An Ex Rocket Man's Take On It: March 2026

The topic here is how experimental experiences, with ramjet ablative combustor liners, relate to the unexpected and unpredicted entry heat shield problems seen on the Artemis-1 test flight. There actually is some correspondence, the overlap suggesting a “fix”. What might actually be done to correct the heat shield problem is presented here.

The author is qualified to speak on this topic, having spent 20 years in aerospace defense new product development work, most of it in solid rockets and ramjet propulsion. He is no narrow specialist, as the next 20 years in teaching (at all levels), in civil engineering, and in aviation work demonstrates. He is now long-retired, but occasionally consults.

The Problem

The Artemis-1 heat shield partially failed by unexpected and unpredicted cratering, but did not have a burn-through. It was constructed of Avcoat tiles bonded to the capsule structure. Chunks of the charring heat shield spalled out, creating alarmingly-deep craters in it. The same Avcoat material was used as the Orion EFT-1 flight test heat shield, and on the Apollo capsule years ago. The difference with Artemis-1 was that the Avcoat of those earlier heat shields was hand-gunned into the numerous cells of a fiberglass-type hex attached directly to the capsule structure. The change to bonded tiles was intended to greatly reduce the labor cost of the hand-gunning, and also reduce its inherent variability.

The bonding of the tiles to the Artemis-1 capsule, and the gaps between them, are not in question here. Those performed quite well. Nor is the cost reduction in question, it is quite real, obtained by casting blocks of Avcoat that were final-machined to tile shapes.

The only real difference is the question explored here: those blocks of Avcoat had no reinforcing hex in them!

A Place to Look is Outside the Organization!

One needs a “fresh pair (or pairs) of eyes” to see past the thinking “ruts”! The problem with that is that one may not really know the people who are from outside. Therefore, one tends to trust such outsiders less! Another aspect of that is the “not invented here” prejudice! One must strongly resist that!

Ramjet Ablatives Is a Related Experience

Unlike solid rockets, ramjets have very long burn times. And unlike rocket motor free volumes out near their case diameters, ramjets have high fluid shear forces all along their combustor walls! The ramjet has long proven to have the more challenging environment for ablative protection. The reinforced-rubber insulations used in rocket cases are quite inadequate for ramjet combustors!

The reentry heat shield sees hotter local effective temperatures, experiences lower surface pressures (because the atmosphere is so thin at entry peak heating altitudes), and quite likely sees fluid shear forces along the heat shield surface that are on the same order as those of the ramjet combustor. That last seems likely because both are subsonic (behind the bow shock for the capsule, with higher sound speeds but lower densities).

This author ran some experimental ablative candidates through short-burn, full-scale tests in a combustor of the size for a ramjet replacement of the AIM-120 AMRAAM motor. In the process of understanding the results, he identified some 5 key issues to consider, as are indicated by notes in red in Figure 1 below. (All figures are at the end of this article.) These issues were porosity for gas release, the polymer pyrolysis temperature, the “glue” effect of viscous melt materials, the “reinforcing aggregate” properties of particulate solids, and the reinforcing effects of fibers reaching into (or within) the char layer.

The baseline ramjet insulator, Dow Corning’s DC 93-104, did the best, as expected. The shortfalls of the experimental alternatives were understood, once those issues were all identified. These issues were all addressed in the formulation of the DC 93-104 (as indicated by the blue notes), but were not all addressed in the other materials. Those included a Japanese “equivalent” to DC 93-104, and two pre-preg cloth layups.

These were short-burn tests, all under 1 minute long. It was already known that many ramjet applications needed longer burn times than that, but that thicker layers of the ablative were simply not a design option, because of the enclosed volume reductions. The solution to the burn time problem came from “in-house”, not Dow Corning, and is illustrated in Figure 2. As used in the 20-inch diameter ASALM-PTV combustor at the same liner thickness, the retention ribbons held the char in place as an insulator, for up to 15 minute burns! While not as “good” an insulator as the virgin material, the retained char was “good enough” to essentially achieve acceptable steady-state results in ground tests and in flight tests!

That kinked-ribbon retention feature was not included in any of the experimental insulation tests discussed here, primarily because all the burn times were under 1 minute! Not being present in these tests, it cannot obscure or impact the fundamental ablation phenomena behind the results obtained! Furthermore, the two pre-preg cloth alternatives had to be wrapped onto an inflatable-bladder mandrel for installation, more like rocket motor insulations. They could not use the kinked-ribbon retention feature anyway! The hope was that the cloth fabric would provide enough retention. But it did not.

