Pages: 1 ... 10 11 [12]
Author Topic: Logging Zeitronix AFR with ME7L  (Read 99975 times)
rnagy86
Sr. Member
****

Karma: +6/-10
Offline Offline

Posts: 360


« Reply #165 on: March 29, 2015, 02:19:16 AM »

Talked to Zeitronix and they suggest that this is most probably a ground offset issue, so I am visiting my wiring guru tomorrow and we'll see. I will report back.  Grin

UPDATE:

I've spent most of the day rewiring and measuring. Now the Zeitronix is grounded together with the ECU which makes no difference in voltage. I have checked the voltages and the output voltage is exactly the same as what the ECU gets, not even a 0.01V difference and yet the issue remains, so the conversion must be wrong, I have no other ideas.
« Last Edit: March 30, 2015, 08:35:28 AM by rnagy86 » Logged
TijnCU
Hero Member
*****

Karma: +54/-3
Offline Offline

Posts: 690


flying brick


« Reply #166 on: April 02, 2015, 06:02:53 AM »

I have tried to set this up with my LC-2 controller, but no luck so far. I found an adress via DDillenger's code approach, but I just get a steady value all the time. I´m guessing that it´s not the correct adress, but I havent gotten into disassembly of the file yet.

To provide a temporary solution, could it be possible to log USHK when I set up the controller to 0/1v for lambda 0.5/1.5 ? If the limit that port can read out through ushk is 1v that should be possible, provided I calculate the correct conversion.

For my lpg controller I have done a similar thing: I programmed the output to be 0,5v=0.5lambda and 1,5v=1.5lambda. That way I dont even need a conversion to read it, it just gives the right value.

Anyone that logs wideband through uushk_w on a 512kb narrowband without rear o2 that could share their *.ecu file for reference?  Roll Eyes

Locations for the 4Z7907551R:

0x380B28 uuhsk_w
0x380B26 uuhsk2_w

To find the locations in your 2.7 (and many other binaries), hex search for this:

f6 f4 ?? f8 05 8f ?? f8 f3 f8 ?? f8 3d 05 f2 f4 ?? ?? f6 f4 ?? f8 ?e 1? 9a 1? 17 ?0 9a 0?

The offset after F2 F4 is what you want. In the 551R it is 28 8B.

28 8B-->8B 28

(8B28+380000)-8000=Your ram address.


In my file I find F6 F4 EA F8 05 8F EC F8 F3 F8 EC F8 3D 05 F2 F4 24 8E F6 F4 E8 F8 EE 11 9A 15 17 F0 9A 01 so I take 24 8E, make it (8E24 +380000)-8000=380E24. But that one doesnt work  Wink
« Last Edit: April 02, 2015, 06:12:09 AM by TijnCU » Logged

Lost
Hero Member
*****

Karma: +20/-14
Offline Offline

Posts: 556


« Reply #167 on: April 02, 2015, 11:12:47 PM »

It is 38000-8000. Not 380000.
Logged
TijnCU
Hero Member
*****

Karma: +54/-3
Offline Offline

Posts: 690


flying brick


« Reply #168 on: April 03, 2015, 12:30:49 AM »

Aah okay! Thanks Mocke, I'll try the new adress. Reading out <1v on ushk did not work.
Logged

rnagy86
Sr. Member
****

Karma: +6/-10
Offline Offline

Posts: 360


« Reply #169 on: April 10, 2015, 01:13:39 PM »

Zeitronix take #2.

So it turns out that if you unplug the wideband sensor from Zeitronix it will show 14.7 / lambda 1 and emit 2.51V on the wideband linear output (white) wire. With ignition on it stays 2.51V forever, then you start the engine, so there is load on the electrical system (alternator, etc.) and suddenly it will emit around 3.01V (for me at least) with the wideband still unplugged from the Zeitronix. Taking 3.01V and substracting the difference from our provided offset -9.6 I am at -10.1 and suddenly my logger values are spot on with actual ones on the Zeitronix LCD. Next week I am going to try to figure out what causes this interference exactly (leaning toward alternator, or TFSI coil conversion since the coils do not have separate ground so high voltage passes through the ECU). So the Zeitronix provided values are perfect in a laboratory environment for sure Smiley

I'd suggest anyone running Zeitronix and logging the linear output to double check their mixture because it will probably be way richer than logged.
Logged
vwaudiguy
Hero Member
*****

Karma: +52/-37
Offline Offline

Posts: 2014



« Reply #170 on: November 04, 2016, 12:17:16 AM »

Zeitronix take #2.

