There was a time when replacing steering columns and steering racks on almost any car was as easy as falling out of a tree. Now, however, replacing steering racks can be a little trickier, but as with most repairs on modern cars, one DIY mechanic’s misfortune can create an even bigger problem for us- the professionals.
Thus, this article recounts the tale of a DIY steering rack replacement on a 2017 Holden Malibu that not only caused a no-start condition but also the likely unnecessary replacement of a brand new OEM steering rack before we discovered the real problem. This cautionary tale has several twists and turns, but before we get to specifics, let us state-
The customer that presented the Malibu to us (on the back of a recovery truck) was not exactly a regular customer, but we knew him well enough to feel somewhat sorry for him when he told us his story. Here is the short version of his tale-
According to this customer, his family needed a second vehicle and since his uncle “knew a bit about cars”, he approached said uncle to help him find a suitable vehicle. After some searching, the uncle recommended this particular, low-mileage privately owned Holden Malibu as an excellent choice since the aged previous owner was not driving much anymore, and he had in fact, not driven the car for more than a year due to ill health.
None of this is really relevant to this story, but what is relevant is the fact that the seller did mention that the car's steering seemed to need some attention. The previous owner was not exactly sure what the problem was, but he did say that the steering felt "a bit wishy-washy". According to the current owner, the seller explained that it felt as if something was missing from the steering system, or that something had become undone because the front wheels would not return to the dead-ahead position automatically upon exiting corners and sharp turns.
One had to actively steer the wheels back to the dead-ahead position, and if the road surface was not absolutely level, keeping the wheels pointing straight ahead could be difficult because the car could veer in any direction at any moment. Then, according to the current owner, two things happened. The first thing was that the seller knocked $500 off the selling price “to help with the repair costs”. The second thing was that the uncle volunteered to do the actual repairs, since “he had been working on his own cars for all his adult life”.
Blinded by his two strokes of good fortune, the current owner bought the car on the spot and asked the uncle to drive it home, which is where his problems started. Long story short; the uncle diagnosed a "sketchy" steering rack and offered to replace it, because he had replaced countless steering racks in his life, after all. This one would just be another in a long list of such replacements, and he would have it done before the end of the week.
In the end, though, replacing the steering rack took almost three weeks, but worse, now the car would not start, and nothing in the instrument cluster worked. Where the car ran perfectly before the steering gear replacement, the engine now only cranked, and no lights on the dashboard came on when the ignition was switched on.
The customer said much else besides but concluded his tale by asking if we would mind "taking a look" at the problem, because he felt he needed to get experts involved. We agreed, and here is-
Our initial inspection did not reveal any major issues with the replacement, in the sense that nothing was obviously missing, and/or broken. We also confirmed the no-start condition, as well as the fact that the instrument cluster was as dead as the proverbial doornail.
However, the bigger problem was that the steering gear was not new. It was used, which was almost certainly the root cause of both the no-start condition and the dead instrument cluster. This car was a Global A (Global Architecture) car, and therefore, the power steering control module, which is incorporated in the steering column, could not have been integrated with any other control modules. Moreover, there was no telling what other issues might have been caused by the non-integrateable power steering control module.
At that point, we knew that it was highly unlikely that the gateway module in the car would communicate with any scan tool in meaningful ways, so there was only one thing to do. We called the customer and told him that since this was a Global Architecture car, the only way to fix the problem was to remove the used steering gear and to have it replaced with a new, OEM part.
To say that the customer was extremely unhappy to hear this would be an understatement. Nonetheless, we were too busy to take on the steering gear replacement just then, but we did manage to calm the customer down somewhat by offering to program and calibrate the new power steering control module at a reduced rate after the replacement. We then arranged for the car to be returned to the customer and did not hear from him again for about three weeks.
Before we get to what happened next, we should perhaps say something about GM’s approach to vehicle security, which is embodied by a strategy known as-
In GM-speak, “Global Architecture” refers to a global or holistic approach to securing all the aspects of a vehicle, as opposed to improving the security of GM products sold around the globe.
This might come across as trite, but the truth is that GM was forced to rethink their approach to vehicle security after a couple of hackers demonstrated in the mid-2000s how easy it was to breach not only the security system of a Jeep, but also to break into the cell phone networks that supported the vehicle's telematics systems.
The hackers used a standard laptop computer to gain remote access to the vehicle, and they could not only start it at will; they could also control the steering and brake systems, disable several critical security systems, and deactivate or reroute telematics communications.
