Auto Enginuity 8.2.1

Hi,

I was about to open a trouble ticket with Auto Enginuity when I noticed they released version 8.2.1 on April 7. I don't like to open trouble tickets unless I'm running the current version and just thought I'd share this with folks.

My trouble ticket will be on the "MG1 torque." The data strongly suggests the units and the values are very, very wrong:

• nearly random sign flips - yes the sign changes when starting the ICE but during normal operation, it should always be of one sign as the torque has to apply against the engine torque to get any torque to flow through the PDS. This might look like a 'mad dance' but it is worse. It is a dance viewed through a kaleidoscope or broken mirror.
• MG1 torque and rpm may be one query - the Graham miniscanner gets both values with each query. If both MG1 torque and rpm are requested in the same series of data requests, the data lines have both from the same time-stamp read. This may be what "info 1" and "info 2" are all about.

I recorded a bunch of data on Sunday with MG1 torque and plan to convert the values into bit-maps and do a statistical analysis. I suspect the data conversion is not right and by looking at the bit patterns, I may be able to figure out what is going wrong. Think of it as trying to decipher a badly encoded message.

RELEASE NOTES SUMMARY:

"
General
Improved Mode 6 retreival for J1850PWM controllers (Requries firmware Version 26)

AE ProLine Firmware (Version 26)
Improved J1850PWM communication
Fixed Audi/VW communication issues on ISO-based modules
Toyota / Lexus / Scion
Added missing Hardtop sensor support
Added missing TPWS support to '06 Prius"

Bob Wilson

Heh, good luck =). I tried reporting 3 or 4 bugs to them and they just told me that it worked on their end and they'd file an official inquiry into the issue...Never heard anything back and they wanted me to pay them to "upgrade" my hardware that may or may not fix the problems.

Andrew

Bob, I have seen this about MG1 torque values being bonkers.
I think part of this is that it remembers the last value if not updated recently. Now with monitoring many things at once, it will take quite a few "samples" before the data will become coherent. Depends how often that particular parameter is actually updated, and data can change a lot in say a few seconds.

I will try increasing the sample rate by sampling only one or two parameters soon by running:
#1 - record Batt HI blk and Batt LO blk voltages only,

and then on another session I plan to:
#2 - record MG1 and MG2 Torque only in one session.

Go to the grid data screen first and make sure only these are selected, and only these have bar graphs, then record. Go for a spin.

Don't forget that the car must be started before opening the interface, so just that fact -you will miss the "cold" state of things.
After startup it immediately goes into charge mode in "P" still.
By the time the software starts up, MG1's already been working
as a generator for a while, and could be at 380rpm.

Regarding ver. 8.2.1, I have downloaded it, but not installed yet.
I did see that the only improvement for toyota hybrids is:

"Added missing TPWS support to '06 Prius"

Just set the key to IGN, don't go to "READY." All of the ECUs will be enabled and it is easy the catch the startup.

The Graham scanner gets things right. I'm beginning to suspect there is something wrong at the lower layers.

I believe Hobbit and Graham Davies both reported the NHW11 OBD is running a non-standard data stream at ~11 kbaud. I'm thinking about setting my Mac to sample at 44 kbs and see if I can read what the Graham scanner is doing versus the Auto Enginuity.

Bob Wilson

Here is a description of the problems experienced this morning:

Auto Enginuity Problem

Bob Wilson

Hi Bob,

Have you tried using this device on a non-Prius vehicle? I'm wondering whether it works satisfactorily on a low-tech car.

I used it on a 2002 Tacoma pickup truck and our ZVW30. It worked great on the Tacoma and OK, generic mode, on the ZVW30.

BACKGROUND:

• 2002 Tacoma cab pickup
• 160,000 miles
• warmed up from lunch, in parking lot

RESULTS:

The Tacoma pickup had a "Check Engine" light and code indicating a failed catalytic converter. Monitoring the O{2} sensors, one side showed the downstream catalytic converter was fubar. He asked if the check engine light could be cleared.