So it turns out that if you unplug the wideband sensor from Zeitronix it will show 14.7 / lambda 1 and emit 2.51V on the wideband linear output (white) wire. With ignition on it stays 2.51V forever, then you start the engine, so there is load on the electrical system (alternator, etc.) and suddenly it will emit around 3.01V (for me at least) with the wideband still unplugged from the Zeitronix. Taking 3.01V and substracting the difference from our provided offset -9.6 I am at -10.1 and suddenly my logger values are spot on with actual ones on the Zeitronix LCD. Next week I am going to try to figure out what causes this interference exactly (leaning toward alternator, or TFSI coil conversion since the coils do not have separate ground so high voltage passes through the ECU). So the Zeitronix provided values are perfect in a laboratory environment for sure Smiley

I'd suggest anyone running Zeitronix and logging the linear output to double check their mixture because it will probably be way richer than logged.


I know, necrobump.  Smiley

Did anything become of this? Posted factors didn't work for me (not surprised). I had to modify them to align logged output with what the gauge displayed. I am using A 30-0300 X Series gauge from AEM. In the manual it states analog voltage from .5 to 4.5. I used the switched 12v+ from the fused side of the heater wire (because lazy), and the signal reference from the ecu (grey wire) to the brown wire on the AEM (analog negative) with the power ground for the controller to the negative terminal of the battery. Anyone else using this particular wideband controller? Output is linear, but wanted to know if my offset and factor might still be slightly off.
« Last Edit: November 04, 2016, 12:26:14 AM by vwaudiguy » Logged

"If you have a chinese turbo, that you are worried is going to blow up when you floor it, then LOL."
BoobieTrap
Full Member
***

Karma: +3/-0
Offline Offline

Posts: 74


« Reply #171 on: November 29, 2017, 12:59:49 AM »

Hey guys, looks like I am late to the party but nevertheless, can someone please tell me where the location for uushk for 1.8T 018H (http://nefariousmotorsports.com/forum/index.php?topic=12278)
I'm trying to log my EGT gauge which outputs 0-5V signal.
Logged
BoobieTrap
Full Member
***

Karma: +3/-0
Offline Offline

Posts: 74


« Reply #172 on: April 11, 2018, 03:48:32 AM »

Hey guys, looks like I am late to the party but nevertheless, can someone please tell me where the location for uushk for 1.8T 018H (http://nefariousmotorsports.com/forum/index.php?topic=12278)
I'm trying to log my EGT gauge which outputs 0-5V signal.

I got it working in the end, on my 018H wideband ECU it ended up being pin 69 and the RAM address was 2 bytes after uulsuv_w.
I checked it against the reading on my gauge and it is surprisingly accurate.

Much better EGTs after the addition of WMI (used to see upwards of 850C)!
Logged
ejg3855
Full Member
***

Karma: +6/-0
Offline Offline

Posts: 118


« Reply #173 on: January 15, 2020, 01:25:47 PM »

If we have uuhsk_w and uuhsk2_w already defined in out Log CFG we don't need to find the Memory Addresses correct?

Resurrecting a dead thread!
Logged
amd is the best
Sr. Member
****

Karma: +9/-5
Offline Offline

Posts: 262



« Reply #174 on: February 17, 2020, 05:58:13 PM »

Does anyone happen to know the proper factor and offset for Innovate gauges in lambda?

0v = .5la
5v = 1.5la

I tried using a calc mentioned earlier in the thread and didn't seem to work.
Logged

2018 Audi RS3
2001 Audi A4 2.8 30v Supercharged
1991 Audi 200 20v
Trango
Newbie
*

Karma: +0/-0
Offline Offline

Posts: 3


« Reply #175 on: January 06, 2021, 06:50:43 PM »

As a bit of a necrobump, I did a neat trick today that might help someone with calibration on logging a WB with ME7.

