Photogrammetry for hardness testing...

[DrB]

I use a Lee Hardness Tester for testing hardness, which seems to give the expected results, and I believe it ranked relatively well on LASC's study of hardness testers.

The indenter is well made and works well. Occasionally a small bullet may shift in the v-block holder, but you can usually tell and discard the reading on the basis of an ellipsoidal indentation.

However, the optical microscope for measuring that comes with the kit gives me absolute fits. I find it difficult to use, even with auxiliary lighting from a flashlight or headlamp held off to the side (or behind, or above, or... I think I've tried all the various contortions). I can either see the gradations, or I can see the surface... I have a difficult time making out both with confidence.

So I was thinking... I've got a pretty nice DSLR camera, with a good macro lens. Why not take the indentation measurements by photogrammetry? Instead of hunching over this $2 silver pen thingy straining my eyes and contorting myself trying to get a little better lighting angle, I can snap a picture and measure hardness sitting comfortably at my laptop.

So far it seems to work well... I'm getting measurements that are repeatable to within about a thousandth of an inch, which is less than a BHN point on the lee conversion table.

Here's how I did it.

I used the diameter of the bullet being tested as a reference, and then as a check I also used my caliper jaws as a reference by setting them to .100". Anything will do for a reference, but I'd make the reference measurement relatively large as compared to the indentation to minimize error due to pixel size... the bullet is likely to be most convenient reference, especially if you've already measured it's diameter with a micrometer as I had. Make the reference object something not much more than an order of magnitude larger relative to the indentation, and not very distant (so as to minimize any lens distortion or perspective differences).

Then measure the diameter of the indentation in your image (using your favorite graphics tool, I use GIMP on a linux platform), as well as the diameter in the picture of the reference using the same approach. To make it easy I placed them at close enough to the same distance from the camera lens that perspective difference would be negligible.


Since we've set everything up to be normal to the lens and for perspective differences to be negligible, t's just a ratio to calculate diameter of the indentation:

diameter indentation = (reference measurement)/(image reference size) * (indentation image size)

My results for the various indentations I measured (I made a number) were as follows (sorry about the table, guys -- suggestions are welcome for a better way to do this):

location Using bullet image 141 Pixels bullet actual measurement Estimated size Average Size Hardness (BHN)
nose 48 48 224 225 48.21
middle 48 47 224 225 47.71
tail 51 51 224 225 51.23
Using bullet image 140 Pixels jaw
nose 45 45 95 100 47.37
middle 46 46 95 100 48.42
tail 49 49 95 100 51.58

So rounding to precision, BHN from these measurements is consistent to within a point at around 23 BHN for the nose measurements and 20 BHN for the tail.

All in all, I'm pretty happy. It beats the heck out of trying to use the silver stick thing that comes with the kit (and which I estimated a hardness of 21 BHN with great difficulty with).

Now that I've done this and it appears to be working to my satisfaction, I'd skip the calliper jaws as a reference object. I am already likely to have a micrometer measurement of the bullets diameter, and as it turns out this works just great as a reference object. If anyone else wants to try this, just take a measurement of your bullet, and use that for your reference dimension in your calculation. Easy as pie.

Attached is a screenshot from the measurement process.


http://castboolits.gunloads.com/imagehosting/thum_154544e193fbbdbb76.png (http://castboolits.gunloads.com/vbimghost.php?do=displayimg&imgid=1430)

[JIMinPHX]

I can't find it anymore, but there was a thread with a bunch of pictures of rigs that people made up to hold the little viewer gizmo for the Lee hardness tester. Some people managed to get a digital camera to take a clear picture of the view through the eye piece. With a little fixtureing, those little Lee pocket scopes are not all that bad. The best cheap & simple fixture that I saw was the top half of a 2 liter coke bottle that had a hole poked through the cap for the Lee scope. the sample to be read went inside the coke bottle & the bottle held the scope in place while you looked through it. right after I saw that picture, I copied it. It really helps a lot.

[DrB]

Jim, if I didn't already have the camera and lens I'd go with that (it would be FAR FAR cheaper, I'm certain).

