BHN Measurement Revisited

  • 4.7K Views
  • Last Post 07 August 2016
mtngun posted this 01 July 2016

If you've been following my 6BR thread you may be aware that I ran into some issues using smaller diameter indenters to measure BHN with my homemade tester, so I'm going back to the drawing board.    First up, my original 10mm indenter had been turned on a lathe, and while the lathe did a pretty darned good job, there were some microscopic imperfections on the tip of the indenter.    The imperfections don't show in this photo so you'll have to take my word for it.   Anyway, I splurged on some 10mm ball bearings and epoxied a ball bearing into a 5/8” thingamajiggy that fits into my push-thru die holder. I used to measure the indentations with calipers and a magnifier, but then someone pointed out that Gimp (and some other free photo editing programs) have a “measure” tool.   Take a photo of the indentation with some sort of scale in the picture, then you can zoom in on the photo and measure the indentation with accuracy limited only by the resolution of the photo. A USB microscope works well for taking the photo, but some cameras and even phone cameras are capable of taking decent closeups.Load the photo in Gimp.   Measure the scale (this one is marked in tenths of an inch).   In this example, the measurement across 0.300” on the scale was 569 pixels.    In other words, the scale of the photo is 569 pixels per 0.300 inch, or 1897 pixels per inch.     Next I measured the diameter of the indentation.   I actually measured in several different places:  257 pixels, 257 pixels, 261 pixels.   Well, the cat was chewing on my arm when I took the last measurement!     Average of the three measurements = 258 pixels.    To convert that to inches, divide 258 pixels by 1897 pixels per inch, = 0.13618 inch.   The applied load was 150 kg. for 30 seconds -- more on how I apply the load later.    Plug the numbers into the formula for BHN (a spreadsheet crunches the numbers for me) and we get 15.6 BHN.    Should we believe that number?   Well, the alloy was 50% lino and 50% clip-on wheelweight, and I was expecting 16 BHN.   That's pretty close.     In the next few days I'll try to post measurements for all my alloys including “known” standards like pure lead and linotype.     Then I'll repeat the tests with different indenters and different loads and see how the results stack up.     Most people buy store-bought BHN testers and that's fine, but like many casters I enjoy making my own stuff whenever possible.    I'm not going to run out and buy a store-bought tester, but if anyone wants to loan me their store-bought tester, I'll be glad to include it in my comparisons.

Attached Files

Order By: Standard | Newest | Votes
onondaga posted this 01 July 2016

http://www.castbulletassoc.org/view_user.php?id=608>mtngun

My favorite ball bearing company in the world is right in my neighborhood will have just the bearing for you:

http://www.acmebearings.com/>http://www.acmebearings.com/

They specify the smoothness of their bearings in microns, and they mean it.

Gary

Attached Files

joeb33050 posted this 01 July 2016

The accuracy of measuring about anything varies with the a number of measurements. The ratio of measured value/true value varies with the number of measurements made. One measurement, group size or BHN or MV, tells us little. Somewhere around 5 measurements the ratio starts getting smaller, 10 is better, 30 is sorta the recommended minimum.

This, the data, illustrates the principle:

http://www.lasc.us/Shay-BHN-Tester-Experiment.htm>http://www.lasc.us/Shay-BHN-Tester-Experiment.htm

joe b.

Attached Files

mtngun posted this 01 July 2016

Thanks for the link, Joe.   Lots of good data there.

Just to address your point, I'll  repeat the 10mm test on 50/50 alloy 10 times.   My guess is that there will be negligible variation with my tester as long as I use the same indenter and the same load, but we'll see.

In real life few of us are going to measure 30 different bullets and take an average, so if a test does not give an accurate result on the first try, what good is it?

Part of the variation in Shay's test is due to using many different testers.   For example,  13 different Cabine Tree testers were used.   For any one Cabine Tree tester, the standard deviation was very low, averaging 0.5 BHN.     But ... the Cabine Tree averaged 4.3 BHN harder than a lab BHN test -- that would be a problem for me.  

If you are only using BHN measurement for your personal use, then I suppose it doesn't matter whether your test result is 4.3 BHN off, as long as it is consistent (and some of the test results in Shay's test were not consistent).   But if you are reporting your BHN test results to the public, as I often do, then I want it to be right, so that the public can believe what I say.