In looking to log the wideband, I used TeraTerm and a Serial connection to monitor the wideband's digital/serial output. I then logged uushk_w via ME7Logger, using the exceptionally helpful instructions posted here. My goal, frankly, was to reduce the hassle that myself and others (Kaleb!) would have to deal with when melding data from ME7Logger along with a data stream of serial WB data. I was also concerned that my ancient AEM UEGO controller would not behave correctly AND also potential voltage shifts, mentioned elsewhere in this thread, got me nervous. So, I wanted to benchmark the analog output against the serial (digital) output well enough to let me rely solely on the analog output.

I performed this benchmarking by comparing the data via old-school video analysis. By that, I mean that I took a video of the two windows as they showed data on my laptop, and then carefully compared the amounts as they appeared in several video frames. This method was admittedly crude but really fast and effective. I did this a few times, and iteratively settled on a correction factor that allow me to use ME7 to log AFR values roughly within 0.1 of serial, and also without the apparent FFT-like smoothing that the serial applies (presumably so that the display doesn't bounce all over the place). I settled on an amount that showed a slightly leaner-than-actual value, with the goal of (if there were any error) being more protective than aggressive. This scale factor was .0093, with an offset of -10.

https://www.youtube.com/watch?v=ypvIm1csuUs

Results of the frame grabs, taken at idle:

Code:
ME7         Serial ME7 high by
14.65 14.6        0.05
14.91 15       -0.09
16.24 16.2       0.04
14.56 14.6      -0.04
14.87 14.8      0.07
14.77 14.7      0.07
14.9 14.9      0
15.06 15      0.06
15.2         15.2      0
14.88 14.8      0.08
13.02 13      0.02
18.39 18.4      -0.01
17.03 17      0.03
12.39 12.3      0.09
18.42 18.4      0.02
12.25 12.2      0.05
12.21 12.2      0.01
11.56 11.5      0.06

On a somewhat unrelated note: I also learned that serial connections to wideband controller 0-5v outputs require cables capable of RS232/5v logic levels to be correctly interpreted. I learned this because I had tried to directly splice in a perfect-looking serial-to-USB cord I got on Amazon, but realized belatedly that it was built for TTL/3.3v (aka 3v3) logic circuits. This connection produced some data but nothing recognizable. Splicing in an RS232 connector, and then using a Prolific chipset RS232-USB UART cable, produced perfect numerical data.

I hope these are of value.
Logged
nyet
Administrator
Hero Member
*****

Karma: +508/-101
Offline Offline

Posts: 11255


WWW
« Reply #176 on: January 06, 2021, 07:01:07 PM »

Excellent information! Did you set this up with .5-1.5 output directly to the ECU or did you need a opamp or ttl buffer?

would love to get more information about what you did, including me7logger params and/or map changes in ME7
Logged

ME7.1 tuning guide (READ FIRST)
ECUx Plot
ME7Sum checksum checker/corrrector for ME7.x

Please do not ask me for tunes. I'm here to help people make their own.

Do not PM me technical questions! Please, ask all questions on the forums! Doing so will ensure the next person with the same issue gets the opportunity to learn from your experience.
Trango
Newbie
*

Karma: +0/-0
Offline Offline

Posts: 3


« Reply #177 on: January 07, 2021, 08:50:39 AM »

Hi Nye,

I packed a ton of information into that post in potentially haphazard fashion, and I'm worried that it may have been unclear.

There are two voltages that emerge from a traditional wideband controller, making for two topics:


Topic 1: The serial, digital signal that is passed from the wideband at 5v logic levels. 5v logic corresponds to the RS232 protocol.

Since few computers have actual serial ports anymore, this serial signal is commonly interpreted by using an RS232-to-usb UART adapter cable. 0-5v logic levels, such as are produced by widebands, apparently require a cable that expects those RS232/5v logic levels. I discovered this by first buying and wiring in a seemingly appropriate but ultimately unsuitable cable that expected 3.3v (3v3 in electronics notation) logic levels, which correspond not to the RS232 protocol but instead the TTL logic level. Amazon is rife with 3v3 "serial-to-USB" cables, which look perfect to wire in, but ultimately will not correctly interpret the 5v logic level.


