By Eben van Tonder
31 July 2026
I have been using Klebemasse, Salzstoß, Hautstoß and Sehnenstoß for a month now as the basic building blocks of sausage, ham and reformed bacon production, without starch, isolate or TVP. What follows are my observations, together with the literature that supports or explains them. Where the literature does not go as far as my observations, I say so plainly.
Not Just the Collagen
When you trim connective tissue off a primal, you never trim connective tissue alone. Some meat comes with it, because silverskin and epimysium sit directly against the muscle fibre and cannot be lifted away without taking a layer of muscle tissue too. So when salt is added to this trim, it is not working on collagen alone. It is working on the meat fraction that clings to the collagen as well.
This is important because salt’s primary functional role in comminuted meat is extraction of the salt soluble, myofibrillar proteins, mainly myosin and actomyosin, not collagen. The literature on protein extraction in meat systems puts the optimum pH for this around 6.5 to 7.0, with an ionic strength above roughly 0.3 M needed before extraction proceeds properly. This is the same mechanism at work whether the meat fraction is a clean primal cut or the thin layer of muscle tissue riding along on a piece of trimmed silverskin. Because that meat fraction is present in Salzstoß trim, the salt has something to extract there too, and this is why the salt makes sense in trim that is mostly collagen. It is working on the meat, not the collagen.
Glue Versus Rigidity
Klebemasse is not built to give a product rigidity. Its job is glue, extra adhesion holding pieces of meat together that would otherwise separate on slicing. Salzstoß has a different primary job. It builds rigidity into the mix, the firmness that holds a formed or reformed product together as one solid piece. I now suspect Salzstoß does more than its primary job. It may bind in the glue sense, the sense Klebemasse is built for, just as well as Klebemasse, and possibly better.
Sehnenstoß Is Different
Sehnenstoß, tendon put through the same fine mince, behaves differently because tendon comes away from the carcase largely free of adhering muscle tissue. A beef sinew is close to pure Type I collagen with very little sarcoplasmic or myofibrillar protein attached. This is why Sehnenstoß is a useful differentiator in the product system. It isolates the connective tissue contribution from the meat protein contribution, so the two mechanisms, myofibrillar extraction on one hand and collagen disruption and later thermal gelatinisation on the other, can be studied and used separately. Salzstoß, by contrast, always carries both mechanisms at once, because it always carries some meat.
Sehnenstoß and Hautstoß also run through a separate hot process, cooked at 85 to 98 degrees C for an extended period, with water added during cooking. This hot process is a different system entirely and should not be confused with Salzstoß or Klebemasse. Extended cooking at this temperature hydrolyses the collagen through to gelatin, and the purpose of that gelatin is water binding, holding water in the finished product, not building a Klebemasse type rigid protein matrix. Salzstoß works cold, through salt extraction and mechanical disruption, to build structural bind. The hot Sehnenstoß and Hautstoß route works through heat and time, to build water holding capacity through gelatin. The two processes share raw materials but not a mechanism, and neither should be substituted for the other.
What the 3mm Plate Does to Collagen
The mechanical action of mincing through a 3mm plate is not just size reduction. It is a directional tearing action. The knife and plate combination shears the collagen sheath, mainly perimysium and epimysium, away from the muscle fibre bundles it surrounds, rather than simply cutting cleanly through it. Epimysium is composed almost entirely of Type I collagen. Bovine epimysial tissue has been measured at around 24 percent collagen on a wet weight basis, with gristle running to about 27 percent, and this collagen is arranged as sheets rather than dispersed fibres. When these sheets are torn rather than sliced, the result I have seen is small collagen aggregates, effectively balls of torn sheath, rather than the large continuous sheets present in the intact muscle.
Why It Binds Better
Intact silverskin sheets present a large, continuous, low surface area barrier inside the product. They do not take up salt evenly, they do not hydrate evenly, and in a finished product they behave as a mechanically distinct phase, a plane of weakness rather than something built into the matrix. Once the same collagen is torn into small aggregates by the fine plate, its surface area rises sharply. This mirrors what is already established for meat particle size in general, where comminution raises surface area and, as a direct consequence, increases the amount of myofibrillar protein available for binding. Small collagen aggregates can be salted, hydrated and coated in extracted myofibrillar protein far more evenly than a sheet ever could. On cooking, this same collagen converts partially to gelatin, and a well dispersed gelatin phase contributes adhesive strength to the matrix instead of sitting inside it as an inert sheet.
This is consistent with the wider finding in the restructured meat literature that a certain amount of well disrupted collagen, mainly from perimysium and endomysium, is not simply tolerated but can make a positive contribution to comminuted and restructured products, precisely because comminution destroys its structural continuity and prevents it from being detected as a separate phase on the palate, once it has been broken up enough that the connective tissue is no longer picked up by the palate as a texture defect. My own observation goes a step further than this literature. I am not just finding that fine milled collagen is tolerated. I am finding that it can match, and possibly exceed, the glue function that Klebemasse is built to provide, on top of the rigidity Salzstoß already gives the mix.
The Salt Extraction Mechanism
The established mechanism for binding in any comminuted meat product runs like this. Salt at sufficient ionic strength solubilises myosin and actomyosin at the cut surface of the meat particle. This extracted protein forms a sticky, viscous surface layer, and where fat is present it forms an interfacial protein film around the fat droplet, which on cooking sets into a continuous gel network that holds particles, fat and water together. This is the accepted basis for the interfacial protein film, the layer of extracted protein that coats fat droplets during comminution under salt and then sets around them into the surrounding gel matrix once the product is cooked. Because extraction happens at the cut surface, the amount of surface area created by comminution is directly proportional to how much salt soluble protein can be recruited into that film in a given mixing time.
