bterwilliger

Yaw sensor apparently = steering wheel position sensor. Replaced under warranty and not a glitch since. Fingers crossed...

inthehunt2

Note: Steering wheel sensor is not the same as the yaw sensor.

The steering wheel sensor measures the angular displacement of the steering wheel (how much the steering wheel is turned).

The yaw sensor is a sensor that measures the yaw of the car (if the car is pointing in the same direction it is actually going - ie. having the car in a drift condition is a yaw).

bterwilliger

Yes, that was evident from the names, but the tech and service manager were using them interchangeably. I'll find out exactly what they meant this week.

Maui

My '05 has 25,000 miles and I have never experienced any of these issues. Neither have I ever experienced DBS. It sounds pretty scary, tho.

Bill Curlee

If the tech and service manager cant figure out the difference between the two,,,,,,RUN FOREST,,,,,,RUN AS FAST AS YOU CAN, from that Stealership!

Sounds like they were EASTER EGGING PARTS to me. Bet ya, the problems are not solved!

BC

inthehunt2

supersecretsquirrel

Being the engineer that I am this sounds like a signal integrity problem. Let me elaborate...

The systems in question work because of high speed electronics. In electronic design when dealing with high speed data paths the issue of crosstalk is one that is always in the forefront of my mind because corrupted data is bad. Now this issue does not just apply to traces on a printed circuit board. It can be found in cable (or wire harness) designs.

Essentially what happens is that when a signal (a voltage) is sent down a wire (or trace) there is a resulting current which induces a magnetic field in that trace/wire. The faster that this signal is, the greater the current flow, and in turn the greater the magnetic field it generates. This is called crosstalk or electromagnetic coupling between signals

If the first trace/wire is close enough to another trace/wire, the magnetic field generated from a fast rising signal on the first trace/wire can induce a signal (voltage) that travels in the opposite direction.

How could this apply to the issue with the C6's? The ESC activates when it receives a signal telling it to do so. It would not surprise me if the cable/wire harness is the problem and not the connector that GM is states is the problem.

For instance, its a given that there is a constant state of signals being transmitted back and forth through that cable/wire harness. If the shielding of the one of those wires is weak enough or perhaps the magnetic field from another wire is strong enough a signal could be generated that makes it way to the ESC computer that causes it to activate the AH and TCS as has been described.

Until this is tested and verified it is just a theory; but from my experience as a electronics engineer and how the issue seems very much like issues I have dealt in high speed data systems that have required a redesign of cables and/or layout change of printed circuit boards, there might be something there.

inthehunt2

Not that I am trying to discount your theory, for it could be plausible. However, since every module is tested (or should have been) tested per CISPR-25 and a vehicle level EMC test is also performed in a full EMC chamber (or should have been), I tend to doubt that it is EMC related. That said however, I have seen instances during a VA/VE that small things like SMD caps were removed from a design and the module that once passed testing would create an issue in the vehicle.

Maybe you're on to something here. Question is, how can that be verified? Wish I had an IFR or something that I could use to test this theory out. Maybe this idea should be forwarded to the DOT for further investigation?

supersecretsquirrel

inthehunt, this is not noise. I am familiar with CISPR-25 and EMC testing and both are only concerned with thresholds due to radiated and conducted emissions (i.e. radio interference). Also those SMD caps you mentioned where most likely part of the output filter on a switching power supply and where there to reduce the output voltage ripple which can radiate and get into other parts of the design.

Testing for crosstalk issues can be done with the proper oscilloscope/logic analyzer with differential probes from someone like Tektronix. If you are wanting to simulate then a 2D field solver would be useful.

inthehunt2

I certainly don't want to turn this thread into and EMC debate. Let me clarify what I was trying to say. When RI testing is done, it is supposed to be done on every signal pin of every module in the vehicle. This RI testing is done from about 20 MHz to 2 GHz by current GM EMC standards. Thus spurious signals on input pins (one sort of source could be cross-muxing of signals/noise on harness wires) are tested to both free-field RI as well as tri-plate and 3-metre line tests. I was trying to suggest that these tests do a fair job of replicating cross-talk (although not perfect). Noise injection on input lines should also have been performed. The trick is to model this crosstalk as you discuss.

Again, I think your idea has merit. The trick is to come up with a way to test it in a way that can be reproduced. Forwarding this info to the DOT would be a great idea. Unfortunately, I do not have a way to test this at home easily and I am sure my company would not like me running testings like this (for personal use) on the company's "dime".

Feel free to PM me. We could discuss further if you'd like.

supersecretsquirrel

inthehunt you are correct this is not the forum an indepth discussion.

My only intent was to suggest that due to the fast rise times of the devices involved that crosstalk could be the issue.

