The basic primary and secondary structures for Godiva have been detailed in the previous two issues of Race Magazine. However I have received quite a few questions of how the honeycomb panels actually attach to the chassis and since there seemed to be such interest I thought that this issue should document my research and discussions so far.
If you look at the panel attachment image below you will see that one of the panel skins has been removed so that the remaining skin overlaps the square section steel tube (SHS) of the chassis. This skin panel has to have its honeycomb-core adhesive removed and the aluminium prepared to the specifications recommended by the adhesive supplier, as does the steel of the spaceframe. The inner skin is attached to the SHS using a 90 degree aluminium extrusion, which is also prepared as required for the adhesive. The aluminium is then riveted to the chassis as described, as a mechanical fixture is will assist joint integrity in an accident.
To ensure that this was a reasonable method of panel attachment I made contact with Robert Chase of Chase Competition Engineering. Chase Competition Engineering is one of the manufacturers of the Daytona Prototypes that use honeycomb panels in their chassis construction. Robert reported that they attach the panels in a similar manner, with the exception being that they use a 90 degree carbon fibre angle, instead of the aluminium extrusion that I am intending to use. Robert Chase also noted that some of the panels are not bonded in as described, but instead use mechanical fixtures so that they are removable (presumably for servicing).
Robert reported that pre and post testing of the chassis showed a 10% increase in the chassis stiffness when using the honeycomb panels. The actual chassis torsional stiffness reported as 9800ft/lb per degree of twist, or in metric terms 13,300NM per degree. According to Robert, each degree of flex equates to 0.4mm of movement from front to rear, which is not very much when you really consider it. Godiva’s chassis has considerably more triangulation than the Daytona chassis and the skin will be quite a bit thicker on the honeycomb panels, so I hope that Godiva will achieve a better torsional performance.
Robert went on to say “we feel that the chassis has no disadvantages, they are extremely safe, inexpensive and easy to repair, and cost effective to build”
However before I pat myself on the back for making a good decision in selecting the panels and hybrid chassis, there are a number of issues we need to recognise when we attach aluminium to steel.
The first issue is that we cannot weld the dissimilar metals together, so we must use mechanical fasteners (rivets etc) and adhesives to perform the same function and we need to select the most appropriate of both.
Secondly they have different expansion rates with changes in temperature.
Finally there will be an electrolytic reaction between the metals if they come into contact with each other.
Adhesives for the bonding of metals have improved enormously in the past 20 years. For those of you reading this and you are somewhat sceptical of the reliability of adhesives for joining metal to metal, remember that all airliners in the world are effectively glued together! Yes I know that some tragically do not continue to fly, but it is not due to a glued joint failure!
Although the actual subject of adhesive use for panel attachment to a spaceframe is a topic in itself, I shall address some of the issues as I see them.
In essence we want the adhesive to:
- Adhere very well to the mild steel of the chassis and to the aluminium of the honeycomb panels. There is obviously no point in using an adhesive that does not adhere well to the substrates (metals being joined), though it is understood that some adhesives may require special metal preparation for optimum performance. This is acceptable if the metal preparation is not expensive or onerous.
- Not degrade with vibration/impact i.e. it must be a ‘toughened’ adhesive. A toughened adhesive will be less likely to degrade in performance with general running stresses and in addition it will withstand sudden shock loads, such as might be experienced in an accident. Predictable structural performance is never more important than when considering accidents!
- Have reasonable retained performance with elevated temperatures of up to 120 degrees Celsius. We do not want reduced performance on a hot day!
- Be as stiff as possible. If the joint is flexible then the panels will not contribute much to the structural stiffness of the chassis. In addition, flexibility of the glue will mean added load on the rivets. Initially these will work well enough, but they will decrease in effectiveness and the stiffness of the chassis will obviously degrade over time and possibly affect the handling of the car. Since I expect that the handling will take some time and money to achieve, this degrading of chassis stiffness is best avoided!
- Have reasonable chemical resistance. This is very useful, as we do not want the structure of the car to be adversely affected from an oil/fuel/brake fluid spill or fumes. Of course we can protect the adhesive joint in various ways, but it would be better if we can find an adhesive that does not need the protection.
