AEM X-Series Wideband - Serial

[dr.mike]

Can you try looking for the signal on the PC, using Hyperterm w/ an RS-232 adapter?
The config should be 9600,8,N,1 ( IRRC )
the signal is just plain text
e.g.

14.7
14.7
14.7
... etc,

BTW. as data goes, the 0-5v is massively superior to the slow, coarse, serial data. It is 500 samples/second, vs. 10 samples/second for the serial. And, it has active ground offset compensation.

[spedracr93]

Can you try looking for the signal on the PC, using Hyperterm w/ an RS-232 adapter?
The config should be 9600,8,N,1 ( IRRC )
the signal is just plain text
e.g.

14.7
14.7
14.7
... etc,

BTW. as data goes, the 0-5v is massively superior to the slow, coarse, serial data. It is 500 samples/second, vs. 10 samples/second for the serial. And, it has active ground offset compensation.

Sorry, Dr. Mike - That definitely exceeds my capabilities as a PC user...

Help me understand the data transmission further, though. On the last page of the manual for the gauge, in the specifications table it lists the following:

ANALOG OUTPUT: Resolution = 10 bit
Update Rate = 500 hz

OBDII OUTPUT: Bit Rate = 250/500 kb/sec
Format = 11/29 bit ID
DLC = 8

SERIAL OUTPUT: Baud Rate = 9600 bps
Data bits = 8
Parity = None
Stop Bits = 1

Due to my current hardware that I'm tuning, OBDII is not an option for use.

Now, this is mostly all lost on me, but it has been my understanding that the serial data was superior to the analog because of the speed of the data; i.e. the faster you can get the data the closer the results are to the event within the engine, thereby decreasing the gap that O2 readings follow the combustion. In comparison of the data above, is 9600 bps not equal to 9.6 kHz?

Not second guessing you; just seeking to understand.

If the analog is better, I'm all for that. I've been using an analog WB for some time now and know I can easily set that up!

[dr.mike]

You are confusing the baud rate ( 9600 ) with the data packet rate (10 ).

The baud rate allows up to 9600 bits per second ( ~950 bytes per second ). But, the wideband only sends 10 sample packets per second. So, there is some space in-between. It's slow enough that you can watch them scroll by on the hyperterm screen. It is this slow, only, because it was supposed to be compatible with the older 30-4100 and 30-2310 devices; which, did not have CAN output. And, really weren't any faster than that anyway.

Additionally, the serial data is only 3-digits. "xx.x" So, there is nothing between 14.7 and 14.8, for example. It's really just a vestigial feature; to allow AEM to phase out the older models, without leaving old ECUs that rely on it, out in the cold.

The analog data is updated 500 times / second. So, it is, literally, 50X faster than the serial data.
And, it has the full 10-bit resolution. IIRC it steps from 14.7 to 14.72, etc.

Here is a trace on the analog output. Going from free-air to 14.7AFR


Name: transient-response-hi-lo.jpg
Views: 0
Size: 140.4 KB
ID: 20749

You can see the steps 2ms apart. Even the CAN data only updates every 10ms. Which is one of those dividing lines.



I would never use the serial data over the analog data on the 30-03xx, unless I had to.

[spedracr93]

I would never use the serial data over the analog data on the 30-03xx, unless I had to.

Enlightening and Interesting.

I'll give the analog a shot.

Thanks!

[Tre-Cool]

I might have a play tonight in the drag car with the analog too. I've got 2 guages in it. My biggest problem with the analog appears to more of an issue with the v2's own grounding problems.

I.e with nothing plugged in some ofthe AD ports are showing .02~05 of a volt.

[dr.mike]

Could just be high-impedance floating inputs. which should be OK.
Or... a complete cluster-f^%k. Either way

[Gregs]

Yea, I'm not sure. I wired it up from this: https://forum.efilive.com/showthread...ging-afr/page2

OP said he had success going to 2 and 5, and then the last post from joecar more clearly states what should be the pinout.


Is there a way to verify serial output other than through the V2. I can read a voltage output on the signal wire, but obviously that doesn't correlate to the transmission like the analog output does. It varied from 0.48 to 0.51 V changing similarly to the rate of the AFR output on the gauge.

Dummy me, I should have checked for voltage while I was there

Can you try looking for the signal on the PC, using Hyperterm w/ an RS-232 adapter?
The config should be 9600,8,N,1 ( IRRC )
the signal is just plain text
e.g.