Thus, in 2010/11, GM engineers responded to the threat by devising a “global” security strategy that secured all, or most of a vehicle’s control modules, as opposed to improving the security of only the modules that controlled the anti-theft/security systems. While not all GM products are thus protected, even in 2021, most GM products sold around the world are fully protected by Global Architecture technology. Here is a list of the most commonly protected control modules on most GM products-
As a practical matter, Global A technology has three “legs” that support the overall security strategy. Here is the short version of how it works-
Security code
This code works in much the same way as the security codes on all other vehicle brands. It is generated at the assembly plant, and it is linked to the vehicle's VIN for the life of the vehicle. Its function is to allow or prohibit certain vehicle security functions, such as enabling the starting circuit if a matching key is used, or disabling the starting circuits if the wrong key is used.
This code is the first line of defence against unauthorised reprogramming of control modules and/or ignition keys.
Environmental ID
In its simplest form, this ID is a code that is programmed into each module. In practice, all the codes in all modules must match a "Master ID" as a means to prevent the swapping of control modules to steal a vehicle or to expedite some repairs.
To make this system work, the Vehicle Theft Deterrent Feature scans all control modules and then matches the environmental ID's of all protected control modules against each other, and the “Master ID” to detect module replacements. If just one mismatch is found, the Vehicle Theft Deterrent Feature will deactivate the starting or ignition circuits, as well as other critical control modules that manage the vehicle’s operation.
Seed and Key Matching
The purpose of this measure is to prevent the unauthorised reprogramming of protected control modules outside of either the assembly plant or an approved environment such as a dealership or authorised service centre.
In practice, each protected control module is programmed with a unique “seed” value and a corresponding “key” value that is stored in a kind of hardened memory. The seed value can be accessed by an authorised programming tool, but programming will only be initiated if the seed value matches the key value. As a practical matter, the key value can only be obtained by a) registering as a vehicle security professional on a site like AutoAuth and b), by purchasing the required software from an OEM approved resource.
Note though that no scan or reprogramming tool can extract any protected control modules’ seed value, and/or display a matching key value.
Taken together, the three legs of Global Architecture makes it extremely difficult, if not always impossible to steal a protected vehicle by swapping out control modules. However, from our perspective as technicians, Global Architecture also makes it impossible to do some repairs unless we replace some control modules.
In practice, it is not possible to replace suspect control modules on Global Architecture vehicles with modules that had previously been programmed. According to official GM sources, replacing any protected control module with anything other than a new, previously un-programmed module will not only damage the replacement module; it can also damage multiple other control modules on the vehicle. At the very least, attempting to bypass Global Architecture technologies will produce one or more of the following symptoms-
Exactly which symptoms appear depends in large part on which module had been replaced. Moreover, since GM is continually expanding or changing the configuration of Global Architecture technology on different vehicles and model ranges, it is almost impossible to keep up with how certain vehicles are protected. In practice, the model, the market(s) in which the vehicle is sold, the production year, the trim level, and the engine fitted all play a role in the actual topology of the Global Architecture on any given GM product.
Thus, the best way to avoid complications with Global Architecture vehicles is simply to never swap control modules between vehicles, which brings us to-
When we next saw the Malibu, it was again on the back of a recovery truck. It was again accompanied by its owner, who had another long story to tell. Here is the shortest possible version-
According to the owner, his uncle had replaced the steering gear with an OEM version, and then took the vehicle to a local GM dealership to have the required programming done. However, for reasons that were not, and still are not clear, the dealership could not get the programming to “stick”, so as a gesture of goodwill, they gave the Malibu’s owner another steering rack, based on their assumption that the new steering gear in the car was somehow defective.
We found this highly unusual, if not downright weird, but based on our experience, this would not have been the first time brand new, out-of-the-box parts were defective, so we did not pursue the point.
Long story shorter; the uncle removed the defective new steering rack, installed the new one, but taking the GM dealership's advice to heart, he did not connect any wiring or plug in any electrical connectors. The dealership told the uncle they would do that when the car comes in for programming, but instead of taking it back to the dealer, he had decided that we should do the programming and wheel alignment instead because as he pointed out, we were likely to be a little cheaper than the dealership.
Our first inclination was to refuse the job, but since this customer was a semi-regular, we agreed. How hard can it be to program and calibrate a new power steering system, after all? We had a new dedicated laptop running GM’s GDS 2 (Global Diagnostics 2) made specifically for Global Architecture vehicles through a new pass-through device, and we had an account to use GM’s SPS (Service Programming Service). Our resident diagnostician volunteered for the job, so we were all set, which brings us to-
Image source: https://www.vehicleservicepros.com/service-repair/undercar/article/21218366/rack-and-pinion-redo#&gid=1&pid=5
The diagnostician began the process by inspecting all the connectors that plugged into the power steering gear. All seemed perfect, and all locked securely into place.
Next, he verified the battery voltage, which was within specs but just to be on the safe side, he connected a clean power supply to prevent voltage or current issues. Next, he connected the pass-through device to the DLC, logged into our SPS account, found the software he wanted, and began the download.
However, the image above shows what happened shortly afterwards. The programming aborted for reasons that were not exactly clear, so we started looking for answers online.
