PDA

View Full Version : (exhaust, gas, oxygen) monitor



shogun
12-07-2005, 03:54 AM
Found this interesting unit on a German E30 board. He bought it from Australia. Any Aussie tried that?
http://www.autospeed.com/cgi-bin/browse.cgi?product=888800044&

Here are pics of the assembled unit from the German board.
It shows you the air flue ratio:
2 red LEDs on the right for too much air ratio
right side the 2 yellow ones for too much fuel ratio
6 green ones for o.k.
http://wp1016621.wp027.webpack.hosteurope.de/fotost/f01358/06.jpg

motorman
12-07-2005, 04:06 AM
Ausie website by the name of silcon chip has one under projects to build, page 8, has an alarm too that goes off as well, company makes good kits, I have used some of their kits and they work fine.

BigKriss
12-07-2005, 10:51 AM
http://www.bimmer.info/forum/showthread.php?t=12733

I have yet to install it, but i thought it only works at WOT?

Guapo
12-07-2005, 10:59 AM
Just like all other A/F gauges, pretty much worthless with a narrow-band O2 sensor.

You need a wideband sensor in order to get accurate fuel ratio for metering purposes.

zygoteer
12-07-2005, 04:15 PM
As different sensors will all have different sensitivities (even the ones with the same part number !) and as they are NOT precision devices (if they were, the item cost would be in the hundreds/thousands etc. ), each one would, strictly speaking, need a different calibration on ANY indicator circuit.
The price of this indicator kit means you have arrange for your own calibration - so what would you calibrate it against ? ( a non-trivial task).
Leaving it uncalibrated means the accuracy of the indicator is then doubtful at best, and you are simply left with pretty (expensive) flashing LED's.

In any event, what action are you really going to take when it indicates an bad AFR ?

Martin in Bellevue
12-07-2005, 04:53 PM
Just like all other A/F gauges, pretty much worthless with a narrow-band O2 sensor.

You need a wideband sensor in order to get accurate fuel ratio for metering purposes.
I've had a split second narrow gauge for some time. It really is very close to useless as a tuning tool, with narrowband O2 sensor information.
When I installed Innovate's LC-1 wideband setup, I could tune the map (afm replacement) box very well. It is now running well with good throttle response everywhere. The narrowband had me running rather rich along the rpm range.

That gauge could be used with the LC-1 for referential readings, to be compared to pc readings. The individual lights can be matched to specific afr values. The LC-1 has 2 seperate programmable outputs.

http://www.dynotunenitrous.com/store/ProdImages/LC1_with_gauge_large.jpg

http://img299.imageshack.us/img299/6213/mapmartin3av.jpg

Jon K
12-07-2005, 05:35 PM
As different sensors will all have different sensitivities (even the ones with the same part number !) and as they are NOT precision devices (if they were, the item cost would be in the hundreds/thousands etc. ), each one would, strictly speaking, need a different calibration on ANY indicator circuit.
The price of this indicator kit means you have arrange for your own calibration - so what would you calibrate it against ? ( a non-trivial task).
Leaving it uncalibrated means the accuracy of the indicator is then doubtful at best, and you are simply left with pretty (expensive) flashing LED's.

In any event, what action are you really going to take when it indicates an bad AFR ?

Not true at all.

The device looks for voltage. An oxygen sends out voltage between 0 - 1v (for narrow band) and this indicates to the ECU the lambda value. Narrowband sensors in general are not accurate for AFR, but serve their purpose. There is no calibration needed for an AFR gauge. There is calibration needed on the ECU side, but since it leaves the stock ECU in tact, theres no calibration at all. If the O2 sensor wire send .9v, then that's the lambda value and there's nothing to calibrate.

Wideband Oxygen sensors (I personally recommend the Innovate LC-1, GENIOUS piece of engeering) operate on a 0 - 5v range and are much more accurate than that of narrowband. They are cheap sensors, about $79, but require a controller to interpret the digital signal ($150+). The LC-1 is a programmable controller that reads an input from a wideband Bosch or NGK snesor and can output in 0 - 1v rescaled for accurate narowband gauges, output in a 0 - 5v typical wideband range, and anything in between (ie. 2 - 4v, or anything that some weirdo standalone might require). It can also replicate a 0-1v narrowband signal for sending to the stock ECU so that you can view your AFR with wideband accuracy on a gauge.