I told him I could clear the check engine light but it would come back. Since we live in Alabama that does not have annual vehicle inspections, he was good with that. Two days later he reported it came back.

I recommended:

• pull and inspect the spark plugs looking for one weak cylinder or one side showing excessive heat from a possible air leak
• changing O{2} sensors was unlikely to fix the problem because the output values on the bad and good side were showing similar values
• replacing the failed catalytic converter is possible from 3d party internet sources and a muffler shop could put it in cheaply but without finding and fixing the root cause, it would fail again

Bob Wilson



Bob Wilson

I'm sorry you're having that problem. I have not seen the "triangle of death" from AE yet on the 03, only from scangaugeII using xgauges. Just make sure it connects with iso-9141 and OBDII standard only. Don't use kwp2000 or can. I have found that if I last ran it on the 06 in can mode, before connecting to the 03 I must re-initialize the hardware to be iso-9141, then connect it to the car not before. Don't have the car back until Sat., but will try the cold "on" first and try to capture Hi blk and LO blk voltages and then record just those two parameters on a ride. Should be interesting.

I suspect there is more going on at layer 1 and 2 than we know about. It is entirely possible my laptop may be part of what is going on. But the first step is to document the facts and data.

That seems to be key to avoiding the problem.

Hummm, so the operational scenario is:

1. connect ProLine to laptop USB
2. boot laptop
3. start AE with communications set to "ISO-9141" and "CAN II"
   1. wait for it to ask for the car to be enabled
4. connect ProLine to DLC connector
5. enable car

I'm pretty sure this works OK but it is different from the order in the flyer:

1. Locate DLC connector
2. Connect ProLine to DLC
3. Connect ProLine to USB port
4. Turn vehicle on

Then comes the section "Operating the Software"

1. Launch ScanTool

I suspect the order in the flyer that came with the manual is the problem and your suggestion of letting the laptop set the ProLine protocol first makes a lot of sense.

Thanks,
Bob Wilson

Bob,
Concerning :Hummm, so the operational scenario is:

1. connect ProLine to laptop USB
2. boot laptop
3. start AE with communications set to "ISO-9141" and "CAN II"
   1. wait for it to ask for the car to be enabled
4. connect ProLine to DLC connector
5. enable car

I do this:
1. Boot laptop
2. After bootup, Connect proline hw USB cable, will hear USB wake up on laptop.
3. Plug Proline onto OBD port in car (car is still off) notice leds..
4. Start AE software and select iso-9141-2 and OBDII proto, not CAN. (for 00-03 only)
5. start car when it is in 2nd detection stage after "proline hw detected".
6. Select which interface you want to observe (Batt ECU, HV ECU etc)
7. Let software detect any DTC's ( it will then detect codes every so often while running. This is Normal, but may be reason why hiccups in data..
8. In "grid data" select which parameters you want to record and make sure they all
have active bargraphs before recording file.
9. Name the record file.
Walla! No triangle of death.

I'm running it on a 1Ghz w/3Gb ram, USB2.0
Actually it only will work with USB2.0
Perhaps you don't have enough ram in the system?
That could really run it into bottleneck mode.
So far on the 03 I've had no problems other than the data being a little squirrly
as you know...

I just turn it to IGN but not READY. This brings up the ECUs so I can capture vehicle startup states. A little rough on the 12 V battery but I've got a good one.

This makes sense.

Hummm, I've got 1.83 Ghz dual Pentium but 512M RAM. I'll see about a memory upgrade. It has USB2.0. I just went through the windows configuration options and enabled "Adjust for best performance."

Bob Wilson

"I just turn it to IGN but not READY. This brings up the ECUs so I can capture vehicle startup states. A little rough on the 12 V battery but I've got a good one."
Will try this on Saturday.
Yes your system sounds good enough except for the 512M ram. If the system runs out of memory, it will start caching to hard disk. This will really slow down the sample rate. A windows thing. I would think at least 1Gb ram, the more the merrier. It's always something with this stuff. BTW my system is WIN XP Pro for tablet, but any version of xp would work.
They say it works with WIN2K as well as WIN 7, but I've only had it on XP pro. I do know that the USB port stuff is handled differently in 2K, don't know about 7 tho...XP is the "holy grail" version at work where we do lots of hardware comms via USB. Must be a reason they won't go to Millenium or 7.