Three hours later, we found that nobody knew why programming procedures on Global Architecture vehicles sometimes aborted, and not even official GM sources could help us. Official GM sources did say, though, that restarting the process sometimes fixed the problem. One source also said that on some Global Architecture vehicles, the GDS 2 (Global Diagnostics 2) scan tool/ SPS tool combinations might indicate a programming failure, when what it meant to say was that the calibration process had failed or aborted.
If you are new to Global Architecture vehicles, you may or may not know that the calibration of a control module or system on such a vehicle happens separately from the programming process. Thus, armed with the knowledge that some scan tools can sometimes report the wrong thing, we repeated the programming, and this time, it completed successfully.
The next thing was calibrating the power steering control module. This involved (among many other steps outlined on the SPS tool), turning the steering wheel from full lock to opposite full lock several times so the system could “learn” both endpoints relative to the dead-ahead position. The control module also needed to learn the characteristics of both the power steering motor and the steering torque sensor. Here is why-
We mentioned elsewhere that the calibration process involved a proverbial laundry list of steps. Well, we followed the process exactly: in fact, we followed it exactly multiple times, but every time we did, the calibration failed.
Out of ideas, we briefly considered referring the car to a dealership, but just then, and shortly before closing time, the diagnostician thought of something worth pursuing. He thought it might be possible to retrieve fault codes and/or live data from the power steering control module to see if (and how) the steering angle sensor responded to steering inputs.
The results were not encouraging. For one thing, he found a new trouble code: C0460-4B “Steering Wheel Angle Sensor Calibration Not Learned”. Worse, though, live data showed that while the steering torque sensor responded to steering inputs in a manner that seemed appropriate, the steering angle sensor did not respond at all; it remained at “0”, regardless of all steering inputs.
This problem was new to us, so the first thing the next morning we purchased a wiring diagram for the vehicle. Our thinking was that an abnormal resistance or some other issue might be preventing the calibration process from completing successfully. So armed with a wiring diagram we traced the relevant wires, which consisted of one power feed, one ground wire, and a few CAN lines.
We verified that both the power supply and ground circuits were good, and since there were no CAN-related fault codes present, we assumed that the implicated CAN lines were all good. Nonetheless, it was during the process of checking the wiring that the diagnostician noticed that the diagram did not show a steering angle sensor. We double-checked and sure enough, there was no steering angle sensor indicated on the diagram.
This problem was also new to us, because as far as we were aware, the stability control, traction control, and autonomous emergency braking, among other safety systems, all used input data from the steering angle sensor to work. We started another online search for answers, and during the late afternoon of the second day, we learned that this Malibu did in fact, not have a steering angle sensor, which begged this question-
It turned out the power steering control module infers the steering angle by comparing data from the steering torque sensor to input data from the power steering drive motor.
In practice, the steering input rate is derived from the steering torque sensor, which is then integrated with the rotation of the drive motor, which is measured in degrees of rotation to either side of the dead-ahead position, which is referenced to the vehicles’ thrust line.
In this case, we knew that the wheel alignment was a factor, but because the front wheels pointed straight ahead when the steering wheel was centred, we assumed that the steering rack had been centred. It turned out that we were wrong about that, which brings us to-
At that point, we needed to free up the lift the Malibu was on. This involved a sharp turn to the left but as we pushed the car through the turn and into an unoccupied bay, we heard a loud grating noise from the left front wheel. Upon investigation, we saw that the inside of the wheel was rubbing against the bodywork.
Turning the steering wheel so the front wheels pointed straight ahead, we turned the steering wheel first to the right, and then to the left. We found that we could turn the steering wheel through more than 2.5 complete revolutions to the left, but less than 2 full revolutions to the right, which, somewhat belatedly, put the problem into perspective for us.
This meant that we had made the wrong assumption because the steering rack was demonstrably not centred. It also made us look foolish, because we had made the most stupid mistake of all; we assumed something without verifying it.
We can gloss over who said what to who after this discovery, beyond saying that everybody involved was equally to blame for wasting two days on this job. Nonetheless, we put the Malibu on the wheel alignment rack, corrected the alignment problem, and repeated the calibration process.
This time, the calibration went off without a hitch, and the subsequent test drive showed no weird steering behaviour. The Malibu tracked straight ahead without any issues, the “wishy-washy” feel to the steering was gone, and there were no new fault codes, which leaves us with this-
Looking back on this episode, it was easy to see why we could not control the power steering control module. The module simply counted the number of revulsions of the power assist motor in both directions, and when the numbers did not match it aborted the calibration process.
The system was working as designed, but the question of why the dealership that declared the first new steering rack to be defective remains unanswered. Maybe the steering rack was really defective, but then again, maybe they were also tripped up by a wrong assumption. We will likely never know for sure.
