Living with Variability


Farinograph & Extensograph Tests

Sourced: Analytical Support to Industry Bulletin No. 95/01

Bread Research Institute of Australia Inc, PO Box 7, North Ryde NSW 2113


Living with Variability


Since the 1930’s the Brabender Farinograph and Extensograph have been “world standard” instruments for measuring and even defining the physical dough properties of flour.  The instruments were originally conceived for the plastics industry.  They also have the capability to measure a flour’s water absorptive properties and a dough’s mixing characteristics, extensibility and resistance to extension.  The graphs on page 2 show the names of some of the measurements used to describe flour/dough quality.  Such measurements have been very useful for many years, firstly as a selective tool in breeding appropriate wheat varieties for different products and secondly in predicting performance of flour under commercial conditions, thus optimising quality for a range of end products, such as bread, biscuits, cakes, etc.

Australian input

While these instruments are manufactured to critical physical standards, differences occur in test results from the same flour measured on different machines.  Australian cereal chemists have been among world leaders in standardising operating conditions and methods, in the endeavour to minimise random and systematic variations caused by machine, operator and laboratory conditions.  Nevertheless a degree of variability between tests is inevitable and expectations of the reliability of test results should be limited to normal levels of achievable precision.

Several major companies, either millers or users of flour, organise regular inter-laboratory testing programs.  These monitor variations in laboratory results and allow a degree of instrumental or procedural adjustment for participants to treat any major problems that appear.  Industry, government and research laboratories also participate in independent inter-laboratory studies, such as those of the RACI Cereal Chemists’ Physical Dough-Testing Subcommittee or the American Association of Cereal Chemists.

Standards/Specifications for Physical Dough Tests

Both flour millers and flour users adopt appropriate quality and precision standards.

Millers must apply specifications in order to select suitable wheat supplies, to monitor mill production for conformity to desired quality characteristics, and to ascribe quality values to batches of flour for the market.

Flour Users have standards for acceptance of ingredients and for adjustment of processes.  Their aim is to ensure production of consistent quality end-products.

Farinograph measurements

Water Absorption (%): is the amount of water required to produce a dough of defined consistency (at the centre or 500BU line).Note: doughs are made from only flour and water, and the Farinograph is quite unlike bakery mixers, nevertheless this measurement gives a good indication of changes in bakery water absorption.

Development tie (min): is the time taken, from first addition of water, for the dough to reach peak consistency (“strength”).  This gives an indication of trends in bakery mixing times.

Stability (min): is the time interval between the top of the mixing curve first reaching the centre (500BU) line and when it first passes below it.  This is related to mixing tolerance.(Some users prefer a smaller time interval, taken where the peak of the curve is “flattest”.)

Breakdown (BU): or Dough Weakening is the amount that the centre of the curve has dropped from the centre-line 5 minutes after the peak (some take 10 minutes).  This is an indication of how rapidly the dough consistency decreases after reaching its maximum.


Extensograph measurements

Maximum Resistance (BU): is the maximum height of the curve, an indication of the resistance to extension (“strength”) of the dough piece.  Note: doughs are made from flour and water plus salt, mixed to fixed consistency (500BU) on a Farinograph, scaled, moulded and rested.

Extensibility (cm): is the length of the curve, an indication of elasticity of the dough piece.  For example this is related to the degree of spread of biscuits.


BRI Studies

Statistical examinations was made of Australian and overseas inter-laboratory test data to evaluate the size and nature of variability of Farinograph and Extensograph test results.  It was found that:

  • Average levels of precision, while appearing to be rather poor, are generally satisfactory in comparison with other analytical measurements.
  • Greater variability occurs between laboratories than within laboratories, as expected.
  • Variability increases as the test result increases (hence precision values shown below are given for two types of flour).
  • Prolonged delays between repeated tests cause variation due to quality changes in storage (say of the order of one month, depending on storage temperature).
  • Comparisons with commercial flour specifications indicated that sometimes greater precision is expected than is achievable – leading to likely problems in flour mill production and/or in client satisfaction.

From these studies precision limits can be given as normally achieved in testing laboratories.  The precision ranges have been halved and expressed as ± values for convenience.

 A.  Realistic expectations/specifications for repeat tests in one laboratory

These values would apply equally for within-laboratory quality assurance and also to customers who rely on quality data supplied by only one laboratory.

