1. Home
  2. Electronics Design
  3. Oxygen sensor negative output

Oxygen sensor negative output

I have been testing a Bosch automotive exhaust gas oxygen sensor of the zirconium dioxide type, which I suspect is faulty. There are four connections to it: two wires to a floating heater element, which require

12v of no particular polarity; one sensor output wire and one sensor earthing connection to the body of the device.

The heater correctly takes about 1.7A when cold, falling below 1A as it warms up. The sensor is supposed to give readings between +0.2V and

+0.8v when exposed to exhaust gases, the voltage becoming more positive with decreasing oxygen level. Instead, it gives readings between -0.8v and -0.2v, with the voltage becoming more positive with decreasing oxygen levels. In other words, the readings are behaving correctly but are displaced about 1.0v negative.

According to various websites, this is typical behaviour for a sensor which has become contaminated with silicone residue, and the obvious solution is to replace it. However, I can find no source for such a contamination and some websites imply that the contamination might be on the reference surface of the zirconium oxide 'thimble', not on the side exposed to the exhaust gasses.

My question is: does anyone know which side of the sensor is the contaminated one when the voltage is displaced negative?

I realise this is not exactly a 'design' problem, but I think that I shall only get a proper answer, as opposed to hearsay and guesswork, by asking designers who have actually worked with such devices.

[I have already eliminated bad earthing contact between the exhaust system and the engine, which is the commonest cause of a spurious voltage readings. The vehicle is not consuming large quantities of coolant, so anti-freeze contamination is not very likely. To the best of my knowledge no repairs have been done with silicone sealants] 
--
~ Adrian Tuddenham ~ 
(Remove the ".invalid"s and add ".co.uk" to reply) 
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham
Loading thread data ...

That's what happens when you square "j" ;-) ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     | 
| Analog Innovations, Inc.                         |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| Phoenix, Arizona  85048    Skype: Contacts Only  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

The theoretical output voltage is: Cell output = (2.303RT/4F)x log ( P1/P2)

where:- R = molar gas constant T = absolute temperature of cell in Kelvin F = Faraday constant P1 = partial pressure of oxygen in the reference gas (air in most cases) P2 = partial pressure of oxygen in the sample

So negative voltages imply the partial pressure of oxygen in the exhaust stream exceeds that of the air at the sensor reference port.

Reply to
bloggs.fredbloggs.fred

Just a FYI. If this problem is causing too rich of a mixture, just replace the parts before they ruin the catalytic converter. I've seen this in defective mass air flow sensors. i don't know how this O2 defect will effect the fuel mixture. [Well, I rather not venture a guess since I'd hate to be the person who said it caused the mix to be lean, which is just a power issue, versus too rich which kills the catalytic converter.]

Basically raw fuel makes the catalytic converter run very hot.

Reply to
miso

[...]
[..]

I had wondered if the reversed reading was caused by an exhaust leak causing gasses to stream past the reference port of the sensor, but there is no sign of a leak near of the sensor. (There is a slight leak a few feet away in the back box.)

The negative reading continues with the engine switched off and the sensor heater sustained by an external battery, so it seems as though either the sensor port is blocked and contaminated or the sensor element is poisoned. As the sensor has only been in use about 18 months, I suspect the latter, but I will try flushing out the port with clean Iso-propanol as a precaution.

--
~ Adrian Tuddenham ~ 
(Remove the ".invalid"s and add ".co.uk" to reply) 
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

[...]

The mixture has been all over the place, with prolonged bouts of power loss interspersed with intermittent surges. This only occurred after driving for a mile or two, which is what made me suspect that it was something to do with the fuel control system switching to 'closed-loop' operation. Surprisingly there has been no visible smoke in the exhaust gasses.

On occasions the exhaust gasses have seemed exceptionally hot, as far as I could tell by placing my hand in the gas flow behind the tailpipe. On other occasions they have seemed normal. The hottest (almost at flame heat) occurred during idling immediately after a severe bout of misfiring, so I suspect that some damage may have already taken place.

I can only estimate the temperature during idling, as my simple temperature testing method obviously cannot be used whilst driving.

--
~ Adrian Tuddenham ~ 
(Remove the ".invalid"s and add ".co.uk" to reply) 
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

t

ter.]

