New figures show the extent that steel reinforcing mesh for house floors has been falling short in crucial tests, with mesh from three out of five companies failing to meet the mark.
The scores of failed tests released by the Commerce Commission under the Official Information Act - and the fact that three top laboratories rated the same pieces of steel differently - have prompted an overhaul of testing, four years after the seismic mesh first came on the market following the Christchurch quakes.
The Ministry of Business, Innovation and Employment (MBIE) has not explained why it did not try to find out sooner if the products and testing were up to scratch, in its role as regulator.
The Commerce Commission has pointed out that the failed tests alone did not prove a company had failed to comply with the standard in the Building Code.
It also said some of the companies were challenging the test results amid fears they would be prosecuted under the Fair Trading Act.
"Some steel mesh suppliers have raised concerns that release of the testing results could prejudice their right to a fair trial," said the Commission, though it added that this was unlikely.
The 70 pages of tests showed the sheets of mesh from three out of five companies failed in almost all cases to reach the required standard of 10 percent ductility, or stretchability.
Among the lows, sheets from Euro Corp scored just 1.7, 3 and 4.5 percent, while Steel and Tube had a run of three sheets with an average score of under 6.5 percent.
Mesh from Brilliance Steel recorded slightly higher test results, but still not above the 10 percent.
These tests resulted in these three companies having to withdraw their seismic mesh from the market for several weeks.
Mesh from Fletcher Building and United Steel passed the tests.
But other types of tests also failed, including for Fletcher, where the welds broke where they should not have on two of its three sheets. Welds also broke too soon on Steel & Tube and Euro Corp mesh.
The tests also showed up discrepancies between the testing laboratories - in this case SAI and SGS as well as Holmes Solutions, which did fewer of the tests.
In one instance, SAI scored a sheet at 8.6 percent ductility in one spot and SGS at 3.5 percent. Another test saw a difference of 7.5 percent versus 1.7 percent.
Across the dozens of results, discrepancies of 2-3 percentage points were common, throwing the averages out.
MBIE was now moving to fix these variations, which have thrown a spanner of uncertainty into four years of mesh production.
It has just issued draft guidelines setting out exactly what part of the mesh to test and how, which was not clear before, although it has refused to provide the draft to RNZ News.
While MBIE was using the term 'guidelines', they would be mandatory for any meshmaker wanting to meet the standard.
It points to MBIE switching from a hands-off approach to regulation that has predominated for years. It remains unclear whether it would push on with compulsory third-party testing, where other building products were judged to be critical.
As for steel mesh, only the smallest of the four grades - which range from 6mm up to 9mm - have been tested by the Commerce Commission.
RNZ News has been told by industry insiders that it was even harder to make the bigger sizes hit the ductility standard.