Bri Bro

I really didn't want to spend $1,000 for a wideband O2 sensor system. The problem is I think I need one.

I found an Australian company that is offering a kit that has the controller and the display for $135 AU(with shipping and insurance). End of March, they will offer assembled units for $165. It is PIC based and has an RS-232 port so it looks like it can data log the results.

If the cross to a NAPA (Echlin part) O2 replacement to the NTK Uego Lamda wideband (listed on the site below) is correct, it will cost a little over $139. Motec.com has the NTK sensor by itself for $400 so something looks wrong. Motec also has the Weld-in ring and a 18mm plug ($20) so you can remove the wideband after the testing is completed (sensor life of about 500 hours). Even with the $400 sensor, the cost will be 1/2 of the system I have found so far.

I contacted the company and here is the response;

The Parts Bin part is the NTK in a Bosch box. The Napa part is the same and I think it comes in an Echlin box.
Hmm, I like Porsches - I have an Alfa Sud, Alfa Sprint and two Alfa 33's - a very poor man's Porsche ;-) (well they are both flat!).
Peter.

I have ordered the parts, it's a risk but it looks promising.

<a href="http://www.techedge.com.au/vehicle/wbo2/default.htm" target="_blank">Wideband AFR Meter</a>

David Floyd

Great info, please keep us posted on your progress

<img src="graemlins/burnout.gif" border="0" alt="[burnout]" /> <img src="graemlins/burnout.gif" border="0" alt="[burnout]" />

Konstantin

yes! tell us more about this :-))

Konstantin

adrial

Sweet...I hope they keep making em/designing new products.

I don't NEED one...but its definetely cool and I may get one just to see the a/f ratio's of different chips...

Bri Bro

Some updates on buying the O2 sensor. The following is info from Peter Gargano at Tech Edge Pty. Ltd.

Question: I ordered the part from NAPA but I think it will have a different connector. How are you dealing with this?
Answer: If it has a different connector then it's most likely the wrong part.

Question:I couldn't find a valid listing under the Part Bin or Petro-Ject.
If The Parts Bin link comes up with "Sorry, no parts found." that means they out of stock.

Whats next on my "to do" list: Call the Parts bin and order the O2 sensor, the NAPA part is on order. Post pictures of the Napa and Parts Bin O2 sensors and maybe the kit (5-10 days in shipping). Working on a way of testing the O2 sensors to see how accurate the system is.

Bri Bro

The NAPA O2 sensor came in, it was in a Echlin box but the part has NTK stamped on the sensor and the connectors was the correct type.
<a href="http://boards.rennlist.com/upload/NAPA1.jpg" target="_blank">NAPA (Echlin) Part</a>
The Parts Bin part was less expensive but was on constant back-order. The meter and display have been ordered from Tech Edge Pty. Ltd. Cost to date.
Sensor = $148.88
Tech Edge = $ 85.80
$234.68
I now have access to a Lab grade O2 monitor to test the end product with. All charts show the Tech Edge meter with the NTK wideband sensor should be very accurate (&lt; 2%). Hope to complete the testing in a couple of weeks.

David Floyd

Good work, things seem to be coming together.

Bri Bro

I am very please with what I have to date. I got a Email for a chap in AU that had an interesting idea. He wanted to know if the wideband could be modified to interface into the normal O2 sensor cable, thus eliminating the need to add a second mounting hole. The more I think about this the more I like it. I did a quick design and the circuit is quite simple.

What do you think about this?

Danno

No way that could possibly ever work. The voltage output of the wideband sensor is completely different from the normal O2-sensor:



Not only is the scale different, the direction of change is opposite as well (rising vs. dropping voltage).

Bri Bro

You are correct; you can not connect the wideband directly to the normal O2 sensor. That is why you need an interface circuit.

If the wideband output voltage, available on Con 3 pin 2 of the interface board, is sent to a comparator Op amp. It is compared to a voltage of 2.5 volts ( Lamda=1 ). The output would then switch from a low (0.1 Volts) to a high (0.8 Volts) just like a normal sensor depending on which side of lamda you are on. If and exact voltage curve fit of the normal O2 voltage would be required, it could be produced with a PIC using the built-in ADC and generating the DAC voltage using an I/O port connected to a resistor ladder.

From the information I have to date, most system working in closed loop only look at the state of the O2 sensor, not at the exact voltage. This is why the O2 sensor voltage is constantly switching back and forth when it is operating correctly. So if you apply the Keep It Simple Stupid rule, the Op amp circuit should do and would be dirt cheap, about a couple of dollars. The PIC circuit would be about $10.

