HRB vs HB 4

[SAmerica]

I work in an automotive manufacturing facility. We get forgings in that have certain hardness specifications. I'm trying to wrap my brain around an issue we're having and cannot find any information. We have a hardness requirement in HB. We use a 2.5mm tip @ 187.5kg load. Upon inspection of the material we've found the hardness to be too high. We contacted the supplier and their records indicate that the material was well within spec during their testing. Their method for measuring hardness is the HRB scale. We've sent samples to two different independent facilities where the HB hardness was well out of specification in the same range that I measured. One of the independent labs measured in HRB but their results showed the hardness to be in specification. The specified tolerance of the steel is 167 - 229HB. The HRB scale shows 86 - 98.2. I guess my question is what is the significant difference between the two measurement methods and why would somebody chose one over the other?

Thanks,
Kevin

 

[OGMetEngr]

The HB, or Brinell method, is good for production hardness testing as it can be done on large or heavy parts with a portable King brinell tester. To do HRB, or Rockwell, the part has to be placed on a table-top hardness tester.

 

[SAmerica]

There isn't much difference in the conversion, steel type, grain structure, etc. to make one chose either or? Both methods the part is mounted and polished and placed onto a table top hardness tester. Any insight to why a sample would measure 25 points above the upper limit of 229HB and using HRB the testing results show it right at or slightly below the upper limit of 98.2. If you convert 257HB to HRB you would get a value of 102HBR. For some reason my test piece that measures 255HB measures 97.8HRB in two different testing labs. Also a total of three testing labs have measured it in the Brinell method and their data matches mine. I'm so confused. I'm sitting on tons (literally like ten tons) of material and I need to know if we're going to burn up tooling if we run it. Are conversions based on a tolerance window so if the material measures higher than that window you'd use a different formula? I've searched all over and cannot find a reason to why my samples are measuring so far apart when using different methods.

Thanks for the help,
Kevin

 

[tbuelna]

SAmerica-

Brinell hardness is measured using a 10mm dia probe and 3000kgf, while Rockwell B-scale is measured using a 1/16" dia probe and 100kgf. For the situation you describe, BHN would seem like the better metric.

However, you note that your inspection procedure involves a large batch of forgings. Typically, there are other factors that matter more with regards to the quality of steel forgings than just surface hardness. Things like forging grain size, material cleanliness, etc, can matter more. With steel forgings, what you should strive for in your as-delivered condition is a forging that has uniform metallurgical properties/quality and is free from retained stresses. If your forging samples are showing excessive hardness at their surface, then you may need to put the forgings thru a stress-relief and normalize after forging.

 

[forging82]

Brinell hardness is more commonly required on forgings where I work over other methods. I've run into problems of hardnesses not matching when sufficient decarb is not ground off. All of the forging specs I've come across have wanted Brinell hardness on forgings.

 

[swall]

SAmerica--does your drawing call out the type of hardness measurement to be used on this part?

 

[MagBen]

when HRB is over 100, this method is not recommended. A scale is more suitable if you have to use Rockwell. Still, Brinell is more representive due to the bigger area it tests. If the probe is smaller than a grain/phase size...when hitting in the middle of grain, or soft phase, you will get a lower reading, when hitting grain boundary, or hard phase, hard particle, you will get a higher reading.

 

[CoryPad]

tbuelna,

While a common Brinell test uses a 10-mm diameter indenter and 3000 kgf applied force, multiple other indenter sizes and applied forces are allowed within the testing standards like ASTM E10 or ISO 6506. SAmerica has provided the test parameters, 2.5-mm indenter diameter and 187.5 kgf applied force. Those are close to the values for Rockwell B testing (1.588-mm diameter and 100 kgf applied force).

SAmerica,

Since the two hardness methods are similar, you should expect similar results. It is confusing that you are seeing large variation. While ISO 18265 shows 102 HRB conversion for 257 HBW, that is outside the defined range and can be used as an estimate only. Since you are near the upper limit for Rockwell B testing, and since it is possible that some labs still use the banned steel indenter balls for Rockwell B testing, I recommend you consider the Brinell testing to be more reliable.