Applying This Knowledge to Ramjet Combustors

The designer must deal with all 5 of the issues in the ramjet environment, but only 1 of the 5 issues actually dominated the picture, as indicated by the gold stars in Figure 3. This dominant issue proved to be randomly-oriented high-temperature-capable fibers tying the char to the virgin beneath. The need for enough porosity to release the pyrolysis gases through the char was not much of an issue in the thin layer designs we were using. In heat shields that are thicker, it might be more important! There is some impact of the presence (or not) of sticky, viscous melt on the surface, to prevent the more rapid erosion of small particles by the fluid shear.

If ramjet combustor design was all that we were considering, then the loss of some char chunks as seen with the Japanese material, might be countered either by the long-burn kinked-ribbon retention feature, or by adjusting the formulation to include more carbon fiber, or both. Even so, it provided useful 1 minute burn time, even with some chunk loss.

Applying This Knowledge to Heat Shields

There is definitely overlap of the physics identified between the ramjet combustor and the heat shield applications! While heat shields do not deal with all the same issues, the porosity and char loss do indeed overlap a little. But the need for radial fiber reinforcement would seem to dominate both applications, preventing as it does the loss of chunks of char! This is illustrated for ramjets in Figure 4.

Quite apparently the real mistake made with the Artemis-2 bonded-tile heat shield design, was deleting the reinforcing hex from those bonded tiles! And that says the most important thing NASA could do is to look for ways and means to put that reinforcing hex back into those very bonded tiles! Preferably doing it without doing the labor-intensive and expensive hand-gunning of the Avcoat into each individual hex cell!

So, the main question here is: can that really be done?

There Really Is a Practical Fix

The answer is “yes”!

But to even consider it, one must avoid the trap of “either/or” thinking! That trap considers only tiles without hex versus doing it hand-gunned completely like Apollo and Orion EFT-1. That thinking trap is illustrated in Figure 5, along with this author’s way out of that trap: use an extrusion press to load all the cells in the hex with Avcoat, all at once!

That way, one still uses bonded tiles, and avoids all the hand-gunning labor! And the tiles with hex in them should cease shedding chunks of char, no matter how the entry is flown, skip or not! Plus, the high micro-balloon content that leads to low densities, but also to high mix viscosity difficulties, is something the extrusion press can handle. All that is needed is the right tile mold tooling to hold the hex in place on the press. That is just tooling design. Plus, the inherent human labor variability is eliminated.

Such a solution is unavailable to people hampered by the bad habit of “either/or” thinking. The author gave this solution to NASA more than a year ago! But they so very clearly did not use it, or even contact him about it!

Final Comments

These ramjet results and their application to heat shields are as much “engineering art” as they are anything! One will not find this in published reports or academic texts. A lot of this is qualitative, not quantitative, and inherently so! The difference between “engineering art” and “engineering science” is explained in Figure 6. Note the large fraction of the necessary knowledge that is “art”, especially in development work!

Note also that the engineering art is passed-on, one-on-one, on-the-job, from the old hands to the “newbies”. That is, it is passed-on if, and only if, there are any old hands on the staff to do that teaching job! Too many organizations prefer to hire only “newbies” that they can under-pay, instead of retaining the “old hands”, excusing this with “because they are too expensive”. But it loses you half (or more) of your essential knowledge!

The current status is this:

It is too late to fix the Artemis-2 heat shield, that rocket is on the pad to launch!

Artemis-2 is forced to risk worse or deeper cratering with a crew

Probability of a fatal burn-through is not zero !

Had almost 2 years, so why not? Schedule and money, same as Challenger!

Must fix the heat shields of Artemis-3-on, to avoid a fatal failure, sooner or later!

Must not repeat Challenger mistake: must value lives above schedule & money

Trust results (even qualitative), not the same sources that failed to predict cratering

What one has to learn and remember is this very harsh lesson: There is nothing as expensive as a dead crew! Especially one dead from a bad management decision!

In different ways, this bad decision thing is really is the root cause of both the Challenger and the Columbia disasters! “Avoid that mistake like plague” is the best advice this author can give you, even if it means listening to outsiders, or overcoming “not invented here” prejudice! Please feel free contact him for more details. He has test data and lots of photos from those old ramjet tests.