Shay's test only tested the one alloy.   My intention is to test the full range, from pure lead up to linotype, and also hardened bullets.   Is the tester accurate over the full range?

One goal in this test is to see if different sizes of indenters and different loads give the same BHN result as the standard 10 mm indenter and a standard load (typically 100 kg, 500 kg, or 3000 kg).   As noted in my 6BR thread, it appears that smaller indenters result in erroneously high BHN's.    Most of the testers in Shay's article also gave high BHN's, so apparently I was not imagining things.   

Assuming it turns out that the smaller indenters give high BHN's, a second goal will be to see if it is possible to come up with a correction factor to adjust the result.

Gary, it's good to know that a few bearings are still made in America.  :)

Attached Files

Ken Campbell Iowa posted this 01 July 2016

a depressing result of mr. mtn's testing is that the formula for sampling BN doesn't appear to work . dangit !!

i made a excel ( ok, open office ) spreadsheet to follow along with the frustration ...if anybody wants a copy pm me with your email . plug in the numbers in kg and mm , just like the formula calls for. mtn did the math right...the results did vary from the formula.

?? story of my life ... heh

ken

Attached Files

mtngun posted this 04 July 2016

As per Joe's suggestion I measured 9 more of the 50% lino / 50% COWW bullets, using 10mm/150 kg, so now there are 10 data points for that combination.

BHN's:  14.8, 15.1, 17.1, 16.8, 14.5, 15.8, 14.9, 15.5, 15.6, 15.6

AVERAGE:  15.6

E.S.:  2.6

S.D.: 0.84

Attached Files

RicinYakima posted this 04 July 2016

I have a Cabin Tree tester, and that is about my spread. It can vary 2 BHn within one ingot. FWIW, Ric

Attached Files

mtngun posted this 04 July 2016

Ric, I admire the design of the Cabin Tree tester.    The Lee appears to be a decent design, too, though perhaps challenging to read its itty bitty indentation.   Lee users might benefit from using digital photo measurement.    More data using the 10mm ball bearing and a 150 kg load.   The first set of linotype measurements were made without filing the part line off the bullets -- and the mold that I use to cast the 45 caliber cylindrical test slugs does have a distinctive part line.   Well, the part line seemed to cause a lot of variation on the high side, so I filed off the part line and tried again, and got much better results. In general it is looking like the homemade tester does a bang-up job as long as I use a 10mm indenter and as long as I file the part line off hard bullets.   The tentative plan is to collect a bunch of data with both the 10mm ball bearing and with a 4mm ball bearing, and then see if I can come up with a correction factor for the 4mm results.

Attached Files

RicinYakima posted this 04 July 2016

Yep, the only real issue I have with the Cabin Tree is that the back has to be flat also, otherwise it rocks and screws up the measurement. There is a learning curve to make this tooling work.

Attached Files

mtngun posted this 07 July 2016

One of the goals of this project was to try to find a way to get decent results with a 4mm ball, so that I'd have a way to measure 22 & 243 caliber bullets.    Also I wanted to try a 4mm / 60 pound combination because I read somewhere that the Lee tester uses 4mm / 60 pounds (actually 5/32” but that's nearly the same as 4mm). Here's some results for 4mm / 60 pounds, and as you can see, they stunk.    Not only were the BHN's goofy, but the variation was awful.    I.e. linotype measured anywhere from 28.3 BHN to 75.5 BHN.   :X         I've never used a Lee hardness tester so I can't comment on how well it works, or if it really uses 5/32” / 60 pounds.     If anyone wants to mail me their Lee tester on loan for a week or two I'll be glad to run it through its paces. So then I tried increasing the weight on the 4mm ball to 50.3 kg (111 pounds).    No joy, it still gave nonsense numbers with lots of variation.    Again I increased the weight on the 4mm ball, this time to 71.8 kg (158 pounds), but the BHN numbers were still nonsense.    Maybe I'm missing something, but I haven't been able to get a 4mm test to work decently.   I'm going to give up on 4mm for the time being.    Continued in the next post ...

