Tired of Wideband/HPT Discrepancy: Going to Try Something

[Z06SUPERCHARGED]

Good read. I just posted over on the GEN 3 about thinking my AEM being off at high rpms in PE as well.

[LSxpwrdZ]

Another note on voltage fluctuations... If a widebands voltage input changes it will affect wideband readings through the analog 0-5v output as well. A solution to this can be to place a capacitor much like a stereo capacitor to the power source your wideband is hooked to that way it will smooth the voltage and supply current where needed instantly.

My wideband isn't a permanent install either. I only put it in the car when I want to tune on it a little. I am using the cig lighter for my LM-1 at the moment and at a constant afr cruising the analog data is bouncing around like an oscilliscope. The digital readout on the LM-1 stay's constant though. So I'm going to play around with hooking it up directly to the battery to see if that helps and I will also try a capacitor to see if that will help. Also note that I have an MPVI Pro cable and I have the analog output grounded to the cable as well.

[DSteck]

In terms of the AFR bouncing around in HPTuners, have you tried slowing down the update rate of the sensor?

[5_Liter_Eater]

Good idea but not possible with the AEM.

[mattgallan]

First off - Thank you 5 Liter for all the info and help for this. Here is my setup and it works perfectly (+/-.1 AFR all the time)

'05 GTO E40 ECM
LM-1 Wideband Controller (Bosch Sensor)
XD-16 Gauge
- wired in through AC Pressure Sensor
- Programmed LM-1 for 0.00V @ 10.0 AFR and 5.00V @ 20.0 AFR
- Programmed analog out for update every 1/12 second (appears to average, which is awesome - and seriously smoothed out the signal)
- SPDT switch for switching between AC and Tuning
- LM-1 common grounded with AC/PS input
- .1uF tantalum capacitor to filter wacky transients on AC/PS Signal line.

The setup is working awesome for me with the equation: (([PID.7101])/.5)+10
It looks to be DEAD on.

[5_Liter_Eater]

You're welcome although you didn't really need this thread. Your PID equation is spot on with what the theoretical is; slope and offset.

I wish I could change the output rate of the AEM, or at least know what it is.

I'm glad others are getting benefit out of this because I'm not. Yesterday my AEM and HPT were off by more than half a point again...WTF?

So now I'm going to focus on cleaning up the power. Thinking of building a regulator similar to this to give the WB controller a constant 12v.

http://linuxcar.sone.jp/reg.en.html

Can you give more details on the capacitor? Where you placed it and maybe a source for it. I can try that first as it's mucho easier.

[mattgallan]

First off I would try the grounding issue. It isn't so much that you ground the wideband controller to the battery but the same ground as the egr/acp signal input. I just grounded my controller to a ground strap that went to the ECM. The +12 is still actually coming from the cigarette lighter, but that is only because I don't feel like running a relay now. It doesn't matter though, the issue here is ground voltage offset.

If that doesn't work then you could try the capacitor. The idea here is to keep it as close to the input to the ECM as possible. In reality it probably doesn't do much, but I put it there for piece of mind I would use tantalum due to it's greater ability to tame sharp voltage spikes compare to electrolytic. That's what they taught me in engineering school anyways
http://mouser.com/Search/ProductDeta...i%2fulZ7Uyw%3d

The main difference in smoothess(as opposed to bouncing all over the place) I saw was from the sampling rate change from the controller. It really sucks that you can't change that.

The power filter you specified should work well, but you have to remember these wideband controllers were designed to work in car electronics with the incurred noise. They "should" have their own filtering methods to keep signals clean. I would definitely try the stuff above before building a filter..lol

[5_Liter_Eater]

Grounding to the AC pressure sensor ground is a good idea. Thanks for the info on the cap.

[KLUG'S SS]

Grounding to the AC pressure sensor ground is a good idea. Thanks for the info on the cap.

Yeah it is!! I hadn't thought of that.

[5_Liter_Eater]

Well it's back into calibration by changing the offset. Now I have [PID.7101]/.474+8.6. 8.6 is really low but whatever. Whats interesting is the slope is pretty consistent @ ~.474. Waiting for it to change on me again...

[mattgallan]

Oh and another thing if all else fails is to put a resistor, probably about 10k ohm from the +5 on the sensor to the sensor output with the engine off. Measure the voltage the ECM is seeing with a multimeter and confirm the ECM is reading it correctly via HPT. I did this when I was installing everything and it was +/-0.01 V

[2therock]

Hi Guys,

I'm trying to straighten my PLX WB out. This is what I have done to attempt about a 2.5 % error between the gauge and HPT.

I have not tried it yet. Does it look OK for starters?

[5_Liter_Eater]

If it gives you accurate results ~14.7 then it seems like it's fine for starters. I don't think you need to go out to eight decimal places though.

[2therock]

Thanks, And For Both Replies,

Looks like I have two options. We will see.

I changed it once and there was no change, maybe I goofed. I'll be going in again.

[carlrx7]

5 liter,

is this only required if your running the wb threw the EGT or A/C Pressure Switch? or is the pro having the same issues..

im getting the pro for logging and already have the aem UEGO permently installed.

[5_Liter_Eater]

I can't say for sure because I have never dealt with EIO/Pro but I do know that there are not many forum members who are able to use the built in wide band PIDs and have them match the gauge. Most need to create user defined EIO PIDs to change the offset. EIO or AC/EGR; If you having to change the offset you should verify the slope too.

[JimMueller]

I seem to be having STFT/WB discrepancy and I thought I'd try this way because I don't remember how to use the old AFR offset custom input.

So you're using a video camera to capture the WB voltage and AFR... and have the laptop next to it showing the reported AFR. I'm using an LM-1 analog output connected to EIO... what do I scan to get the HPT volts value for your spreadsheet?

