some external switch device to toggle between O2 sensor and the 0.38v clamping signal.

Not to be contrary on this but I do know what I'm doing with the car side of things. Since the supercharger is aftermarket, the ECU has no clue what to do. A piggy back computer called a Powercard also senses boost and modifies the injector pulse to add more fuel to the increased charge air volume. The narrowband O2 sensor, as you said, can't measure this properly and just sees it as rich and it tells the ECU to dial back fuel. Now you have an ECU dialing back fuel and a Powercard adding fuel and they get into a fight the result of which is a lean condition until the engine gets to 4000 rpm and open loop (the ECU no longer uses the O2 sensor signal in open loop). Clamping the O2 signal at, in my case 0.38v, prevents the lean condition. In your example then Adamey, 0.38v would be interrogated by the ECU to equate to the 5.0ms injector pulse and it would stop making adjustments until boost disappears and the O2 signal is again passed unmodified to the ECU.

 

[GaryMachin]

Hello again.
Mike, I went to measure the 0.5v but it appears my problem got worse. The 0.5v signal is no longer outputting. Worse from my point of view, the 0.38v signal has failed too. Without these functional, building a circuit is kind of moot. I'm going to look at Plan B - a vacuum/pressure DPST switch, a couple of resistors and a pot to make a voltage divider for my 0.38v. Thanks for offering a circuit. Time to start saving for a new J&S.