[help needed] 4-wire narrow band control

[AndreyB]

Let's assume a generic narrow band O2 sensor like http://www.ebay.com/itm/281240356403

What do we know about heated element control in such a sensor? Apparently sometimes it's not just constant +12v, sometimes they use headed element own resistance or EGT temp for control? https://www.google.com/patents/US7467628

[alfadriver]

Unless you want to find the heater specs from the manufacturer, you are better off finding a sensor from the early 2000's where it's an on/off heater control without a duty cycle. The cycled ones are far more heat than you need- they are used to get the senors lit off quick, and allow closed loop fuel earlier. For emissions.

12V heater, find one that draws about 2A, and it will be hot, but not so hot that it will burn itself out.

[AndreyB]

If I am reading what you are saying, this particular one from the first post - it works with plenty of cars from 1992 to 2002 - would be just simple constant +12v, but it would take some time to warm-up.

Also some newer ones while still using 4 wires would have a more powerful heating element which would allow for a faster start-up, but would require PWM later.

Is that close?

[alfadriver]

Exactly. I wish I had a database of the sensors that would work, but I bet that the MS dudes would have them.

Just got to find the not quite so powerful heater sensors that are out there.

[kb1gtt]

I'm a bit ignorant about NO2, but I think the PWM heaters are often based on the exhaust temperatures more than being a specific sensor technology. If you take a NO2 that's intended for down stream, and you put it close to the engine, you'll overheat and cause the sensor to fail early. I believe MFG's try to make a best guess about the sensors location, but then after the car is built, and they take several measurements, I think they discover the sensor is either to hot and failing to uphold warranty or to cold and failing emissions. If to cold, the easy solution is to either turn it on 100% or PWM it to get the sensor back within temperature specification. I believe they take the car and run it through a series of tests with external calibrated equipment. Then they tune the sensor accordingly and PWM as required. We don't have that external equipment, so we can't make adjustments that way.

I seem to recall you can get single wire NO2 sensors, then place them near the engine, and they will eventually self heat with no heaters, but you need to get it close to the engine such that it gets warm enough. For emission reasons, if you place the NO2 in a self heating region, like you would do with a single wire, they you heat it with internal heaters for the first say 5 minutes, you'll get valid readings faster than if you wait for the exhaust to warm up. Then if you place it to far down stream like after the cat, you might never get warm enough, so you have to keep the sensor either full on or PWM on. It all depends on where the sensor is and how hot that part of the exhaust gets. I don't think the differences in the sensors are very significant.

If you know the resistance coefficient, you can determine temperature by measuring the resistance. See http://www.omega.com/prodinfo/rtd.html?gclid=CNeP7snz_b0CFcyhOgod9lkARQ

I've seen soldering iron temperature controls that measure the irons resistance to determine the temperature. The issue we have is that we don't know what resistance correlates to what temperature, or what kind of tolerance that resistor will have across multiple sensors.

I don't think NO2 is really very worth while, as it's mostly for emissions, and we don't have control over the exhaust design to ensure the sensor is functioning correctly. I guess we could make a hacked attempt at tuning and simply turn it on 100%, then once the tuning is done we can stop using it all together. However I don't think that would work very well.

[alfadriver]

HEGO (Heated Exhaut Gas Oxygen) sensors are indeed heated based on the exhaust temp- or more correctly modeled exhaust temp.

But up until about 2002, all HEGO's were heated on or off- no PWM to heat and maintain heat. That was a technology added to enhance emissions for PZEV vehicles. And those sensors generally are fully operational within 10 seconds of starting- so they are strong heaters.

There are a lot of good sensors that are just on/off heaters. For this application, I would use one of those- no need to go closed loop to quickly. They are a LOT better than the single wire sensors.

As for their worthiness- given the choice of nothing or one of these, I would chose a sensor. But given the availability of WB sensors, that's the way to go.

[puff]

i'd use wb sensor for initial setup and then move to cheaper nb heated sensors. getting an expensive one shocked when diving into a pool is insane and not that rare on 4wd vehicles.
how do we tell if some partiular nb sensor is heated simply or through pwm?

[alfadriver]

Odds are that the MS guys have found a good supply of easy to use HEGOs. Either that, or use ones from roughly 1998-2001.

[rus084]

and at what distance from the outlet of the cylinder head to place it? (72cc engine)

or where it is more convenient?

[puff]

the same question for my 3.5l v8

[kb1gtt]

As for NO2 heater control, I guess the question is how we regulate the temperature. Such a feature will most likely be MOSFET driven, so we can PWM or ON/OFF easy enough. However if someone chooses to have such an option, how do we know a proper PWM? I guess you'd have to install a thermocouple and measure how hot it get's or figure out it's thermal resistance when the engine is running. Perhaps they control the thermal constant for these types of senors and we can reliably correlate resistance into temperature. If so I can propose some circuits for how we can measure the resistance. However as it stands now I'm ignorant about the sensor details, so I can't make much of a suggestion in terms of how the heat can be measured reliably.