14.7
14.7
14.7
... etc,

BTW. as data goes, the 0-5v is massively superior to the slow, coarse, serial data. It is 500 samples/second, vs. 10 samples/second for the serial. And, it has active ground offset compensation.

I can do this

You are confusing the baud rate ( 9600 ) with the data packet rate (10 ).

The baud rate allows up to 9600 bits per second ( ~950 bytes per second ). But, the wideband only sends 10 sample packets per second. So, there is some space in-between. It's slow enough that you can watch them scroll by on the hyperterm screen. It is this slow, only, because it was supposed to be compatible with the older 30-4100 and 30-2310 devices; which, did not have CAN output. And, really weren't any faster than that anyway.

Additionally, the serial data is only 3-digits. "xx.x" So, there is nothing between 14.7 and 14.8, for example. It's really just a vestigial feature; to allow AEM to phase out the older models, without leaving old ECUs that rely on it, out in the cold.

The analog data is updated 500 times / second. So, it is, literally, 50X faster than the serial data.
And, it has the full 10-bit resolution. IIRC it steps from 14.7 to 14.72, etc.

Here is a trace on the analog output. Going from free-air to 14.7AFR


Name: transient-response-hi-lo.jpg
Views: 0
Size: 140.4 KB
ID: 20749

You can see the steps 2ms apart. Even the CAN data only updates every 10ms. Which is one of those dividing lines.



I would never use the serial data over the analog data on the 30-03xx, unless I had to.

Mike, Thanks for the great info and explanation. Coming from the LC-1 days we were always told to use serial because it was better. Seems funny to be moving to digital->analog for better performance. It's unfortunate that the serial connection is a simpler connection to make vs those orange connectors on the bottom. Due to these findings i'm going to move to analog after I get this serial issue figured out. It will drive me crazy if I don't figure out the solution.

If memory serves me right I'll have to go into some config file to setup what voltage=AFR for this series of gauge. Does anyone have that info handy?

One big thing just hit me, I don't want an AFR output I want a lambda output. Can we change this?

[Tre-Cool]

if you use the analog then your calc pid would be configured for the input/output type.

*CLC-00-2058
V 0.0 5.0 .1 "{EXT.AD4}"
AFR 10.0 20.0 .2 "({EXT.AD4}*2.3750+7.3125"
lambda 0.5 1.5 .3 "({EXT.AD4}*0.1621)+0.4990"

CALC.AEM_AD4 F001 CLC-00-2058 "AFR,lambda,V" WO2-Analog "Wide Band AEM AD4"

if you use the serial port, u need to leave the display set to afr otherwise you get weird data come through.

[dr.mike]

Coming from the LC-1 days we were always told to use serial because it was better. Seems funny to be moving to digital->analog for better performance.

The LC-1 was great, if you knew how to set it up properly ( few did). Otherwise, you had to deal with ground offset issues and HUGE noise problems. So, the serial data was used allot. Even though it cut down the response speed to 12Hz.

This is about the best possible output from an LC-1. With the cables cut down to about 6" and, using 12-gauge wires, with isolated signal ground and differential input. 99% of installations were MUCH noisier than this.


vs. the 30-0300 w/ the full length harness and standard ( single-ended ) input, like the MPVI. Out of the box.


Both are the same transition. From free-air to Lambda 1.00

One big thing just hit me, I don't want an AFR output I want a lambda output. Can we change this?

if you use the serial port, u need to leave the display set to afr otherwise you get weird data come through.

Kinda the same issue here. There is a bug. Where, if you switch the display to Lambda, the serial data, also, switches to Lambda... kinda.
The decimal point ends up in the wrong place. So lambda 1.00 displays as 10.0 etc. Of course, some genius started using that. So, now it can't be changed.

[joecar]

Yes, with the LC-1 analog output, you could use the second channel of the FSV1/FSV2 to subtract out the offset!


Back in the day, with the LS1 VPW PCM's, the pid logging rate is 10 Hz, this is matched to the LC-1 serial comm's 12 Hz sample rate.

But with the CAN bus ECM's, the pid logging rate is 20-40 Hz, so you need a wideband with higher sample rate.


The 03-0330 analog output does look much cleaner... a little bit more overshoot.