Attached Files

mtngun posted this 07 July 2016

From “Engineering Properties and Applications of Lead Alloys” by Sivaraman Guruswamy:The takeaway is that his recommendation for measuring the BHN of lead alloys is 10mmm / 125 kg for 3 - 39 BHN, or 10mm / 250 kg for 5.6 - 78.8 BHN. 

Attached Files

mtngun posted this 07 July 2016

The good news is that I continue to get decent results using a 10mm ball bearing and 150 kg load.    The numbers for known alloys like pure lead and linotype look good, and the other alloys are at least in the ballpark.    Repeatability is fair. One thing I learned while doing these tests is that imperfections on the the surface of the bullet can affect the results quite a bit.   I got into the habit of filing the surface to remove any part line and to make the surface reasonably flat, but the file leaves microscopic striations and even a filed surface sometimes is still not flat.   Striations or non-flatness seem to result in BHN variation on the high side.    FYI “R2RS” = Batch #2 of Rotometals Reclaimed Shot, and “HTR2RS” = Oven Treated Batch #2 Rotometals Reclaimed Shot. I've been experimenting with different webcam software to operate my cheap USB microscope (in Linux).   For this photo I used “Cheese", with the resolution set at 1600 pixels wide.   That gives you some respectable resolution for measuring even small indentations like on this heat treated reclaimed shot.        Another thing I noticed is that if I measured a particular indentation with digital photo measurement, and then checked the measurement with calipers under a magnifying lamp, the calipers would always read 0.003” - 0.009” larger.    Which is the correct measurement?     I'm not sure, but I'm going to stick with digital photo measurement rather than relying on my not-so-good eyesight to read calipers.       A customer kindly offered to loan me his Cabin Tree tester for this evaluation, and when it shows up I'll run it through its paces and see how it compares to my homemade tester. Otherwise, my plan is to stick with the 10mm / 150 kg homemade tester.    That means I will not be able to test small caliber bullets, but oh well.  I'm thinking about making another homemade tester, based on a nutcracker design, that will have more leverage so it can use 200 - 300 kg load, for a bigger indentation on heat treated alloys.

Attached Files

John Alexander posted this 07 July 2016

I have a lee tester and would be glad to put it in the mail to you if someone else hasn't already offered and your offer is still open.  Same with my LBT tester that I believe you asked for earlier.   I have used the LBt tester for a long time and when I use it on Linotype or Lyman #2 it seems to give reasonable results although you have to use carefully and take several readings.

John

Attached Files

mtngun posted this 07 July 2016

John, PM sent.

Attached Files

mtngun posted this 07 July 2016

I added 20:1 alloy to the chart for the homemade tester with 10mm ball & 150 kg load. Most sources put 20:1 at 10 BHN, and my homemade tester averaged 9.6 BHN.   It looks I'll be getting the loan of John's Lee hardness tester (if he can find it! )  so it'll be interesting to see if Lee has better luck with a 4mm ball than I did. 

Attached Files

frnkeore posted this 08 July 2016

Can you compare your results to the process that I use?

I first got it out of the old NRA Cast Bullet publication.

"pure lead diameter / sample diameter, squared, times 5"

I have always used a 5/8” diameter ball, on igots and was able to get accurate measurements when compared to BHN harness.

I use a arbor press to make the indent but, as pressure doesn't seem important with this medthod, a lot of people use a vise. I haven't tested other size balls but, it's not suppose to be important either.

Frank

Attached Files

mtngun posted this 08 July 2016

frnkeore wrote: Can you compare your results to the process that I use?

I first got it out of the old NRA Cast Bullet publication.

"pure lead diameter / sample diameter, squared, times 5"

I have always used a 5/8” diameter ball, on igots and was able to get accurate measurements when compared to BHN harness.

I use a arbor press to make the indent but, as pressure doesn't seem important with this medthod, a lot of people use a vise. I haven't tested other size balls but, it's not suppose to be important either.

FrankMy 10mm / 150 kg indentation for pure lead = 0.232" My 10mm / 150 kg indentation for oven treated reclaimed shot = 0.083"

Plugging those numbers into your suggested formula yields 39 BHN for the oven treated reclaimed shot, vs. 42.6 BHN using the standard equation for BHN.    So your process appears to produce reasonable results as long as a 10mm ball is used.     I have not tried a 5/8” ball.