- Minimise the galvanic reaction between the steel of the chassis and the aluminium of the honeycomb panels.
- Be able to tolerate a reasonable gap between the substrates. A gap tolerant adhesive will be able to bridge a gap of 1-2mm, so that the joint integrity is not overly compromised by a surface irregularity or a very minor tolerance error.
- Cure fully at room temperature and in a controllable manner. We do not want an adhesive which has to be cured at an elevated temperature for an extended time to achieve an acceptable performance. Granted, such adhesives are often the best performing, but their use in Godiva would create too many problems in initial construction and also in repair. In addition to this, we want to be able to have enough time to assemble the parts accurately before the adhesive cures.
As soon as we say what the substrates (the materials to be glued) are, the number of adhesives available to us decreases markedly and this is especially so when we add in our ‘desirable characteristics’ as above.
As Race Magazine will cover the specifics of suitable adhesives in much more depth in the future, we shall just outline the basic types of adhesives for now. The information below is summarising the research so far. Due to the massive number of adhesives available it is by no means comprehensive and you should always check with the adhesive manufacturers, as there will always be exceptions to the information summarised below.
The adhesives that are thought to be suitable are epoxies, structural acrylics and filled polyurethanes.
Room cure epoxies (so called because they cure fully at room temperature) are generally two part adhesives, which I am sure most readers would be familiar with them in some form, probably from using them around the home or in the workshop. They involve mixing two (often but not always equal) quantities of part A and part B, one of which is the hardener. The ratio of the two components is critical for good performance, as is correct surface preparation, with metals requiring different preparation as denoted by the adhesive supplier. Room cure epoxies are available in quite large quantities and some manufacturers can provide the adhesive in twin tube cartridges which have a mixing nozzle as part of the cartridge. This eliminates some of the error possible, in what is often a very messy and sticky process. From the research so far completed, room temperature epoxies can be very stiff, quite tough, have reasonable elevated temperature performance, generally good chemical resistance, and reasonable gap filling properties require specific metal preparation and they are generally electrically insulating. They are also expensive adhesives, require significant care in their use and are intolerant of poor application and surface preparation.
Structural acrylics are two or three part, toughened adhesives, where the adhesive is applied to one surface and the ‘initiator’ or catalyst is applied to the other. The parts are then assembled and the adhesive cures without the need for elevated temperatures. The structural acrylics are often very tolerant to long ‘open times’ where the adhesive and initiator are exposed to the air for longer periods prior to assembly. Structural acrylic adhesive is also more tolerant of limited surface preparation according to one supplier. Structural acrylics can have very good elevated temperature performance, are quite stiff, reasonable gap filling properties, generally electrically insulating and excellent impact resistance. However some structural acrylics can have limited chemical resistance and have very short working times (from 2-5 minutes), which could have expensive consequences if I make an error attaching a panel to the chassis! Structural acrylics are reportedly less expensive than epoxies, but I was unable to specifically compare them prior to print.
Polyurethanes have become amazingly popular in vehicle construction in the past 15 years for a very good reason, they are simply a very good adhesive for a wide range of purposes. They have become very popular in the motorsport arena for attaching a wide range of body panels to chassis and indeed I intend to use them for attaching the bodywork to Godiva. However we are considering them here for a more structural purpose as many people are using them for this purpose on clubman style cars.
Polyurethanes are single part adhesives, usually supplied in cartridges for ease of handling and application to the bond surfaces. They cure well at room temperatures, with the isocyanite base reacting to airborne moisture to effect a cure. Polyurethanes adhere well to both metals with limited surface preparation required; have excellent gap filling properties and excellent impact resistance. However they are not very stiff and this can be exacerbated by elevated temperatures. As noted above we want an adhesive which is as stiff as possible and in this respect polyurethane does not compare well to either structural acrylics or epoxies.
At this point in time epoxy adhesive has been chosen. It is fills all of the desired criteria and most importantly, those with much more knowledge than I (e.g. Chase engineering, Lola, Mallock, FSAE teams etc) have chosen epoxy adhesive to attach aluminium honeycomb panels to their chassis!
