Rockwell Testing rounding criteria 1

[LEB1]

Rockwell Hardness Testing to ASTM E18 :2014
Clause 5.2.4 states:
Rockwell hardness measurements shall be rounded in accordance with Practice E29 with a resolution no greater than the hardness values display of the testing machine.

we have recently upgraded our machine and the display resolution is to 1 decimal place. Technically we report to this level and have now been questioned from one customer who is requesting this rounded to the nearest whole integer. he is citing Note 2 of the standard which states that common practice for acceptance testing is to round to the nearest whole number, as we are a commercial testing facility we are not performing acceptance testing ( which I believe is related to production facilities )

my question is - if our Rockwell testing machine displays to 1 decimal place can we round to the nearest whole number and still be compliant with ASTM E18:14

thanks

 

[metengr]

Yes, provided reference is made to E29 and the method selected for rounding is stated by note or other in your test results. Also, you should have a lab procedure for hardness testing regarding, method, calibration, accuracy and reporting of test results.

 

[mrfailure]

Our lab rounds to the nearest integer as SOP. You should have no problem with that particular customer's request. Personally, I think reporting to whole numbers is a relic from the days when only analog dial testers were available, and this footnote really should be eliminated as it just breeds confusion.

 

[LEB1]

Many thanks for the responses

my only concern is that the clause 5.2.4 clearly states the resolution of the machine is to be used , rounded to E29

the Note related to this clause does reference the rounding to whole number for acceptance testing , we as a commercial test house do not do acceptance testing , I believe this is essentially production testing.

Liam

 

[redpicker]

I suspect your client is, in fact, using your for acceptance testing, which is why he is making this request.

The wording in 5.2.4 states "with a resolution no greater than the hardness values display", it certainly can be less.

This is becoming more common, as more and more third party inspectors who have little specific knowledge are reviewing the test results. A standard states a hardness range of, say, 24-28 HRC. The hardness test report states 28.1 HRC. Is this acceptable or does it represent an out-of-tolerance condition? How about if the other testing is acceptable (tensile, impact, etc...)? Obviously, if the other tests are not acceptable, the hardness value is irrelevant. The controlling document is the material standard. If it required 24.0-28.0, then the value is not acceptable. However, if it states the requirement as 24-28, then it should be acceptable to round per E29 prior to determining acceptance, which is what it sounds like your client is asking.

rp

 

[mrfailure]

The safe way as a lab is to report to the 0.1 precision of the machine (if you use a digital tester)as your default. This then covers situations where you have specs written to 0.1 precision (fastener specs being a good example). Then you can always report less precision if the customer insists.

I also wanted to use Redpicker's example in another way: Suppose your average hardness measures 23.9 for hardness specified as 24 - 28? That would still be acceptable because the precision of the hardness specification dictates the precision used in E29 rounding. (This is true for other physical properties, like tensile requirements). I usually explain why the test is acceptable even if the number in that situation is less than minimum.