You guys are having a great discussion. Memory is so cheap, it makes sense to install the max that your laptop can support.

I guess I'm in no rush to buy this AE software, will allow more time for the bugs to be worked out (unless my 2004 decides to turn on a warning light.)

I am enjoying keeping up with this great discussion also. I am looking forward to this fall when AE has the software ready for the Gen III. Maybe by then Bob and the others will have most of the bugs worked out and AE will respond and fix their issues

Interesting, I'm getting push back from AE support. They really think I'm seeing legitimate ABS errors. It is a curious kind of error that shows up depending upon the order and configuration of the ProLine connector and AE. Yet if I don't use the ProLine connector, I get no flashing ABS light. But there was a recent posting by Dan Hyatt in "Prius Technical Stuff" where he speculates about a weak AUX battery inducing problems.

We've known for some time that a weak AUX battery can lead to a number of intermittent and random problem reports. It has become standard practice to recommend AUX battery replacement if it is OEM and more than 5 years old. Now I'm beginning to wonder if it makes sense to look closer at 12 V battery and Prius startup.

As for AE, I'm of mixed mind. I don't have 'a dog in that fight' and if we develop effective, user procedures to make this tool "s*ck less," it makes sense. On the positive side, the tool queries many elements not reported by either the Graham scanner or much less, over the counter scanners.

I think it makes sense for us to figure out what works and what doesn't work and share that among ourselves. I don't think AE has much interest in our 2001-03 Prius and their tool. Sad, but not everyone is a Graham Davies.

Bob Wilson

Hi,

The latest not from Auto Enginuity indicates they still think the ABS codes are legitimate. So this morning, I again used the Safe Startup:

1. Start laptop and Auto Enginuity with ProLine USB connected.
2. Wait for Auto Enginuity to ask about the car ON status.
3. Car must be off before connecting ProLine to DCL.
4. Turn Car to IGN or START
5. Auto Enginuity completes initialization

To change which controller:

1. Use menu "disconnect"
2. Use menu "connect" and wait for Auto Enginuity to finish communications
3. Use menu "Change controller" - it should respond
4. Select new controller
5. Wait of initialization to complete

I was able to change to the "ABS" controller and e-mailed the recorded data file to them. We're still in discussions and hopefully, I won't have to start making MPEG movies showing what is going on. <GRINS>

Bob Wilson

When I took the car to Firestone, they also read the codes and found nothing. I faxed the report to Auto Enginuity. I'm not holding my breath.

This evening, I used "ISO-9141" and "KWP2000," the variable length version. I remember Graham Davies posting something about this in PTS. There still is an initialization problem and it looks like the first use hangs. But if I pull over and restart it, it becomes stable. I recorded at least 20 minutes of data this evening, a personal record.

I don't know the voltage levels involved but will use my audio recorder trick to get an idea and verify the timing. Wiki reports it uses 10.5 kbaud and if so, this is easily handled using an MSP430 kit. I think when I start sending them layer 1/2 based reports, maybe they'll try to fix the problems.

It is sad to deal with a supposed technical organization and find their first level, tech support is . . .

Bob Wilson

FYI, the support staff at Auto Enginuity insist that the ABS codes are real and dismissed the Firestone report. They insist only Toyota and their tool can read these mysterious codes that go away by power cycling the car.

Thursday, I began testing "Speed Tracer" on a drive back. My first efforts looked at the MPG trace:

• 7:28:00 - 7:29:59 :: barely two minutes of data before the ProLine stopped sending data. But I was already headed home and didn't have time to fiddle with it.
• 7:45:10 - 7:52:40 :: this time over seven minutes of data before it stopped. Again, I was on the road and didn't have time to fiddle with it for another 10 minutes. I had to power cycle the ProLine by disconnecting from both the vehicle and USB.
• 3d time the charm :: again, a power reset of the ProLine only this time it continued running for about an hour before my next stop.