I, however, will be running standalone fuel management soon. I am using the Innovate LC-1 in addition to the stock narrowband sensor. This will feed the ECU the true narrowband reading while allowing me to use one output on the LC-1 to feed the MegaSquirt standalone ECU and additionally run my digital AFR gauge. I don't know what else you could ask!

But yeah, lambda reading is an industry standard... 0 - 1v narrowband, 0 - 5v wideband... if yours differs, sell the car its retarded.

bill g
12-08-2005, 01:41 AM
I made an adaptor to plug in at the oxy sensor connector (under starter motor on Aussie M30) so I can hook up a low impedance digital multimeter to read the voltage signal from the sensor - I believe this gives me similar information as the coloured lights arrangement, perhaps not as conveniently. Of course for this to be any use you need to know what the voltages should be.
With the adaptor I also have the facility to temporarily disconnect the oxy sensor output to the ECU (so no feedback). The output voltage then gives me some idea of the fuel mixture without it having been adjusted by the ECU in response to the oxy sensor output.


Bill G Melbourne Australia

zygoteer
12-08-2005, 02:37 PM
Not true at all..... Your text actually confirmed what I said mostly ... thank you, so I really fail to understand your erroneous comment.
I am somewhat dissappointed that a seemingly experienced contributor here could make such a sweeping statement on an opening line.

But it seems you mis-read my words/grammar ... I did not say calibrate the sensor ... I said calibrate the measuring device.

The 'measuring device' in question is merely a simple circuit using a chip from the "LM3914/LM3915" family made by National Semiconductor (one is linear and the other logarithmic - tests done years back confirm that the divider chain and reference have big inconsistencies) ... very old technology (pre-E34 even !) using an untrimmed voltage divider and an untrimmed internal voltage reference - about 1.4USD retail 1-off.
Such a circuit will not be accurate unless there is calibration/compensation circuitry as well (there isn't any on that 'measuring device' !)

However, if you still believe that a 12USD device will give an accurate readout, or if you also think that anything automotive is of real 'instrument accuracy' then you ought to take advice on design & manufacturing economics.

You then go on to say you recommend a type of sensor that requires a 150USD controller to interpret the signal.
Does this really mean you don't believe your own words here, and have decided to spend 10 times more money for a readout etc. for yourself ?

As regards, the consistencies of sensors, if you have actually measured the real output of a selection of sensors under a test condition, and are still convinced they will all give the same spec sheet parameter value, then maybe you need a new or better DVM, or alternatively no fact that I or anyone else could relay is going to change your mind.

I also note that you still refrain from giving your 'raison d'etre' for a real-time AFR monitor, in the face of my question of : "what action are you really going to take when it indicates an bad AFR " ... all this whilst you are driving eh ?

Jon K
12-08-2005, 03:55 PM
Your text actually confirmed what I said mostly ... thank you, so I really fail to understand your erroneous comment.
I am somewhat dissappointed that a seemingly experienced contributor here could make such a sweeping statement on an opening line.

But it seems you mis-read my words/grammar ... I did not say calibrate the sensor ... I said calibrate the measuring device.

The 'measuring device' in question is merely a simple circuit using a chip from the "LM3914/LM3915" family made by National Semiconductor (one is linear and the other logarithmic - tests done years back confirm that the divider chain and reference have big inconsistencies) ... very old technology (pre-E34 even !) using an untrimmed voltage divider and an untrimmed internal voltage reference - about 1.4USD retail 1-off.
Such a circuit will not be accurate unless there is calibration/compensation circuitry as well (there isn't any on that 'measuring device' !)

However, if you still believe that a 12USD device will give an accurate readout, or if you also think that anything automotive is of real 'instrument accuracy' then you ought to take advice on design & manufacturing economics.

You then go on to say you recommend a type of sensor that requires a 150USD controller to interpret the signal.
Does this really mean you don't believe your own words here, and have decided to spend 10 times more money for a readout etc. for yourself ?