From what I can gather from online diagnostics manuals, most fuel controlle rs are aiming for +0.45V average out of the O2S, with the output swinging f rom +0.2 to 0.8V somewhere in the range of 0.5-5Hz rate. Excursions into ne gative output voltages are okay as long as the time average comes to +0.45V . And you're right, it is the excessively hot exhaust that damages the conv erter.

Reply to
bloggs.fredbloggs.fred

You really need the shop manuals for voltages. I have 4 O2 sensors, and not all have the same voltage range.

Frying the catalytic converter is far more expensive than taking it to a garage. This is one of those no win situations, where you pick the lesser of two evils. Kind of like voting in a Republican primary. ;-)

This reminds me of the old radio show Car Talk, where the callers had put black tape over the check engine light for a year before getting around to fixing the problem.

Reply to
miso

There are two basic types: a voltage-generating system base on Zirconium and a variable resistance system based on Titanium. I have managed to establish that this is a Zirconium-based sensor which means that the voltages are determined entirely by the electrochemistry of the process (see other replies to this thread).

Some vehicles have multiple sensors at different places in the exhaust system and I suspect that their readings would differ according to what they were sensing, which would account for your observations. However, no Zirconium-based sensors are supposed to give negative readings, no matter what the exhaust conditions, which is what mine has been doing.

Luckily my vehicle was made in a year where a catalytic convertor was not mandatory and the emissions limits fairly lax. As long as it passes the appropriate emisssions test for a vehicle of that age, I may be able to get away without a functioning convertor.

There would be a lot less of that sort of thing if the manufacturers provided proper diagnostics which were easy to use without specialist equipment. When pintable [pinball] machines first started using microprocessors, there was a simple built-in diagnostic system which allowed the mechanic to read the signal from every sensor and to test-operate every solenoid. With the elaborate dashboard displays in modern cars, a similar system could be added for the cost of a couple of switches and a bit of software.

--
~ Adrian Tuddenham ~ 
(Remove the ".invalid"s and add ".co.uk" to reply) 
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

Many thanks for the advice. I think I have now tracked down most of the causes of the problem. The sequence of events seems to have been:

1) The fuel filter was choked, also the pump may have been worn out. This meant that, on heavy fuel demand such as climbing a hill, the pressure at the injectors fell. 2) The engine management system compensated for the reduced fuel pressure by lengthening the injection period. When the throttle was shut at the top of the hill, the fuel pressure rose quickly and the engine flooded. 3) The extra fuel burned off in the exhaust system and damaged the oxygen sensor. Also, the engine management system either went off-scale or got in a tangle, which caused it to lose track of what was really going on.

Replacing the fuel filter, fuel pump and oxygen sensor has put things right but revealed a trivial underlying fault, which must have been confusing the situation even more: the throttle stop switch was oily and only made intermittent contact.

Once again, thanks to all contributors on this and other threads for helping to save the vehicle from the scrapyard (for now, at least).

--
~ Adrian Tuddenham ~ 
(Remove the ".invalid"s and add ".co.uk" to reply) 
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

he

ire

it

d

ve

0.8v

but

a

on

de

by

est

t

ter.]

e

does it even measure fuel pressure? didn't think many petrol cars did they just have a mechanical pressure regulator recirculating fuel to the tank

and I don't think it would explain flooding, when you shut the trottle the injector period go to zero almost instantly, low or high fuel pressure the engine would floor with full on injectors and no air

e

Generally the ECU won't adjust more than a limited percentage based on the sensor, it can run reasonably in open loop

sometime sensors just go bad, wear out or what ever, every second year or so one of the two on my car needs replacement

-Lasse

Reply to
langwadt

It does have a mechanical pressure regulator. Perhaps I didn't explain myself very well. I think the fuel pressure drop during acceleration caused a weak mixture which the oxygen sensor picked up. The ECU then attempted to compensate for that by lengthening the injector period.

The engine was cutting out and giving intermittent low power, but the inertia of the car kept it turning. The faulty throttle-stop switch meant that the air by-pass valve was fully open a lot of the time anyway, so there was always some air coming through. Once the injectors reacted and stopped flooding, the engine would return to a fast idling condition which allowed me time to change gear and to keep the vehicle moving.

Drivers behind me were often annoyed.

It ran much better open-loop when I eventually disconnected the oxygen sensor altogether.

I will keep an eye on things in future, and treat even recently-replaced items with suspicion.

--
~ Adrian Tuddenham ~ 
(Remove the ".invalid"s and add ".co.uk" to reply) 
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.