Roof Rack Functional Failure: An Analysis

Even before I had improved my micro tower, shared here, I recognized that something was not quite right about my roof rack. I couldn’t find any damage, but could see one tiny sign that the rear crossbar was not quite the way it was when I mounted it months ago. The rack was still sturdy; so, I continued using it for my micro tower and other antennas. Later, I opted to strip everything off my roof so that the car would blend-in while parked unattended for a week. I also planned an inspection. That’s when I discovered a hidden failure.

I’ve shared a video version of this article here. This article is essentially a script for the video. However, the video adds some visual aids which may make things easier to follow.

Before I start, I want to express a few disclaimers before the keyboard warriors launch their rants about my foolishness: First, I knew from the start that putting a micro tower on my roof rack was exposing the rack to forces which may be beyond its design limitation. Sure, the setup weighs ~1/3 less than the rack’s weight limit, but the taller components leverage greater forces against the mounting points than the typical rooftop load. Second, I went into this experiment with the expectation that the crossbar’s mounting screws might be the system’s weakest link and planned to watch them carefully. It turns out that my prediction was close, but not exact. I’ll explain later. Finally, and perhaps most important, I’m not an engineer! I think like one and can tend to over analyze things. However, I do not have the training to calculate the forces applied to certain points along the rack system. I may use a few engineering terms properly, but that has more to do with some of the engineering-like work that I have done in my career, rather than me being a trained engineer.

One more thing that I’d like to clarify is that this was not a structural failure; not completely, anyway. Instead, this was a “functional failure.” What does that mean? When a product is in its design phase, a list of “functions” is created so that engineers know what kind of design they should consider. I don’t know what engineers who design roof racks list for functions; so, I’ll provide some notional examples. Some of the functions that a roof rack might need to perform could include:

• Carry 165 lbs without detaching from vehicle.
• Carry 165 lbs without damaging vehicle.
• Carry 165 lbs without deformation of rack components.

My rack was able to carry the 92-lb tower without a failure in the first two functions; however, a failure occurred in the third notional requirement. With that shared, let’s dive-in and see what happened. This will be a long story since I’m presenting a lot of info. My roof rack is an aftermarket setup by Yakima. The brand name is not important. The key point is that the rack is made to attach to a “naked roof,” or a roof without factory roof rails. Nearly all aftermarket roof racks of this type are similar; so, I’m not badmouthing Yakima here. In nearly any setup, the crossbars hold the cargo and are bolted to two pairs of towers, which support the crossbars and provide an interface with the car’s roof in the form of soft pads. Each tower is clamped to the roof by a clip which pulls downward from the door jamb.

I had two concerns for weak points in this system. The most obvious point would seem to be the clamps at each door jamb. I’ve seen examples of clips that had slipped outward when not tightened properly (or perhaps on an overloaded rack). My Yakima Baseline Towers feature tension screws that prevent the clips from slipping. It would seem that the only way for the clips to slip is if they are somehow bent. Still, I inspect the clips for proper position from time to time, especially the driver’s door since I see it every time I enter the car.

Next, a hidden weak point, is the internal mounting screws which fasten the crossbar to each tower. To my surprise, each tower has just one M4 screw to serve this purpose. I knew this from the beginning. Yes, there is some metal in the assembly to distribute the clamping force, but it still got my attention. By way of comparison, a factory roof rail on a typical wagon, SUV, or minivan features TWO M5 or M6 screws at each corner and sometimes one or two more in the middle of the rail, all of which are usually attached through a structural member instead of just sheet metal. That presents a much more secure mount!

I researched the tensile strength of an M4 screw and learned most screws test at about 3500 lbs! Using a “10% rule” as a guide, that could mean that an M4 screw’s working load is around 350 lbs.  That knowledge helped my confidence that the screws would hold my load, but I still kept an eye on things. Eventually, I noticed a 1-2 mm gap between the rear towers and crossbar. I think the casual user might not have noticed that for a very long time. Remember, I’ve been watching for problems! This photo shows how I was able to slip a snap ring between the tower and crossbar when there had previously been no room. I inspected more closely when I removed the rack and discovered that the gap was because the tower screws no longer achieved clamping force with the crossbar.

All four towers were able to slide freely along the crossbar despite being tightened to specs. I think it was holding decently when tensioned to the car since I could not feel any play or rack motion. My initial thought was that the crossbar T-slots might have deformed at the tower mating points. I removed the crossbar from the tower and didn’t see any apparent deformation of the crossbar. Instead, I saw a bulge in the mounting flange within the tower. The bulges in each tower were similar, suggesting that they may have been designed that way and that one of several tree limb strikes was not the cause. I’d have to await new towers to be certain of the failure mode.

The new towers arrived and I confirmed that the old mounting faces are indeed deformed. The reason I’m dismissing the tree limb strikes as the cause is because an impact from any obstacle would pull upward on the front mounts (tension) and compress the rear mounts. Compressing the towers is impossible due to the large mating faces within the crossbars. Compression forces would likely transmit to the roof. Instead, all four towers showed evidence of tension damage. How did this happen? Sure, a tree limb strike could over-tension the front mounts. But another likely cause, especially for the rear mounts, is the applied forces from braking, acceleration, and corning. These forces, although much smaller than a tree limb impact, have been applied over the course of thousands of miles and easily can account for the amount of time it took for the damage to reveal itself.

Then there are the forces that we don’t think about because they’re not impacts at all. Think of when you drive over a railroad track or other uneven surface in a construction zone. The car pitches and rolls, sometimes uncomfortably, but still only a few degrees, and we continue driving because, “Hey, the car is suspended.” But the top of the micro tower sees more motion. An inch or two in the car could be six inches at the top of the tower! The motion then swiftly reverses to return to a normal state. I think that rapid motion, perhaps exceeding 1G, could exert a lot of force at the base and rack towers. It’s a simple matter of leverage and applied force. How much? I’m not sure. The heaviest part of the structure sits just six inches over the roof and right on top of the crossbars. Next, is the weight of the 38-lb tower distributed over 38 inches. The 9-lb mast is 76 inches high, but that weight is distributed, too. Each of the Yagis weight about 4 lbs and sit and different heights along the mast.