As regards, the consistencies of sensors, if you have actually measured the real output of a selection of sensors under a test condition, and are still convinced they will all give the same spec sheet parameter value, then maybe you need a new or better DVM, or alternatively no fact that I or anyone else could relay is going to change your mind.

I also note that you still refrain from giving your 'raison d'etre' for a real-time AFR monitor, in the face of my question of : "what action are you really going to take when it indicates an bad AFR " ... all this whilst you are driving eh ?

I don't think you understand what i said. We agree, a narrowband O2 signal ranges from 0 - 1v. That doesn't offer much accuracy. In fact, narrowband sensors in vehicles aren't even depended on, theyre simply used for extremes, for instance if the car is in open or closed loop, it will check based on a poll of the narrowband that the car is in fact running as such, and additionally will attempt to correct itself in different running conditions. Automotive sensors CAN and ARE sometimes completely useful as measuring devices. VW/Audi use wideband sensors in their NA cars... that's accurate.

Narrowband AFR gauges with narrowband sensors are useful in that they will show if your O2 is coming up to temp (initial readings will be at a certain voltage) as well as if the are is running the WOT map, the narrowband sensor, as "useless" as it is, can tell the AFR of a WOT map ;)

I have use O2 simulators, which simply mock an O2 signal to the ecu so it runs without one. The $12USD gauge is EXTREMELY accurate!!! The sensor is not! There is not enough resolution from the narrow bands 0 - 1v operation to determine close AFR ratios. They are simply a guideline. For instance, if you go WOT the O2 is to confirm the WOT map is being run simply by checking that the car is returning [at some voltage] across the narrow O2.

The "$12" measuring device is in fact accurate. It is the narrowband sensor that is not. There is not enough resolution. So, what you do, is use the LC-1 wideband controller, which uses a 0 - 5v (hence wideband resolution) oxygen sensor, and converter one of the outputs to a narrowband signal. For instance:

Let's say 0v = 7.35 AFR and 5v = 22 AFR... the LC-1 will measure the sensor lets say at 1.5v = 14.7 AFR (Stoich) for idle. Because that is over the 1v INPUT of a narrowband sensor, it wills cale it. a stoich voltage on a voltmeter/narrowband/etc is somehwere around .8v or so.

The LC-1 is accurate to the .1 AFR, which is like .11 volts, which it scales to narrowband and thus can turn what you're referring to as a useless device, the narrowband $12 gauge, into a COMPLETELY functional meter of your AFR.

Why? You've evidentally yet to tune a car. Whether NA or FI, as soon as you do anything to the fueling, you are going to want to see what is happening. The reponse of the outputs are "Instant" "1/12 sec delay" "1/6 sec" and "1/3". The key here is, you can average your readings across 1/6 second of a refresh. If you're driving around part throttle, and you just changed your fuel map, set it to 1/12 or 1/6 and watch your AFR. The "$150" controller is also loggable... so you can record a session, and tweak it out based on RPM. I don't want to sound like an ass here, but I've used 4 Innovate LM-1's and there are perhaps the most valuable tools when messing with a motor. As per the thing regarding the values being different across sensors... that again has never been my finding. We have 3 Bosch and an NGK sensor. The NGK seems to run .1 richer... which indicates a reading of like .01 volts off.

I recommend the $150 controller not because a narrowband won't get the job done for an ECU (because obviously they do, for OEM use) but I would invest the $150 to the controller because it can output a narrowband simulated signal to both the stock ECU and or narrowband gauge. This adds: Wideband accuracy to a narrowband gauge, Wideband accuracy to a narrowband ECU. Sure if you want to go get a wideband gauge, decent ones start at $229+ and you still need the $150 controller $80 sensor... Thats why I said it's worth getting the controller. $30 gauge, $150 controller, $80 sensor... do the math, and it's just as accurate!

Also, I am not sure if you're aware or not, but your O2 should be recalibrated when changing altitude. You cannot do this with a stock ECU. So once the car is adapted to one range of readings, it could run differently +5000 ft.

zygoteer
12-08-2005, 04:42 PM
I understood it completely, but you are missing my point, but it seems your concept of accurate is a lot different to mine.


oh well