rcldesign

It is my opinion that Tony's solution is better because it gives you a number, rather than having to mentally interpolate between ticks on a dial.

2+2

What about the modified gauges from <a href="http://www.gadgetseller.com" target="_blank">www.gadgetseller.com</a> that has a full throttle operation only?

Bri Bro

Tony's solution is better. And you are NOT getting .01V of 'accuracy', you're getting either .019 (~.02) or .0097 (~.01). Odds are its .02 since I don't think the A/D is 9-bits (8 is more typical). And, its not 'accuracy' its 'resolution'

Here are some voltages vs AFR for a WB, from a AFR of 10 to 17 the voltage span is 1.5 Volts to 2.7 Volts or a change of 1.2 volts. Next step is to change the scale from 1.5 to 2.7 volts to 0.5 to 1.7 volts with an Op amp.

If you use an 8 bit A/D, you have 256 step of resolution. So lets say you use a reference voltage range of 2.5 Volts to more then cover the range. The resolution is 2.5/256 = .0097 volts per step. An 8 bit ADC will give 0.0097 Volts resolution.

Vout: AFR: Lambda
1.4 10.08 0.69
1.45 10.23 0.70
1.5 10.38 0.71
1.55 10.53 0.72
1.6 10.69 0.73
1.65 10.86 0.74
1.7 11.03 0.75
1.75 11.2 0.76
1.8 11.38 0.77
1.85 11.57 0.79
1.9 11.76 0.80
1.95 11.96 0.81
2 12.17 0.83
2.05 12.38 0.84
2.1 12.6 0.86
2.15 12.83 0.87
2.2 13.07 0.89
2.25 13.31 0.91
2.3 13.57 0.92
2.35 13.84 0.94
2.4 14.11 0.96
2.45 14.4 0.98
2.5 14.7 1.00
2.55 15.25 1.04
2.6 15.84 1.08
2.65 16.48 1.12
2.7 17.18 1.17
2.75 17.93 1.22
2.8 18.76 1.28
2.85 19.66 1.34
2.9 20.66 1.41

TonyG

beab951

I'm not sure which 02 sensor you have experience with or which a/f gauge you are referring to, but my true wide band setup is clearly very different from yours with a much higher resolution.

Here's the voltage, A/F chart

A/F - Voltage

10 - 1.70
11 - 1.95
12 - 2.20
13 - 2.33
14 - 2.49
14.7- 2.57
15 - 2.73
16 - 3.09
17 - 3.39
18 - 3.59
19 - 3.99
20 - 4.25

Bri Bro

The sensor is a NTK L1H1. The area you need resolution is between AFR 10-14.7, in this range they are almost the same values, your controller range is 0.87 Volts and it looks like the one I have is 1.1V. After 14.7, the O2 sensor curve gets less linear and the gain in the controller plays a factor in increasing step size. In the end, it all get mapped by the micro to the display or the gauge in a lookup table.

I was just showing that a 8 bit ADC can give 0.01 resolution in this application.

TonyG

Your sensor

10-1 a/f = 0.69V
14.7-1 a/f = 1.00v
................
Change .31V from way rich to full stoich

My sensor

10-1 a/f = 1.70v
14.7 -1 a/f = 2.57v

...................
Change = .87V from way rich to full stoich

That's about 3x higher resolution.

Far from the same.

Now factor in Vt delta. Probably much higher. That, and the 1/3 resolution of your sensor, can and does your setup not very accurate.

Bri Bro

What do you mean "my sensor". What sensor are you using? The voltages I listed are coming out of the control unit. The WB sensor generates a pump current, the controller converts this into a voltage output. All of the WB control units I have looked at have a little different output curve but they are very repeatable.

Your sensor
10-1 a/f = 0.69V
14.7-1 a/f = 1.00v

Look again and get your fields straight, you are looking at the Lamda numbers for voltage.

That should be:
AFR=10/1 (Lamda=0.69)= 1.4Volt
AFR=14.7/1(Lamda=1.0) = 2.5 Volts

The delta is then 2.5-1.4V =1.1 volts

My sensor change = .87V from way rich to full stoich
So, 1.1 volts is greater then 0.87 volts by 26%.

That, and the 1/3 resolution of your sensor, can and does your setup not very accurate.

So, since the resolution of your unit using the correct numbers is lower, and using your logic above, your unit is less accurate.

Truth be known, both will be close to the same in the way they function, the step size is just to small to detect on a gauge or a three digit digit display. Tony, you are sounding like a person tryng to sell a product.

Danno

"When can we get our hands on that, and how much?"

Ok, they're in stock. Pricing is $290/each +10 shipping. They're pre-calibrated for the standard O2-sensor by default. The second calibration is for a 0-5v output from a wideband controller unit. The wideband controller that Link is working on will most likely support both an UEGO/NTK-L1H1 sensor as well as the new Bosch LSU4. The controller's output will then match the wideband table that's programmed into the gauge.

Send me an email if you'd like one of these gauges... time to go update the website with this new product..

m42racer

Just to add a smaal something here. Any gauge is better than none at all, and any differences in output of any sensor/interpertation of this data is only bas good as the condition of the sensor. At the end of the day, all final acceptance of A/F should always be based upon Spark Plug appearance. This will then make any system valid, as data supplied would be based upon a difference or change.

I think a good companion to any of these A/F units would be the Knock gauge Travis is selling.

I must admit, I do like the idea of not having to look at a number to tell me if I'm Rich or Lean. In my own experience in tuning EFI systems, getting in the Ball park is always the hard part. Final tuning to get the desired A/F was the easy part. Its just good to see new parts are now becoming avialable to us 951 owners too.