After speaking with Andrea, K2EZ, I realized that the upward force applied to the front crossbar can be computed using leverage calculations. First, I assumed a minimum force of 1G to be exerted along the mast during maneuvers. Next, I estimated the weight of each 1-ft section of the tower, shown in this image. From there, I took each height along the tower and multiplied it by the crossbar spread. For example, the ratio at one foot high would be 1:2.67, or 0.376, and the ratio at the top of the tower would be 7:267, or 2.622. I then took the ratio and multiplied it by the weight of the applicable section of tower to see its upward force on the front crossbar. For example;

(1÷2.67) x 17.5 = 2.25 lbs and
(7÷2.67) x 5.5 = 18.41 lbs

Once I calculated the upward force at each level of the tower, I then added all of the values to estimate the total upward force on the front crossbar at 1G, which is ~80 lbs (see photo above). Divide that number by two to calculate the force on each crossbar tower. Keep in mind that this calculation is for force applied in only one direction and at 1G. Obviously, forces may be lower on a smooth road, but they could be two or three times that over road irregularities! My calculations do not take wind loading into consideration. For the record, Andrea calculated that I’d have to be traveling at over 100 mph before the tower would feel 1G of force from the wind alone.

How can 40 lbs be enough to deform my rack tower’s mating faces? Well, either the forces are greater than 1G or the rack’s towers are designed to tolerate less force. For example, placing the same 92 lbs on the rack in the form of a rooftop cargo box brings the load much lower and within the parameters that the rack engineers had in mind when they specified rack tower mating flanges that are only 1.5-2.0 mm thick. Assuming that the weight within my 14-inch tall cargo box is centered six inches over the crossbar results in the following: (0.5÷2.67) * 92 = 17.23 lbs

Divide that by two and you get 8.6 lbs per crossbar tower! Even with a full 165-lb load (box and cargo combined), the force increases to just 30 lbs, or 15 lbs per crossbar tower. The difference is pretty amazing! I’d suggest that I might have avoided my problems if I had done these calculations BEFORE putting a tower on my roof. However, I wouldn’t have believed that 40 lbs of force was enough to deform my rack tower’s internal mating surfaces. Now I know.

Engineers are pretty smart and probably specified a grade and thickness of mating surface that would allow it to bend rather than break or split. The bending of rack tower mating surface likely prevented a catastrophic failure that could have resulted in cargo being scattered all over the highway or into the windshields of other drivers. The “soft failure” of my rack also limited the damage to the rack itself rather than damaging my roof. My roof is still perfect!

My guy lines are an element that I intentionally left out of this equation. You see, I use the guy lines as mild stabilizers and retention in the event of a structural failure. Note: This includes the case of a rack tower’s mounting flange splitting in two. As a result, they are not tensioned to the point of seizing all tower motion when the car is moving. Attaching the guys to the rack would do nothing to limit the lateral forces that are applied to the towers. Guying the tower to the roof, as I have done with my Seasuckers, could help if I was willing to tension them to a high degree. However, I do not want my roof deformed for the sake of routine stabilization. My guy lines were taught, but not tensioned enough to sing when plucked.

What is my next step? I’ve already purchased new crossbar towers so that I could diagnose my problem. I could put everything back on the car and keep the party going. However, that’s just another $220 mistake waiting to happen. Instead, I’ll take my WORN, but undamaged, crossbars and bolt them directly to the side rails of my “baby trailer.” I’ve been asked many times whether a trailer solution would be a better choice for my tower. Well? I’m going to find out!

Everything shown in this photo will bolt directly to the trailer. The main difference is that the overall tower height will be reduced by nearly two feet! That’s good for tree avoidance. The lower platform may also lend itself to a telescoping mast solution at some point. For now, I’ll work with my existing system on my trailer. My only concern about a tower on the trailer is with the way the trailer bounces violently over some obstacles. All of my comments about tower motion shared earlier certainly apply, perhaps to an even greater extent since the trailer’s movement can be more sudden. I can improve the ride by adding weight over the axle to preload the suspension. I often use either cinder blocks or jugs of water when transporting fragile cargo.

My other antennas, including the verticals and the horizontal loop module, will remain with the car and on new crossbars. Only the tower will ride on the trailer. Leaving the tower on the trailer provides two additional benefits: First, I now have the option of mounting my rooftop cargo box on the car while moving a contest tower. Second, and perhaps more of a benefit, I can assemble and erect the tower on my trailer at a leisurely pace without any impact on my daily driving, including if I need to enter a parking structure or drive my usual routes with low-hanging tree limbs. The trailer can sit in my driveway until the instant it’s time to take it on a contest rove. I may even store the tower this way and make it ready to deploy at a moment’s notice for other needs. Sure, I still need to investigate extending some feed lines and rotator controller cable. I think that will be easy!

Here’s a photo of the tower on my trailer for a test drive. I’ve shared two more below. I’m doing a minor overhaul of the trailer to include paint, a new wood deck, bearing maintenance, new tires, and experimenting with different rack setups. So, this is just a short-term test. I’ll share more photos of the trailer setup on the coming weeks.

What do you think? Am I nuts? Perhaps a bit foolhardy? Or am I onto something with this trailer setup? Watch a video about my baby trailer to see just how versatile I’ve made it. Being small, like my car, there’s almost no drawback to bringing it with me on a rove.

Always Adapting!

Scott