95% of repeat test results on one flour in one laboratory would lie within the following average limits:

  For weaker flours For stronger flours
Water Absorption ← ± 0.7%  → ← ± 1.2% →
Development time ← ± 0.4 min  → ← ± 0.6 min→
Stability ← ± 0.9 min  → ← ± 1.0 min →
Breakdown ← ± 14 BU  → ← ± 11 BU →
Extensograph (45' pull)    
Maximum Resistance ← ± 20 BU  → ← ± 43 BU →
Extensibility ← ± 1.1 cm  → ← ± 1.3 cm →



B.  Realistic expectations/specifications for between-laboratory test results

These values apply for comparison of test results on one sample from each of two laboratories.  That is, if more than one laboratory is involved in testing a particular sample, this greater level of variability must be expected.

          95% of test results on one flour from each of two laboratories would lie within the following limits:

  For weaker flours For stronger flours
Water Absorption ← ± 1.1% → ← ± 1.7% →
Development time ← ± 0.9 min → ← ± 1.3 min→
Stability ← ± 1.2 min → ← ± 3.0 min →
Breakdown ← ± 20 BU → ← ± 17 BU →
Extensograph (45' pull)    
Maximum Resistance ← ± 30 BU → ← ± 60 BU →
Extensibility ← ± 1.3 cm → ← ± 1.9 cm →


      PS When the second “pull” (90’ pull) of the Extensograph is measured, the variability in results is approximately double the above (45’ pull) values.

      What this variability means – examples:

      • Consider a difference of 60 BU between Extensograph maximum resistance (Rmax) from two tests, repeated in one laboratory on one bakers’ flour.  This does not necessarily mean that the flour has variable Rmax, but it is likely that the difference is due only to normal within-laboratory variations (as shown above in A. “± 43BU” for Rmax in “stronger flours”).
      • Consider a difference of 1.5min in Farinograph development times measured on one biscuit flour in two laboratories.  It can’t therefore be concluded that the results are any more different than is normally expected for tests in two laboratories, as this difference is less than the value of “± 0.9min” shown in B. above for development time in “weaker flours”.

      The average variability described here may be considered inadequate or too large, however this is realistic for present laboratory instruments and performance in Australia and overseas.  Nevertheless there is the possibility that precision could improve, by means such as:

      • Improved performance of test instruments and standardisation of procedures.
      • Checking performance of particular laboratories and then using those that have better than average test accuracy and precision.
      • Using only one testing laboratory, provided that its results are known to be related to the desired end-product qualities.
      • Using maximum or minimum specifications where possible, instead of specification ranges, because this reduces expected variability values.

      On the other hand, it should be noted that while laboratory tests are performed under conditions of controlled temperature and humidity, bakeries have no such control.


      Living with Variability

      To begin with, wheat is naturally a variable commodity.  Even the qualities of one particular wheat variety can dramatically change as a consequence of differences in growing region, climate or agricultural practice.  Laboratory tests are the only way of ensuring that newly purchased wheat resembles previous mill grists as closely as possible, in order to maintain consistent flour quality.  This Bulletin has drawn attention to real variability in quality testing despite continuing efforts at its minimisation.

      It is hoped that the precision values reported here will be used by

      • Flour mill customers and flour millers – as background for discussions as to whether mutual specifications are satisfactory or realistic.
      • Laboratory managers – for reviewing in-house quality assurance limits of test results.

      It is important that understanding be reached on normal variability and how this fits with a laboratory’s or purchaser’s flour specifications.  Any specification that is too close to, or exceeds the limits of normally observed variability may raise tensions between laboratory staff or between supplier and customer.  Such tensions may not be resolved by check testing.  This is because of the normal variability between repeat tests, especially check testing done in a second laboratory, or because of intervening storage-related changes in flour qualities.

      Considered discussions are required between suppliers and clients to optimise conditions for achieving desired qualities in end-products based on realistic specifications for raw materials.

      Specifications for flour and dough quality need to be drafted with a clear understanding of the actual variability associated with the test procedures.  With expectations of test results based on the precision levels actually achieved, it may be possible to reduce the number of tests required to check quality values.  It should also be possible to avoid unnecessary disputes between users and suppliers of flour.

      For the future, the BRI is involved in development work which may lead to new methods of defining quality characteristics of flour.