Mechanical Fasteners and Galvanic Corrosion
In a similar manner, the issue of what type of mechanical fastener/rivets is also of interest, though for the sake of brevity I will shorten this discussion to the type of rivet chosen and the rivet material. But let us also discuss why use rivets at all, particularly if I am indicating that the adhesive is so good!
Epoxy adhesive performs very well in shear (see pic.1) where the force applied to the joint is in the same plane as the joint itself. Epoxies (and many other adhesives) do not perform nearly as well in peel, where the force is applied perpendicular to the joint (see pic.2).
As with many things we can predict most of the forces the panel will experience in normal use up to a certain point, however given all of the variables, we cannot predict what would happen in an accident. We can assume that the panel will be subject to peel forces which are less than ideal for our adhesive joint.
Thus we use mechanical fasteners to give additional security to a critical structural element and to ensure that the peel forces are limited by the mechanical fasteners themselves.
Rivets for this purpose are therefore a safeguard and really must be used, but what type and what metal should they be made from?
As noted above, I will be joining two dissimilar metals that often corrode (galvanic corrosion) when they have contact with each other and are in the presence of an electrolyte (water etc). The basis for this reaction is explained in the Galvanic Reaction sidebar.
This galvanic reaction is a significant concern and something we need to control if at all possible, as we do not want our chassis gradually decaying before our eyes…or even worse, becoming less structurally secure and safe without us knowing about it!
We want to ensure that the most important elements that contribute to the chassis strength are the least affected by any possible galvanic corrosion that may occur. Naturally this means that we want the steel spaceframe to corrode the least, particularly as this incorporates our CAMS/FIA approved roll over protection system and is clearly the most important structural element!
With respect to the honeycomb panel attachment we certainly do not want the hundreds of rivets to corrode, as they are small and must maintain their structure to perform as intended. Corrosion will thus affect them quite quickly if we choose the wrong material and their periodic replacement would be time consuming and involve dismantling the majority of Godiva…best avoided.
As we are joining mild steel to aluminium we would be best to choose a stainless steel rivet and following on from accepted racecar practice (in particular the well written advice of Carroll Smith in his book Engineer to Win), I will choose a rivet that retain the forming anvil through the entire joint. This type of rivet is by far the strongest in shear when compared to other rivets that do not have this solid steel insert through the whole joint (see pic.4).
Stainless steel was chosen because it is far more ‘noble’ than the aluminium alloy used for the panel skins and thus our panels skins become the anode. The advantage with respect to these panels being the anode is that they are easy to inspect.
As you can see in the Galvanic Reaction sidebar, the greater the cathode metal contact as a ratio to the overall anode area, the greater the galvanic corrosion. In Godiva the area of contact between the stainless steel rivets and the panel skins will be very small in comparison to the overall panel surface area (a good area ratio). This means that we will have the ‘least undesirable’ arrangement and that the amount of galvanic corrosion should be manageable.
We do not want to use an attachment method that would result in the honeycomb core becoming an anode, as the core material is made from very thin aluminium and any corrosion would result in a significant loss of structural performance. There is little danger of the core becoming part of the galvanic reaction with the attachment method as described, as the honeycomb core is not in contact with the panel skins due to the adhesive that bonds both skins to the honeycomb. This is one reason why only the skins are attached to the steel of the chassis and we are not directly passing any structural load into the actual core material.
The type of aluminium skins used on the panels can also affect the amount of galvanic corrosion to a small amount, but I am again fortunate here in that the Ayres honeycomb panels are intended for a marine environment and thus are made with a corrosion resistant grade of alloy (5051/5251). It has been reported to me that some cars that used much more aerospace grades of alloy suffered extensive galvanic corrosion issues, though I cannot be sure that this was solely due to the alloy (6005) used for the panels skins or due to the attachment method.
The galvanic corrosion issue may also have some implications for the adhesive used as some adhesives are more conductive than others. I have been informed that to further toughen and stiffen some acrylic and polyurethane adhesives, the manufacturers add increasing amounts of different types of fillers. Should you ever decide to use such adhesives, it would be worth contacting the adhesive manufacturer to find out which of their products are the most suitable and which, if any, are unsuitable.