Topic 2: The typical analog voltages put out by wideband controllers on a different wire than the serial logic. It is this analog voltage that ME7 can be tweaked to log via the ADC's used for the rear, post-cat oxygen sensors (pins 69 and 11, for bank 1 and bank 2 respectively). In many cases, widebands emit 0-5v, and the typical scale factors (often around .01) and offsets (usually around -10) applied to uushk_w and uushk_w2 handle this voltage in a way that produces AFR values from 10:1-20:1.

My goal was to calibrate the analog voltages found in topic 2 using the serial voltages found in topic 1 as a benchmark / reference standard. I took this path to deliberately avoid using values and voltages that the wideband controller documentation would cause one to expect, instead favoring experimental data from the serial (topic 1) voltages. Perhaps it's better to say that I wasn't trying to completely avoid using or discard the math suggested by the documentation, but I rather didn't want to rely solely on what was to be "expected".



My discoveries from this exercise:

- As mentioned before, serial-to-USB cables must be at the RS232/5v level - TTL/3v3 cables are unsuitable and will send you on a goose chase.

- The scale factor that my AEM 30-4110 controller required to create near-alignment with the serial signal was different than any value I had seen in this thread. Who knows why that was, but it highlights the value of performing this personalized benchmarking. To confirm, I applied a .0093 scale factor and a -10 offset to the uushk_w variable found at 0x381116. This is not very different from other scale/offset calcs found elsewhere in this thread, but it was different enough. This unique and personalized value was required to calibrate the interpreted analog output to the raw serial values.

- In reviewing the documentation for my WB controller: https://www.aemelectronics.com/files/instructions/30-4110%20Digital%20Wideband%20UEGO%20Gauge.pdf, there are a few calibration settings (listed on page 11) that create analog voltage outputs comparable to the voltages produced by narrowband. This *may* allow logging on ushk or another variable that expects or interprets low voltage and doesn't require finding where uushk_w lives for a particular ECU. I don't know what value there is for this but it was intriguing for about 20 seconds.

- The analog signal seems to be far more resolute and faster to respond than the serial data. My suspicion is that the serial signal that comes out of the back of the WB controller (which is also the gauge, in the case of the AEM system) appears subject to the same smoothing that the gauge experiences. If I had to guess further, AEM has probably realized that users are expecting a slower, less-jumpy gauge behavior, and so they've applied signal processing to the actual UEGO WB sensor data to smooth out the numbers seen on the gauge. In contrast, the analog outputs from at least this WB controller seem more aligned with the actual WB sensor data, making them superior for logging purposes.


Hopefully this at least clarifies some of the stuff I had posted previously.


« Last Edit: January 07, 2021, 09:16:02 AM by Trango » Logged
nyet
Administrator
Hero Member
*****

Karma: +508/-101
Offline Offline

Posts: 11255


WWW
« Reply #178 on: January 07, 2021, 02:11:17 PM »

It does and i would avoid using the serial data for anything but dedicated logging of JUST wb data anyway.

I'm much more interested in the topic of allowing OEM stereo O2 narrowband control to co-exist with mono (or stereo) analog WB voltages directly into the ECU, and using existing (rear o2?) voltage inputs with a minimum of analog hardware (opamps and the like, even it it requires oddball .ecu scaling factors) so not requiring any rs232 silliness (and all of the latency and excess signal processing) and worrying about ttl vs 3v vs 5v.

its easy enough for me to do signal processing in ecuxplot anyway, if needed.
Logged

ME7.1 tuning guide (READ FIRST)
ECUx Plot
ME7Sum checksum checker/corrrector for ME7.x

Please do not ask me for tunes. I'm here to help people make their own.

Do not PM me technical questions! Please, ask all questions on the forums! Doing so will ensure the next person with the same issue gets the opportunity to learn from your experience.
Pages: 1 ... 10 11 [12]
  Print  
 
Jump to:  

Powered by SMF 1.1.21 | SMF © 2015, Simple Machines Page created in 0.027 seconds with 17 queries. (Pretty URLs adds 0s, 0q)