Some quick initial observations.

The PLX seems to work. I connected and started taking readings with KOEO. I started the truck. The device seemed to have turned off once the truck started. When I was done, I stopped torque and left the PLX plugged in. It did seem to power down before I checked back (15-25 min or so).

The gauges that I was displaying worked and the data I had logged worked. The PLX is visibly slower that the MX. You can see in in the green data light that's in each gauge. I'm not sure if that's a deal breaker, but it was interesting to see such a difference.

Torque reports PID/sec:

PLX 9.6
MX: 69.4

So the MX seems to be about 7x faster there.

I went for a ~15 minute test drive and then counted the logged rows per second based on the logged device time.

PLX: 0.711
MX : 2.125

I'm not quite sure why I don't see the factor of 7 difference in the .csv logging.

 

AE is about in line with the PLX, and I never understood why it's possible to sample much faster - but AE doesn't do it. For those with Infinity, you'll see a ton more data logging, but it records the same reading three times. Going off memory, I believe the Infinity samples the PIDs at about the same rate as the MX, but it fills the logs with the exact same readings times 3.

As I read that, it's not real clear. Here's a demonstration:

ICP on AE:

14:05:03.125, 500
14:05:04.625, 495
14:05:06.125, 498

ICP on MX with Torque:

14:05:03.125, 500
14:05:03.625, 498
14:05:04.125, 495

ICP on Infinity:

14:05:03.125, 500
14:05:03.325, 500
14:05:03.525, 500
14:05:03.725, 498
14:05:03.925, 498
14:05:04.125, 498

At first glance, the data looks really fast, then one realizes it's just logging really fast, but the sampling is no better than the MX. The speed limit is the truck's communication bus, not the OBDII adapter. If the Infinity were to be installed on a more modern vehicle (CAN bus), I theorize the logging and sampling would be in sync... but I don't have Infinity to test this.

 

You have a good point on AE being slow too. I don't think speed is a deal killer for me to continue with this path. The other down side is that you need to hit a button to turn the device on, but it does auto power off. Since the device is on the end of a cable I think I can locate it in a spot where I can easily reach the on button.

When I did my 15 minute test, I was logging quite a few sensors. I cut out just two columns as an example. I took a 10 second window from each log as an example to compare samples.

PLX:
Device Time,#FTE_Injector Control Pressure(psi)
27-Jun-2015 13:29:50.587,536.94000244
27-Jun-2015 13:29:51.928,507.86999512
27-Jun-2015 13:29:53.085,518.13000488
27-Jun-2015 13:29:54.232,524.9699707
27-Jun-2015 13:29:55.436,558.59997559
27-Jun-2015 13:29:56.692,527.82000732
27-Jun-2015 13:29:57.894,534.65997314
27-Jun-2015 13:29:59.073,502.73999023

MX
Device Time,#FTE_Injector Control Pressure(psi)
27-Jun-2015 14:03:50.271,474.80999756
27-Jun-2015 14:03:50.696,489.05999756
27-Jun-2015 14:03:51.242,495.32998657
27-Jun-2015 14:03:51.736,496.47000122
27-Jun-2015 14:03:52.161,475.95001221
27-Jun-2015 14:03:52.572,478.23001099
27-Jun-2015 14:03:53.014,499.32000732
27-Jun-2015 14:03:53.518,494.19000244
27-Jun-2015 14:03:53.941,486.77999878
27-Jun-2015 14:03:54.379,503.88000488
27-Jun-2015 14:03:54.776,486.20999146
27-Jun-2015 14:03:55.378,482.79000854
27-Jun-2015 14:03:55.881,497.04000854
27-Jun-2015 14:03:56.364,481.6499939
27-Jun-2015 14:03:56.856,474.80999756
27-Jun-2015 14:03:57.318,472.52999878
27-Jun-2015 14:03:57.791,467.97000122
27-Jun-2015 14:03:58.267,477.08999634
27-Jun-2015 14:03:58.664,478.23001099
27-Jun-2015 14:03:59.081,486.77999878
27-Jun-2015 14:03:59.552,459.42001343
27-Jun-2015 14:03:59.987,465.11999512

 

As you mentioned earlier, the display is a whole different thing. Here's AE during a first-crank after an injector job:





Here's Torque with the MX on a simple blast:

 

TorquePro has a setting to modify the sampling rate.

Under 'Data Logging and Upload', 'File Logging Interval'

I think the default might have been 1 sec, but I set mine at 0.5 seconds.

 

Hmm, it's greyed out for me. I have synchronous logging set (which I think give me 1 row per sample set vs 1 row per pid?).

 

I don't have synchronous logging on. I don't really see a benefit to is

Maybe that is why it is greyed out?

IIRC my connection shows 41.7k or 47.1k or something like that as the connection speed. It shows in the notification area at the top when connecting.

I have not tried the higher speed option, haven't really needed that kind of resolution yet. Maybe it would let me log more sensors at a time?

How fast is the bus on the truck?

 

I believe the synchronous will write a row after it has polled all the sensors. I suspect the other option will just write what it has at the requested interval. I don't know if it leaves unsampled sensors blank or it repeats the last value or what.


Yes, it's enabled at that point. Now I have to see the difference....

That's the protocol/speed 41.6Kbaud.

I suspect the higher speed setting is related to the bluetooth speed. I doubt the speed at the OBD-II connector can be altered.

I assume 41.6 KBits/second.

 

To put this in perspective: Remember modems? 56K baud - faster than the bus on our trucks.

 

The PLX seems stable so far. It also does stay connected from KOEO through starting and running and shutting off back to KOEO. I'm not sure why it didn't seem to stay connected the first time I tried it.

Is there any specific test or PID that I should try/verify before going to the next step?

 

I've had hiccups with the MX and it would stall the truck when I loaded the EOT PID. OS upgrades/updates, app updates, and software/settings reload shook the snow globe pretty good. I now have no issues with the system. Whatever issue you might have had, it can clear with an automatic update - leaving you with "Funny - it didn't used to do that."

I say load it up with PIDs (one at a time while testing each one) and see how far you can push it.

 

A quick update.

1) Torque will read a single PLX adapter.

Issues:
a) With the digital/analog gauges it only works on the digital portion. This means no peak hold readings.

b) Logging does not work. It doesn't even seem to try even thought it accepts the request to log the sensor.


2) The probe. The NPT tap is welded to the probe. It has 1/2" exposed past the end of the NPT threads. Given the thickness of the manifold this means that the probe will only extend about 1/4" to 3/8" into the manifold.

I know that people suggest that the probe tip be in the center, but I have to believe that temp difference would not be that much different anywhere inside the manifold. Does anyone have a specific reason as to why this could be a real problem?

 

Logging is a function of the app, not the adapter. Do you have Torque, or Torque Pro?

 

Torque pro.

It's hard to speculate as to why it doesn't work. The data is coming in from a different data stream and the software needs to context switch between OBD sampling and iMFD (PLX) sampling.

I guess it's possible that the the iMFD data doesn't fit nicely into the data structures set up for OBD so some "paths" don't work.

I can't imagine why the digital portion of a gauge would work while the analog portion does not.

 

The last time I had issues with Torque Pro, I did an uninstall and an install, then set my parameters by hand again. It was a pain, but the results gave me the reliability and full function I was looking for. This may or may not work in your situation - I don't know.