PM on its way.

radiercks

inthehunt & supersecretsqurriel;

I suspect that you both are zeroing in on the problem. Having worked on some aeronautical instrumentation in the past, along with industrial automation products, any movement of a wiring harness can introduce problems. Considering '05 was a major model year change, I wonder if they just didn't life test this particular area enough.
This thread I stumbled upon now gives me pause in my current search for a 2005-06 convertible.

Thanks for your extensive discussion on this topic.

B-Vette

I own a C5 and was looking at upgrading to a late model C6 because of some mechanical issues, (not traction control related though - not yet any how). Might have to rethink that one!!

GregC6

I received a response from them, and they told me to do the tsb first, and if the problem still occurs get back in contact with them

vahpr

It appears asking GM whether the issue has been addressed for 09 models gets you nowhere - the cust svc robots will only assist with setting an appointment at a dealer to have it looked at, and would not help with the question of whether or not GM has done anything with the 09's to alleviate this. Nor would they provide any info on who I might talk to. The search goes on........

2006c6keller

Sounds very ? after going over the edge of a mountain road!

bterwilliger

Confirmed it was the yaw sensor they replaced. Issue has not reoccurred, but it may take a bit of time.

The data rate for these signals can't be anything more than a few hundred hertz, which at 12v (possibly 5v) means inductive crosstalk is highly unlikely. But as long as we are speculating, I’m going with a sensor issue, coupled with a loophole in the firmware (which regardless of sensor or signal issues, should absolutely not let anything like this happen).

supersecretsquirrel

bterwilliger,
data rate and system voltage has nothing to do with crosstalk (inductive or capacitive). Crosstalk is a function of signal rise time; more specifically the 10-90% window for signal rise and fall times.

Cross talk happens when the amplitude of the signal that is the aggressor is great enough to exceed noise difference between (a) the intentional signal observed at the receiver, and (b) the receiver threshold, measured over the setup and hold window.

Because the transistors used today have lower gate capacitance values, the devices will transition faster then they used to 20 years ago; even at 5V or 12V.

This faster transition is why termination resistors are needed to attenuate the amplitude of the signal before it is reflected back to the source. This combined with proper grounding and spacing will keep crosstalk from being an issue.

Just an FYI I have seen several 5V and 12V applications with data rates of only a few hundred kilohertz have cross talk issues.

BTW inthehunt2 and I are having a great discussion about this via PMs. You are welcome to join in if you are interested.

bterwilliger

Sorry for the OT, but the relationship between voltage and inductive crosstalk is not up for disagreement unless we are going to deny electric field theory (take that up with Faraday). This is fundamental to electronics (not least, transformers).

supersecretsquirrel

Im not denying the relationship between voltage and crosstalk (inductive or capacitive). Im saying that your posting about crosstalk being unlikely due to data rates being only a few hundred kilohertz or the voltage seen on the signal lines being only 5V or 12V is wrong.

EM theory clearly does state (and I can’t disagree) that a voltage on a wire causes a current to flow due to the resistive properties of the wire. You also have capacitance due to two conducting bodies being at different potentials (the wire and the voltage source). These two are ordinary inductance and capacitance.

There are two others, mutual inductance and capacitance. It’s this mutual inductance and capacitance that is the issue; not data rate or voltage as you are describing it.

Whenever you have two circuits you have mutual capacitance (i.e. parasitic capacitance). The voltage on one circuit creates an E Field that affects the other.
Mutual inductance exists whenever there are two current carrying loops (wires or traces that are near each other). The current in one creates a magnetic field that affects the other.

What you haven’t stated (or grasped) is that the rate of change of the voltage step height is a function of the driving waveform time (i.e. rise time of the signal). This time value allows you to determine the mutual capacitive current that flows from point A to point B. This multiplied by the impedance of the receiving circuit and divided by the rise time of the signal (there is again rise time) allows you to determine cross talk voltage levels. The smaller the signal rise time of the signal being generated, the greater the cross talk voltage.

Mutual inductance injects a noise value into the receiving circuit that is proportional to the rate of change of current. Fast changes in current in circuit A can induce large voltages in circuit B. Changes in current are a function of signal rise time.
As with mutual capacitance, the inductive cross talk is a function of signal rise time. Dividing this rate of change of voltage by the signal rise time multiplied by the wire resistance allows you to determine inductive cross talk levels.

Interference levels as low as say 2% can contribute as much as 500mV of noise in TTL system. These levels can be generated by both mutual capacitance and mutual inductance (with mutual inductance being the bigger offender) and they can sum together.

So tell me how a 5V TTL signal could not be a source of cross talk?

Again I invite you to the discussion on this with inthehunt2 and I via PMs.