DIYEFI.org Forum

It's easy to take a distributor, put it on the table and see what happens while you are spinning it by hand. I can also put the thermistors into the fridge (both ECT and IAT match viewtopic.php?f=24&t=100 and btw I have created a script to calculate the coefficients for the fancy equation but that's off-topic - http://rusefi.com/Steinhart-Hart.html )... but

O2 sensor? Mine is a generic 1-wire one. How do I know what voltage is 14.7 and what voltage is something else?

What about a random MAP sensor? I am looking to use Toyota stuff - like 89420-02010 or 89420-20190, both read 3.6Volts at sea level. Googling did not help so far. Do I have to get something well know like GM? Do I have to Google better?

Re your script:

viewforum.php?f=31
https://github.com/fredcooke/FreeTherm

One wire 02 sensors are not worth your time. They are also very steep at the only point you care about, so you just have to know/find that. Many/most are 0-1V but understand that you only know "rich" or "lean" from it. They can not tell you stoich or any other value in any absolute way. They're near useless.

MAP, use vacuum and/or air compressor + accurate gauge and/or reference sensor of known calibration.

http://builds.freeems.org/firmware/mast ... tml#l00071

Get one from a Honda, they're a known quantity.

Fred.

I've grabbed a Honda civic sensor - 079800-2540, it reads 2.8volts with air pressure so it is your 1.8 bar sensor.
Connected it to my Toyota sensors - 89420-02010 and 89420-20190 with a Y-fitting, checked what they read at different pressures, did some math... So these Toyota sensors are probably P=V*36.1503419-17.0492, that make them 1.6 bar sensors I guess?

If not a single-wire, what would be a better option? A 4-wire sensor is only different because it would heat itself to start reading faster, right? Would it still be the same narrow rich/ok/lean output? Are any OEM sensors "wider" than others?

EDIT: 98+ Accord, 00+ Nissan, 01+ Cherokee say they use wide band OEM? Do we know the formula for any of these? What would be the oldest vehicle with an OEM wideband?

It's not rich/ok/lean, it's just rich/lean. It's binary. And near useless.

Why piss about wasting time with unknown worn out oem wbo2 crap when brand new stuff is so cheap that even a bum can afford it.

http://www.14point7.com/

105 for spartan < no brainer...
125 for slc_oem eval kit < diy option...

Fred.

Linear Output; 0[v] @ 0.68[Lambda] linear to 5[v] @ 1.36[Lambda], 10[AFR] to 20[AFR] for gasoline

The picture looks like an OEM Bosch sensor

Thank you for the responses, a lot of great info here!

Alan might have to correct me, but I think his controller is just inside the plug and that it is a standard sensor. IE, it takes no more space than a normal plug that you need no matter which controller, just without the external extra controller box.

russian wrote:Linear Output; 0[v] @ 0.68[Lambda] linear to 5[v] @ 1.36[Lambda], 10[AFR] to 20[AFR] for gasoline

The picture looks like an OEM Bosch sensor

Thank you for the responses, a lot of great info here!

Yes the picture has a oem bosch sensor in it but you need a controller to be able to use the wideband sensor, without a controller you won't get any useful data. The controller is in the connector to the other wiring. The controller is outputting .68 lambda at 0v, and 1.36 lambda at 5v.

Even if you found a wideband sensor off a oem car, you would still need a controller for it.

Despite what fred says a narrowband sensor can be useful, I wouldn't use it for tuning. But its application as a diagnostics tool is very useful.

Is it true that the only job of the controller is keep the sensor at the right temperature by adjusting the heating voltage based on the sensor temperature? I am definitely not looking to implement a WB controller at this time, just trying to understand the dark magic of it a bit. If that's not the only job, what else does the controller do? Does it do anything with the actual sensor output, or is it already linear (assuming the temperature is taken care of)?

Hentai wrote:Despite what fred says a narrowband sensor can be useful

Fred wrote:And near useless.

:-p

What do I miss with the inverse of the http://en.wikipedia.org/wiki/Steinhart% ... t_equation ?

So, FreeTherm for the default values gives
A=0.000968129884085426
B=0.000330078990232376
C=-3.05403208764295E-07

these reverse formulas give, for T=303
Y=(A-1/T)/2/C=3818.231246414

x=sqrt(POWER((B/3/C);3)+Y*Y)
and here I am in trouble - the parameter for sqrt is negative, it is -32180527.9148721
Do I read the formulas wrong?