[20div0]

Did anyone used this:

http://www.14point7.com/products/slc-oem

It's $40 and works with this wideband O2 sensor for around $50.

http://www.ebay.com/itm/Bosch-17014-Oxygen-Sensor-/400684159884?pt=Motors_Car_Truck_Parts_Accessories&hash=item5d4aa3138c&vxp=mtr

I bought this several years ago and planning to use it when I eventually get around to it.

Andy

[puff]

looks nice and qite cheap. i got a spartan model from the same developer - haven't tried it tho... but the one i have is somewhat a highly-integrated model - once it degrades, i'd need to buy a new one...

[AndreyB]

Back to the original subject, here are some pictures of the O2 heater wire on my 1995 accord.

This looks like "PWM depending on RPM" to me? Not sure what are these short drops @ 1000rpm.

[kb1gtt]

It would appear, it is kind of PWM, but perhaps not quite how you are thinking. I would guess 1kRPM was also a cold engine, then 2.5kRPM is warmer. Perhaps they choose to use the RPM clocks for the PWM. As well, I would wager a guess that if you let it get warmer, I'd bent you get really close to 0% duty. It may need a short duty to measure voltage and current, such that it can predict if the heater is warm or cold. A short pulse can allow the that measurement to happen.

[AndreyB]

I was switching between 1000 / 1500 / 2500 fast and the reaction was instantaneous.

[AndreyB]

I want to revisit the narrow band control subject, now in light of the 2003 Neon.

Hi-side heater. With stock ECU I see ~1.4A during the 1st 5 seconds and then heating current goes down to ~0.8A. Is there hope to drive this with the same VND5E025AKTR-E ?

[kb1gtt]

I'll assume you mean the upstream only, and the other channel will be the alternator field. If so that is a yes you should be able to drive each independently.

However if you mean drive both before cat and after cat with one wire, that's probably a no go. The after cat sensor runs much colder and the heater is usually on far more. You'll probably want individual PWM control over each sensor heater. However your not doing emissions stuff, so I'm assuming you'll ignore the after cat sensor.

[abecedarian]

Most O2 sensors are not PWM heated. Rather, the heater is either on or off. When it's on, the current is measured to determine the heater's temperature and if the temperature is proper, the heater is turned off. The periodic 'spikes' seen on the heater wire are caused by the ECU turning the heater on to measure the temperature and is a function of RPM. However, if you simply turn the key on, you should read a constant voltage until the ECU determines the heater is hot, after which it should shut off and periodically turn on again to check the temp and re-heat if necessary.

The post-cat O2 is solely used to determine if the converter is functioning, and a narrow-band O2 use post cat is dubious at best since the converter should be doing redox on HC and CO, and thus would be reducing the amount of free O2 in the exhaust so if the post cat reflects O2 is less than the pre-cat O2, everything's working... in theory.

[AndreyB]

When it's on, the current is measured to determine the heater's temperature and if the temperature is proper, the heater is turned off. The periodic 'spikes' seen on the heater wire are caused by the ECU turning the heater on to measure the temperature and is a function of RPM.

This defenitely fits the pictures. Is there a trivial schematics we can use to control narrow band the same way?

[abecedarian]

When it's on, the current is measured to determine the heater's temperature and if the temperature is proper, the heater is turned off. The periodic 'spikes' seen on the heater wire are caused by the ECU turning the heater on to measure the temperature and is a function of RPM.

This defenitely fits the pictures. Is there a trivial schematics we can use to control narrow band the same way?

This is what I'm working with at the moment:

O2_Heat.png (3.67 KiB) Viewed 18243 times

You'll probably want some clamping on the O2HEAT_SENSE_P signal, for protection of course.
... and might need to feed into an op-amp for some gain.

[kb1gtt]

The NEON is high side drive. So that circuit won't work unless you have an op-amp that can handle the higher voltages. You could probably isolate by using a hall sensor to feed the analog signal back to the ECU. Something like this should work https://www.pololu.com/product/1185

Also 1A at .02 ohms is .02V. So you'll want at least a gain of 100 if you use the resistor sensor thing. You'll also have to software compensate for variations in gain due to resistor tolerances and offset gain, as 0 amps won't produce 0V out from the op-amp.

[abecedarian]

Realistically, unless Chrysler decided to ground the heater through the sensor body to the exhaust pipe (bad idea anyhow) it doesn't matter how the existing ECU drives the O2 heater- there should be V+ and V- connections in the harness for the sensor(s) so however rusefi decides to drive it is how it goes.

[AndreyB]

should be V+ and V- connections in the harness for the sensor(s) so however rusefi decides to drive it is how it goes.

I want to keep stock harness and not hunt this black G102 wire

[abecedarian]

Yep, I see that now, so stand corrected.

The circuit would be similar, but instead of measuring voltages near ground potential, they'd be near 14+ volts- you could theoretically invert the circuit I posted.
That would influence how the ADC would be set up, as in it wouldn't have a common V- reference with other sensors without some other circuit wizardry, which I'll leave to Jared.

[kb1gtt]

Did that high side driver chip have current sense, or was it just a trip point. I wonder if we get good enough data from that chip that we don't have to worry about how to sense the current.

[AndreyB]