This is exactly why particle size and plate hole size are treated as primary process variables in the literature, not secondary ones. Reducing plate hole size and comminuting more finely is repeatedly shown to raise water binding capacity and reduce cook loss. Studies varying the hole diameter of grinder plates have found that a more uniform, finer particle distribution improves the meat’s water binding capacity, which is exactly what cooked sausage needs for a strong structure and a good yield with minimal mass loss on cooking. Cookout itself has been shown to fall as particle size becomes smaller. This lines up with general grinding practice, where a fine plate in the 2 to 3mm range is treated as the range for a smooth bind and classic emulsified products, while a medium plate in the 4 to 6mm range is reserved for everyday burgers and meatballs, and a coarse plate for rustic, chunky fillings.
A 3mm Salzstoß mix has more cut surface per unit weight than a 4mm clean meat mix, so more myofibrillar protein is available for extraction within the same paddle mixing time, and mixing or massaging time itself is a recognised process variable, governing how much of that available protein is actually recruited into the binding film during the time the product spends in the paddle mixer. Comparing a 3mm Salzstoß mix to a 4mm clean meat mix therefore changes two variables at once, plate size and the collagen content of the raw material. The collagen story cannot claim credit for all of the difference.
Two Jobs, One Overlap
The current Klebemasse protocol uses clean primal meat fractions minced through the same 3mm plate and paddle mixed for the same time. Klebemasse is meant to glue, not to add rigidity. In my hands, the Salzstoß trim, silverskin with its adhering meat, has produced a glue effect at least equal to this clean meat Klebemasse, and by my impression, somewhat stronger. If the plate size and mixing time are genuinely held constant between the two, as they are in the current comparison, then the difference in raw material is the variable left standing, and the most likely explanation sits in the collagen fraction itself.
Two things are happening together in the Salzstoß mix that are not happening, or are happening to a much lesser extent, in clean meat Klebemasse. First, myofibrillar protein extraction proceeds from the meat fraction exactly as it would in clean meat. Second, the torn collagen aggregates, once hydrated and partially gelatinised on cooking, add a second adhesive phase that clean meat, being nearly free of collagen, simply does not have available to it in the same concentration. Whole muscle collagen studies show that collagen content and fibre diameter correlate strongly with the mechanical properties of the tissue they are measured in, with both collagen fibre diameter and collagen content in the perimysium and endomysium tracking meat toughness closely, and while that literature is describing toughness in intact meat rather than bind in a comminuted matrix, it confirms that collagen is mechanically active and concentration dependent, which is consistent with a well disrupted collagen fraction contributing measurably to the strength of a cooked matrix rather than merely being tolerated in it.
The restructuring literature establishes that disrupted collagen is not harmful in comminuted products and can be a positive contributor. It does not, on the evidence I have found so far, go as far as demonstrating that torn silverskin collagen from a 3mm plate matches or exceeds clean meat protein’s glue function. That specific comparison is my own finding, not yet a published one.
An Unexpected Conclusion
Trimming silverskin off ham cuts or reformed bacon primals is necessary regardless of what is later done with the trim, because intact sheets of silverskin inside a whole muscle product are a known defect. Put through a 3mm plate and paddle mixed for 20 minutes, that trim has, in my hands, matched or bettered the glue effect of the clean primal meat Klebemasse protocol I had been using until now, while still doing the rigidity job Salzstoß is built for. If this holds under proper testing, it means the trim stream from ham and bacon production is not simply a lower value by-product to be minced into economy sausage. It may do the work of two separate binding systems at once.
What Still Needs Testing
The salt extraction mechanism, the surface area effect of comminution, the identity and distribution of collagen types in epimysium and perimysium, and the general tolerance of disrupted collagen in comminuted products are all supported by the meat science literature cited above. What is not yet supported by anything beyond my own observation is the specific, measured comparison between the glue effect of Salzstoß and the glue effect of clean meat Klebemasse. That comparison needs instrumental gel strength or shear testing, run with plate size, mixing time, temperature and salt level held constant across both raw materials, before it can be stated as more than a strong practical impression from a month’s work.
References
Acton, J.C. (1972), as cited in Zhuang, H. and Zhou, G. (2019), review of restructuring technology in meat products, Journal of Food Science and Technology.
Chesney, P.J., Mandigo, R.W. and Campbell, J.F. (1978), cookout and particle size relationship in restructured meat, as cited in Zhuang and Zhou (2019).
Strange, E.D. et al., contribution of collagen to the properties of comminuted and restructured meat products, American Meat Science Association, Reciprocal Meat Conference proceedings, 1989.
Light, N., Champion, A.E., Voyle, C. and Bailey, A.J. (1985), the role of epimysial, perimysial and endomysial collagen in determining texture in six bovine muscles, Meat Science, 13(3), 137 to 149.
Gordon, A. and Barbut, S. (1992), effect of chloride salts on protein extraction and interfacial protein film formation in meat batters, Journal of the Science of Food and Agriculture.
Food and Agriculture Organization, small scale sausage production, chapter on comminution and protein extraction, fao.org.
Development of the design of plate with variable diameters of holes and its impact on meat grinding quality and efficiency, Processes, 2024.
Effect of grinding method on textural properties of cooked ground beef, meat science literature on plate size, temperature and mixing time.
Note on method. The mechanisms of salt extraction, collagen structure and comminution effects above are drawn from the peer reviewed and industry literature cited. The comparative claim that Salzstoß matches or exceeds clean meat Klebemasse’s glue effect is my own observation from one month of product development work and awaits formal instrumental validation.