Variability Testing

The SAGDesign Suite includes two variability tests which are abbreviated versions of the full SAGDesign Test (SDT). The two tests, SAG Variaiblity Testing (SVT) and Bond Variability Testing (BVT), are intended to provide test results for large quantities of samples such as those taken in geometallurgical programs. This is achieved through the use of a calibrated regression of the SAGDesign test database which allows for the accurate prediction of SDT results.

SAG Variability Test (SVT)

SVT test procedure is identical to SDT but is completed after only 3 grinding cycles at 60% passing 1.7 mm. The database regression can then be used to predict the final revolutions required to 80% passing 1.7 mm. From a practical perspective, this offers the opportunity to upgrade test accuracy to SDT standards at a later date by simply, prior to proceeding to ball milling, by continuing the test with additional grind cycles to the final grind target of 80% passing 1.7 mm. As such, SVT can also be used as a preliminary/exploratory test prior to proceeding with design quality test work.

SVT equation




AG/SAG pinion energy from F80 152.4 mm to T80 1.7 mm (kWh/t)

Predicted Revs


Cumulative number of SAG test mill revolutions to T80 1.7 mm, predicted by SVT



SAG mill feed sample mass (g)

Predicted Revs Equation

SVT b parameter equation


Slope = Slope of the grinding curve (wt. % passing 1.7 mm/rev)
Passing / Cycle = Average change in wt. % passing 1.7 mm per grind cycle
Init. Passing = Initial wt. % passing 1.7 mm in fresh SAG feed
a0-6 = Regression model parameters


A condition of using the SVT equation in the calculation of the WSVT is that the initial fines (<1.7 mm) in the SAG mill feed must be no greater than 40%.  If the fines account for more than 40% of the feed, the test must be brought to 80% passing 1.7 mm rather than 60% passing 1.7 mm, i.e. the full SAG test must be performed to generate a reliable test result. This is unlikely and generally only occurs on especially soft or friable ores.

Bond Variability Test

Similar to SVT, the BVT procedure is identical to the Bond ball work index test but is concluded after only three grinding cycles. The database regression then predicts final grams/rev of the equivalent full Bond BWI at the specified closing screen.

Sv-BVI equationWhere:

P1 = Closing screen (µm)
Predicted Gpr = Grams per revolution predicted from SAGDesign BVI


Predicted Gpr equationWhere:

Gpr3 = Grams per revolution measured in ball mill cycle three
 A and B = Regression model parameters chosen by screen size.


For more information on the SAG or Bond Variability tests, or to request a quote, please contact Starkey & Associates.

Translate »