 

[dbooker630]

If you are concerned about damaging tools during machining, perhaps you should section one of the forgings to take a look at the structure. You have not mentioned whether the forgings are annealed after forging, used in the as-forged condition, or subjected to a controlled air cooling after forging. There could be an inconsistent microstructure that may be associated with the hardness variation. If nothing else you will gain confidence as to whether you will burn up tooling.

 

[SAmerica]

Wow. Thanks for all of the responses. The material is SAE 1049S. It undergoes normalizing per TSH5101G. The forgings have an external and internal requirement which is the same (167 - 229HB or 86 - 98.2HRB). The external specification is well within spec however the internal specification is the one that's way out of specification using the Brinell method. The suppliers specification on their drawing is in HB or HRB. They choose to use HRB. The normalizing of the material is air cooled in a controlled environment. Unfortunately we don't have the means to analyze structure in the material. I believe the supplier is measuring their own. With the internal testing we're cross sectioning an area and measuring the hardness 3 - 5mm from the surface. It seems more of a normalizing issue than decarb issue. I just wish I could get some concrete data to show the supplier so we can have the parts reprocessed at no cost. Again, thanks so much for all of the responses so far. I'll be using this forum from now on. If you have any more insight regarding the information I just provided let me know!

Thanks,
Kevin

 

[SAmerica]

Are banned steel indenter ball obsolete?

 

[weldstan]

Is the internal hardness too low? If so you may need to cool from the normalizing temp with forced air

 

[SAmerica]

It's a large company with a very controlled normalizing process. They have been very reliable to us for many years. There may have been issues in their process of normalizing. I'm just wondering now if grain structure as described above is causing my data in HB to be so far out when the same sample is measured in spec using HBR.

 

[dbooker630]

It certainly could be a normalizing issue, I agree. There are numerous reasons for differences between surface and core microstructure, that are associated with material, time at heat, size of load, cooling rate, starting and finishing temperatures, just to name a few. Since you are in automotive, your quality department should have a approval package or PPAP that may have a met lab report with microstructures, as well as a control plan. Ask the supplier if anything has changed recently. If you can get some microstructures and post them here, many of us on this forum may be able to offer their thoughts on the heat treating.

 

[SAmerica]

Our engineer asked about structure analysis and they said everything checked out fine. If I had a way to check it myself I'd like to see with my own eyes.

 

[CoryPad]

Standards like ASTM E18 and ISO 6508 require tungsten carbide indenters, so the steel indenters are obsolete.

 

[dbooker630]

Is the normalized structure and hardness part of the final product customer spec, or it is your spec that is required only for machinability purposes? You mentioned a concern about tool life, ultimately if you cannot machine the stock something has to change. Can you try machining a small lot, maybe there will be enough pieces to make a call on machining.

 

[Maui]

Heat Treat facilities that perform normalizing operations typically stack as much material as possible on each furnace load to maximize productivity. This usually doesn't present any significant problems during the heating stages of the normalizing cycle. But it can create issues after the parts are pulled out of the furnace for air cooling. If they are not physically separated from each other so that they can properly air cool, then the forgings on the outside of the tray will cool faster than the forgings near the center of the load. And this can cause the forgings near the center of the load to cool so slowly that they anneal. For forgings between the center and outside of the load, you will often find a mixed microstructure that has been poorly normalized/poorly annealed. And these forgings may also show hardness variations within themselves, depending upon where they are tested. Take another look at the microstructure from two different areas of the same forging where significant hardness differences were detected. You will likely see a corresponding difference in the microstructures at these locations.

Maui

http://www.EngineeringMetallurgy.com

 

[mrfailure]

It is important - really important - to understand that hardness conversions are performed per ASTM E140, which explicitly states that conversions are estimates that do not necessarily reflect the hardness values when directly measured. I have seen major deviations when using the conversion compared with actual measurements. Since you have found such a deviation, it is incumbent on your vendor to perform Brinell testing directly to determine compliance to your specs - do not accept their conversions.

 

[SAmerica]

Unfortunately our drawing allows the use of both methods (HB and HRB). They aren't converting the data. To them their making the argument that the parts are within the specified limits given their method of choice (which is the HRB method). Just mind boggling that it measures at the upper limit on the HRB method but way over on the HB method when taking actual measurements on the same sample. I'll have to talk with the QM and QE to see if a trial run can be performed. Guess we'll call this one a mystery.

Thanks again for all the help guys,
Kevin