If you use a PIC, It could also be connected to a couple of other signals, it is fast enough to do several other functions. It could monitor the injector pulses , RMP and wheel speed. The data could sent to a RS-232 port. You could then log the following info:
A/F ratio (Lamda)
Injector duty cycle
RMP
Speed

Konstantin

if you cann do it work go for it.
I am interested!!!

Konstantin

Danno

You are correct; you can not connect the wideband directly to the normal O2 sensor. That is why you need an interface circuit.



Ok, so you can massage the wideand O2 signal, invert and scale it to have the same range (limits) as the normal O2-sensor:



But your DME will completely mis-interpret the signals; its algorithm won't be able to cope with the linear signal. Notice that the voltage output of the normal O2-sensor is extremely centered on the 14.7:1 air-fuel ratio. The "state" of the exhaust indicated by the sensor is strongly on the "lean" &lt; 0.2v or "rich" &gt; 0.8v side. When the DME inspects "state" of the O2-sensor, it will easily be on the rich or lean side (&lt; 0.2v or &gt; 0.8v) and it will adjust the air-fuel mixture accordingly. And the dithering pattern is closely centered around 14.7:1 with a variation of 14.0:1 to 15.5:1.

If you hook up the wideband O2-sensor and massage the output to fit within the range of the normal O2-sensor, the shape of the output is still different. Now when the DME inspects the sensor, it will find a "too-rich" condition all the way from 11:1 to 18:1 air-fuel ratio. It will completely lean out the mixture all the way up to 19:1 before it gets a &lt; 0.2v signal which indicates a "too-lean" mixture. Then it will add fuel until a "too-rich" state is indidated. It will have to add fuel all the way up to a 11:1 mixture before the state is "flipped" back into "too-rich".

So the range of air-fuel ratios that will result with the wideband O2-sensor is between 11:1 to 19:1 if the DME is still assuming that the stock O2 sensor is hooked up. And it will dither between 11:1 to 19:1 air-fuel ratios instead of the stockrange of 14.0:1 to 15.5:1.

Massaging the wideband O2-sensor to try to make it work with the stock DME also defeats the primary purpose of the wideband anyway; that is to have an accurate display of actual air-fuel ratios. Your wideband O2-sensor output won't work with your wideband display anymore, so you can't precisely tune your car.

Just weld on a additional bung for the wideband O2-sensor and leave the stock systems intact.

Bri Bro

Danno, You have drawn a somewhat linear line (inverse slope of the wideband voltage curve??) between one and zero volts on the y axis and all your statements are based on this line(pink line on chart). This is not the signal a comparator would generate. A comparator (LM339 for one) is not a linear device; it switches states when the input voltage equals the reference voltage. The output will be a step function at Lamda=1, just like a normal O2 sensor.

Next question, why would you want to do this? It allows the wideband O2 sensor to drive both the wideband display and supply the original O2 signal. You don't loose the wideband display because it will use a separate analog monitor output on the wideband system board, the display still works as before.

I just thought the Australian chap came up with a good idea, combine two things into one sensor. All I want to know is if this a function that people would like to have?

BTW, I am an Electronic Engineer and have been around the track a few times. I don't think I can ever learn enough about a Porsche, it is an amazing car.

951Gary

[quote] I just thought the Australian chap came up with a good idea, combine two things into one sensor. All I want to know is if this a function that people would like to have?

<hr></blockquote>
I think the answer must be yes! In fact, I'm wondering if some programing wizards could set up the DME to stay in closed loop for all throttle settings (would it be fast enough?) and use your comparator widget to shift the wideband O2 (use 1.8 V rather than 2.5 V)so that the DME dithers around and A/F of 12.5 instead of Stoic....

If this could be done, forget figuring out the fuel map for your new turbo, etc., just let the DME handle it...

possible?
Gary

Danno

I'm wondering if some programing wizards could set up the DME to stay in closed loop for all throttle settings (would it be fast enough?) and use your comparator widget to shift the wideband O2 (use 1.8 V rather than 2.5 V)so that the DME dithers around and A/F of 12.5 instead of Stoic....

I was just wondering this exact same thing. Figured you could trap the full-throttle switch signal between the KLR to DME. Have that swap over to a second comparator or reference voltage to center the O2-sensor state around a 12.5:1 mixture instead.

However, I'm unclear as to what conditions the DME/KLR combo uses to drop into open-loop operation. I've heard that the KLR uses an TPS acceleration analysis and if you open the throttle faster than a certain rate, it would trigger the full-open TPS signal to the DME even if you're not at full-throttle yet. Also heard that the RPM plays a role with the DME dropping into open-loop as well. In which case, trapping the KLR full-throttle signal to the DME won't help either...