UEGO 30-2310 Evoscan scaling

[Jimvr4]

Thread name is a misnomer - the scaling change is done in the Chrome ROM, NOT Evoscan

I'm using the UEGO 23-10 wideband which comes without a gauge to keep the sleeper look.

The AEM documentation states AFR = 2.375*V + 7.3125 and that 0.5V corresponds to 8.5 AFR and 4.5V corresponds to 18.0 AFR. These are for gaseline of course. If you are on E85 or some other fuel the AFR equation is different so check the AEM documentation.

Now what we have to do in Chrome 2 is open ECU Flash, load your current .bin file and go to the Sensor Scaling section.

Under WB Scaled in the Chrome 2 ROM I generated the values to correspond to the equation (2.375 * (255/x) + 7.3125). The table comes out to

Volts AFR
0 7.3
1 9.7
2 12.1
3 14.5
4 16.8
5 19.2

Flash these values into the ROM and then you are ready to test it with Evoscan.

In the Wideband Scaled request BF in Evoscan the scaling formula should be left as x * 0.0627 + 7. DO NOT CHANGE IT. This is where I initially screwed up.

With this I now have 14.5 - 14.7 AFR at idle

[99 vr4]

Assuming you are using the fuel tank pressure differential input.

The hardware does not use a precision 5v reference. The hardware does have a 5vdc source which is the sensor impressed voltage (in Mitsubishi terms) and then uses a separate ground reference called sensor ground. This 5v source is plus or minus .02 volts.

In measuring this just now on four ECUs:

Supply 13.500 vdc

5.0130
5.0254
5.0114
5.0203



Sent from my SPH-L720 using Tapatalk

[Jimvr4]

Thanks Brett. Could be a terminology thing but ±.02V from 5V is certainly precision. That's around ± 0.5%.

[99 vr4]

I should have said precision per say. .. Since O2 is measured in tenths of a volt, +/- .02V represents a significant change. Correct?

Sent from my SPH-L720 using Tapatalk

[Jimvr4]

Well, 8 bit measurements have a quantization of 0.019607V per bit for a perfect 5V system. But if the voltage reference is 5.02V then the quantization is 0.019682 per bit. Not very significant.

[99 vr4]

Good to know!

Sent from my SPH-L720 using Tapatalk

[Greg E]

Well, 8 bit measurements have a quantization of 0.019607V per bit for a perfect 5V system. But if the voltage reference is 5.02V then the quantization is 0.019682 per bit. Not very significant.

You got it Jim. When I first setup the sensor scaling tables I was concerned about there being enough resolution for the range. I soon realized the sensors readings never go high enough for there to be an issue. ECUFlash won't even let us increment that fine.

Right now I'm working on redoing the scaling for engine load to go beyond the current 2 byte limitations. Again, in a 16bit world there's a lot of space between 1 and 65535.

[Chris@Rvengeperformance]

this is probably just another way, but this is what an aem tech sent me:

"Try this scaling in Evoscan, this is what I’m using for the same 30-2310 controller and it’s dead on for me. Make sure that the brown wire is tapped into the sensor ground on the stock ECU, not to the chassis.

(2.375*((x/256)*5))+7.3125"

It matches my inline gauge I got from dynotunenitrous.com (he actually built it for me, I was the first one to ask for one for the inline aem):

https://www.dynotunenitrous.com/stor...?idproduct=387

[Jimvr4]

this is probably just another way, but this is what an aem tech sent me:

"Try this scaling in Evoscan, this is what I’m using for the same 30-2310 controller and it’s dead on for me. Make sure that the brown wire is tapped into the sensor ground on the stock ECU, not to the chassis.

(2.375*((x/256)*5))+7.3125"

It matches my inline gauge I got from dynotunenitrous.com (he actually built it for me, I was the first one to ask for one for the inline aem):

https://www.dynotunenitrous.com/stor...?idproduct=387

Not sure what their product is but the equation can't possibly be correct for the 30-2310. It would indicate 8.5 AFR for 0V and 18 AFR for 5V which is not what the sensor does.

[Greg E]

this is probably just another way, but this is what an aem tech sent me:

"Try this scaling in Evoscan, this is what I’m using for the same 30-2310 controller and it’s dead on for me. Make sure that the brown wire is tapped into the sensor ground on the stock ECU, not to the chassis.

(2.375*((x/256)*5))+7.3125"

It matches my inline gauge I got from dynotunenitrous.com (he actually built it for me, I was the first one to ask for one for the inline aem):

https://www.dynotunenitrous.com/stor...?idproduct=387

You don't want to change the scaling for logging the wideband in chrome. That formula is setup to read the output of the scaling tables. Using that formula will only show some odd readings.

I would also shy away from using the ECU ground for some of these aftermarket sensors as I've run into issues with it offsetting all the other sensors readings.