If I set my LM-1 to report, say, 3V at all AFR's, then I could just average the difference between what HPT reports as the voltage reading and 3V, right? Or are you saying that the voltage offset fluctuates under other conditions?

I did just that, then created a custom EIO linear PID: (Volts/1)+0. I created a histogram and watched the average. It began around 3.2 and over 5000 frames it went up to 3.32 and was still slowly climbing. Not sure if it would have ever stabilized.

So if there aren't any problems with that method, then am I correct in thinking that my LM1 custom PID should be:

(Volts/(volt range/AFR range)+(lower AFR limit + offset)
(volts/((5-0)/(17-11)+(11+0.32)
(volts/(5/6)+(11.32)
(volts/(.833)+11.32

?

[5_Liter_Eater]

So you're using a video camera to capture the WB voltage and AFR... and have the laptop next to it showing the reported AFR. I'm using an LM-1 analog output connected to EIO... what do I scan to get the HPT volts value for your spreadsheet?

I'm displaying the AC voltage because that is what I have my WB piped in through. You would create a custom EIO PID. Use the linear function and just use Volts/1+0. This will give you the raw voltage that HPT is seeing. Then use the video camera to see what AFR the LM-1 says and Volts HPT see's at that time and plug those values into the spreadsheet to generate the slope. Then change your AFR equation to use that new slope and use a digital camera to take some snapshots of the LM-1's AFR and HPT's AFR to derive a new offset.

If I set my LM-1 to report, say, 3V at all AFR's, then I could just average the difference between what HPT reports as the voltage reading and 3V, right? Or are you saying that the voltage offset fluctuates under other conditions?

I did just that, then created a custom EIO linear PID: (Volts/1)+0. I created a histogram and watched the average. It began around 3.2 and over 5000 frames it went up to 3.32 and was still slowly climbing. Not sure if it would have ever stabilized.

I'm not sure that will work. It will definitely give you an offset but doesn't do anything to help you verify the slope. My theoretical slope was .5 and my new slope based on the data was .475.

So if there aren't any problems with that method, then am I correct in thinking that my LM1 custom PID should be:

(Volts/(volt range/AFR range)+(lower AFR limit + offset)
(volts/((5-0)/(17-11)+(11+0.32)
(volts/(5/6)+(11.32)
(volts/(.833)+11.32

?

Again, this method is good to find the offset but my problem was also that my slope was not true to the equation and thats what I was trying to verify. Assuming the slope is correct then the offset is easy to find out by taking several digital pics of the LM-1's AFR and HPT's AFR and find an average difference.

[schnellTTZ06]

This is a good thread and made me decide to troubleshoot my own AEM because its always been off from .5 to 1 point at different afr's. I just kinda took this into account as just being like that and to make mental note. Anyway I do have an electronics background so the ground offset made me curious. Here are my findings. Withe the AEM WB you can disconnect the 6 pin connector and it is supposed to give a reading of 2.35 volts. I found this to be off, actual reading 2.448 volts when I used a chassis ground. And it was 2.438 thru the MVPI ground. I also ohmed out the ground between the MVPI and the chassis and found a 7 ohm resitance. I then tested what HP tuners was reading under the "raw voltage" section and found that it read 2.45 volts. I also ran the car up to about 5k rpm to see if the voltage would drift up or down and it seems that the voltage regulator on the alternator does a good job, the voltage remained stable. AEM also has a voltage breakout for each range from 0.00 - 5 vdc in a strangly non linear way in increments out to the hunredths position. I dont know exactly how the engineers at HP tuners setup their predefined wideband selections, but AEM is definitely not precisely linear. So I used the list feature for every voltage range. I found that my guage (still 6 pin connector disconnected) reading a solid 14.8 (2.4 vdc) was about .1 off, thus the scanner should have read 14.8 but read 14.9 so I adjusted until they were the same and then applied the difference to all the other values. I suppose you could use the camera method as stated in an earlier post and not use the slope method as its not very accurate. Here is what my list ended up being. I hope this all makes sense, I would also appreciate feedback on this, I havent made any runs yet to see how accurate this is but Im sure Im on the right track. Also, I think a jumper could also be intalled on the MVPI to compare the separate WB formulas simultaneously.
0.0,10.000
0.1,10.220
0.2,10.370
0.3,10.520
0.4,10.840
0.5,10.990
0.6,11.140
0.7,11.460
0.8,11.620
0.9,11.780
1.0,12.080
1.1,12.240
1.2,12.400
1.3,12.450
1.4,12.700
1.5,13.020
1.6,13.180
1.7,13.340
1.8,13.640
1.9,13.800
2.0,13.960
2.1,14.260
2.2,14.420
2.3,14.580
2.4,14.700
2.5,14.900
2.6,15.200
2.7,15.360
2.8,15.520
2.9,15.820
3.0,15.980
3.1,16.140
3.2,16.440
3.3,16.600
3.4,16.760
3.5,17.080
3.6,17.230
3.7,17.380
3.8,17.540
3.9,17.700
4.0,18.000
4.1,18.160
4.2,18.320
4.3,18.640
4.4,18.790
4.5,18.940
4.6,19.260
4.7,19.410
4.8,19.560
4.9,19.880
5.0,20.000

[JimMueller]

I'm sorry, I just can't get my head around how to show the entire voltage range..

- if I use 11-17AFR= 0-5V to represent my normal analog output, then I would need to at least idle the car from 11-17AFR... is it safe to be that lean even at idle?

- The LM1 display shows lambda and afr, but supposedly if I use the internal LM1 logging it will record voltage, but then I can't compare it to the HPT voltage

- It doesn't appear I can log LM1 AFR and LM1 voltage simultaneously within HPT because they are both using the same analog connection.

Any suggestions?