This is the pattern of either a thermal or marginal electrolytic capacitor. As a test, I'm leaving the ProLine USB connected over night and will see if this leads to reliable operation when first connected to the Prius. I'm looking forward to testing our Prius warm-up.

My warm-up protocol involves cutting through the neighborhood and using "N" to coast as much as possible. This tends to reduce the warm-up cost in the first 1-2 miles. The engine continues to run so it warms up as before. However, I avoid the higher fuel consumption that happens in "D". This may be a great way to document the effectiveness of this protocol.

Bob Wilson

Only use ISO-9141 and standard OBDII on the gen1. I have found that with anything else it will spike pins 6 and 14 while doing its mode query, in other words it's looking for a can bus first. If you configure the interface only to ISO-9141 and have it running before connecting to the system (starting the car) there's no problems then.

As recently described on a recent post, the scangaugeII will also cause the spike and "ABS" warnings, however now that we've found out the remedy (p6 and 14 removed) all is well now.

Now that the pig-tail gives stable operation, I can analyze the data to look for data quality and quantity. For example, I enabled all "HV ECU" data fields and drove around while recording the data, over 300 values. But only these 38 changed during this test drive:
Signal name|units|range
Master Cylinder Control Torque |(Nm)|[-512.00 - 508.00]
Engine Speed |(RPM)|[0 - 8000]
Target Engine Revolution |(RPM)|[0 - 8000]
Request Power |(W)|[0 - 45000]
Generator - MG1 Revolution |(Nm)|[-16384 - 16383]
Generator - MG1 Torque |(Nm)|[-500.00 - 500.01]
Regenerative Brake Torque |(Nm)|[0 - 160]
Request Regen. Brake Torque |(Nm)|[0 - 160]
Motor - MG2 Revolution |(RPM)|[-16384 - 16383]
Motor - MG2 Torque |(Nm)|[-512.00 - 508.00]
Generator Current V |(A)|[-250.00 - 250.01]
Generator Current W |(A)|[-250.00 - 250.01]
Motor Current V |(A)|[-750.00 - 750.01]
Motor Current W |(A)|[-750.00 - 750.01]
Motor Temperature - MG1 |(C)|[-50 - 205]
Motor Temperature - MG2 |(C)|[-50 - 205]
Inverter Temperature - MG1 |(C)|[-50 - 205]
Inverter Temperature - MG2 |(C)|[-50 - 205]
Drive Situation ID |(#)|[0 - 6]
SOC |(%)|[0.00 - 100.00]
Auxiliary Battery Voltage |(V)|[0.00 - 20.00]
Vehicle Speed |(kph)|[-256 - 255]
Battery Cell Temperature Max |(C)|[-128 - 127]
Battery Cell Temperature Min |(C)|[-128 - 127]
Acceleration Sensor Main |(V)|[0.00 - 5.00]
Acceleration Sensor Sub |(V)|[0.00 - 5.00]
Shift Sensor 1 |(State)|[0 - 255]
Shift Sensor 2 |(V)|[0.00 - 5.00]
Power Resource IB |(A)|[-256 - 254]
Power Resource VM |(V)|[0 - 510]
Driving Pattern 3 |(Mask)|[0 - 4]
Driving Pattern 2 |(Mask)|[0 - 4]
Driving Pattern 1 |(Mask)|[0 - 4]
Loading Condition |(Bit)|[0 - 1]
Engine Warming Up Request |(Bit)|[0 - 1]
Aircon Request |(Bit)|[0 - 1]
Idling Request |(Bit)|[0 - 1]
Engine Stop Request |(Bit)|[0 - 1]Only these changed.

The polling interval was 84 seconds for over 300 data points. For slowly changing data fields like temperature, no problem. But the interesting power fields change much faster than we can easily use. Verification and validation will require techniques closer to code breaking.

Bob Wilson