Attached Files

frnkeore posted this 09 July 2016

In thousands of a inch, the error for the HT alloy would only be .0035 (.0795 Vs .083) to make the formula read 42.6.

I'm a retired machinist of 40+ years and I use a dial caliper with a 5X, magnetic loupe attached to it. Even with that, it's hard to define the exact edge of the indent.

Although not available to most anyone (I don't even have one) the best way to measure the indent is with a 20X Comparator.

Frank

Attached Files

mtngun posted this 09 July 2016

Agree, Frank.

The biggest disadvantage I see to the ingots-in-a-vise method is 1) it would be tough to do with individual bullets and 2) ingots usually have a rough surface, or at least my ingots do.   Plus I would encourage anyone using it to use a 10mm ball.   I paid $8 for 100 ball bearings so that's still a bargain compared to a store-bought tester.

My interest in BHN has always centered around heat treating or age hardening of antimonial alloys, so I want to be able to measure bullets, not just ingots.   

I've finished the Cabine Tree tests -- they went very fast because it was very easy to use -- and will try to write up the results later this evening.  

John offered the loan of his LBT tester, but I'd already used one back in the mid 80's and already formed an opinion, so I'm not planning to include the LBT in this comparison.   The shorter version is that I thought there was a lot of variation and that I could make a better tester myself, hence my homemade tester.    On the plus side, the LBT is capable of measuring small meplats and seems to give a result that is at least in the ballpark, which is more than you can say for some testers. :D

Attached Files

mtngun posted this 09 July 2016

The borrowed Cabine Tree tester showed up.     I was impressed with the heft and the quality.     The Cabine Tree does not use a ball indenter, rather it uses a cone with a small meplat.  Similar to a Saeco tester, it doesn't really read BHN, it just reads “something” and that “something” is supposed to correlate to BHN. It was extremely easy to use the Cabine Tree and the tests went very quickly.   The dial indicator eliminates most of the human error. The instructions suggested filing a flat on the surface to be measured, and I agree with that.    I also found it helpful to lightly file the other end of the bullet, to knock off the sprue bump.   If you don't, the sprue bump might compress a little bit during the test, and that will throw the reading off. A supplied chart is supposed to correlate the dial indicator results to BHN  I measured each alloy at least 5 times and took an average, results below. I had high hopes for the Cabine Tree because it looked well built -- and it is well built -- and because the dial indicator removes most of the human error.   But like most storebought hardness testers, the Cabine Tree uses a small indenter rather than the standard 10mm ball.    That's understandable because a 10mm Brinell tester would be big and heavy and expensive.   The problem is that it seems to be difficult to get good results with a small indenter.

Attached Files

mtngun posted this 09 July 2016

Following up on Frnkeore's suggestion:

My first thought was “But that only works on ingots.   How would you use it to measure bullets?"

This is what I came up with.

I use a 45 caliber cylindrical slug for BHN and density measurements.   In any event, 45 is a popular caliber so many casters will have a mold for a flat nose 45.

I cut a piece of 1/2” pex pipe that was slightly shorter than 2 cylinders and the 10mm ball.   The pex serves to hold the cylinder sandwich together during the test.    Put the sandwich in a vise and squash them.

That was easy! :)  However, the pure lead slug obturated so much that it was stuck inside the pex. :(   To get it out, I ended up slitting the pex with a utility knife.   That's OK, the slitted pex still will serve to hold the cylinder sandwich.

Results, as measured with a USB microscope: Dpure = 0.299" Dlino = 0.1535"

5 BHN * (Dpure / Dlino)^2 = 19.2 BHN

That compares to 21.7 BHN for the 10mm / 150 kg test.   At least it's in the ballpark.

I'll run this system through its paces with other alloys and we'll see if it continues to give ballpark accuracy.   

Advantages: --  cheap, cheap, cheap ! --  if it indeed gives reliable ballpark accuracy, that's more than you can say about some hardness testers. :D

Attached Files

  • Liked by
  • papertrl
Show More Posts
Close