Given that I have the camera, and that this is extremely easy and repeatable, I don't think I'm ever going back to the silver gizmo. :)

I can even catalog photos of a sample to show age hardening, or go back and recheck my measurements at a later date.

[DrB]

Two questions -- how do I change the display size of an image, and what's the best way to display a table? Do I need to import it as an image?

[JIMinPHX]

I wasn't trying to say that there was something wrong with your method. I was just trying to point out another option for you to consider. Sorry if it sounded like I was trying to rain on your parade. I wasn't.

The board will accept bmp, doc, gif, jpe, jpeg, jpg, pdf, png, psd, txt, & zip files as direct postable attachments. One possibility may be to put a table in a .doc file & try to post that. Otherwise you may need to make the table into a photo format. One ease way to do that is to use Ctrl+Print Scr to copy what you see on your computer screen to the clipboard, then past into a blank document in MS-Paint & save as a photo format. You can later crop that picture to only show what you want.

When you post a photo directly on the board, then the board decides if it can post the whole photo or if it should shrink it down & just post a thumbnail that you can click on to see the whole photo. If you reference a photo location that resides somewhere else on the Internet, then the board seems to post it at full size most of the time. This even works if you post a picture on one thread where it is thumbnailed, & then reference that photo in another thread, where it will show up full size.

[JIMinPHX]

example - http://castboolits.gunloads.com/imagehosting/154544e193fbbdbb76.png

[DrB]

I wasn't trying to say that there was something wrong with your method. I was just trying to point out another option for you to consider. Sorry if it sounded like I was trying to rain on your parade. I wasn't.

The board will accept bmp, doc, gif, jpe, jpeg, jpg, pdf, png, psd, txt, & zip files as direct postable attachments. One possibility may be to put a table in a .doc file & try to post that. Otherwise you may need to make the table into a photo format. One ease way to do that is to use Ctrl+Print Scr to copy what you see on your computer screen to the clipboard, then past into a blank document in MS-Paint & save as a photo format. You can later crop that picture to only show what you want.

When you post a photo directly on the board, then the board decides if it can post the whole photo or if it should shrink it down & just post a thumbnail that you can click on to see the whole photo. If you reference a photo location that resides somewhere else on the Internet, then the board seems to post it at full size most of the time. This even works if you post a picture on one thread where it is thumbnailed, & then reference that photo in another thread, where it will show up full size.

Jim, my apologies if my reply sounded prickly, I wasn't offended at all by your suggestion. I had come across mention of fixtures for holding lights with the lee microscopes elsewhere, but as I had never found a lighting angle/brightness that made it work well for me, I never got around to the ergonomic aspect of the problem

You know, the biggest problem with my approach that I see is cost. :) If you have a camera/macro lens already, great -- but I don't think anyone's going to go out and buy one just so they don't have to fool with the lee microscope!

Thanks a lot for your suggestions and help. Again, my apologies if my reply came across as brusk.

Best regards,
DrB

[JIMinPHX]

No apologies were needed, but thanks for putting the good foot forward anyway.

Regards,
Jim

[markinalpine]

Celestron, the telescope maker, also makes a USB microscope that includes software for making measurements. I've been on the lookout for one, maybe on sale.

[DrB]

Celestron, the telescope maker, also makes a USB microscope that includes software for making measurements. I've been on the lookout for one, maybe on sale.

That sounds like great fun with the kids, regardless.

You can always measure a mic'ed bullet to make sure it's accurate, and away you go.

[Gunslinger1911]

I just looked on Celestron site, $45 MSRP for that USB microscope. Prob less in the real world.

Lee hardness kit = $45 from factory outlet .

Damn, anybody want to buy a slighty used Lee tester ? LOL

[DrB]

I just looked on Celestron site, $45 MSRP for that USB microscope. Prob less in the real world.

Lee hardness kit = $45 from factory outlet .

Damn, anybody want to buy a slighty used Lee tester ? LOL

Hold on there, partner! :) you'll still need the lee indenter and vblock so as to make your indentation to measure with the microscope.

If anyone tries the celestron, please let us knew how it works for you. That would be a lot cheaper than a camera and macro lens for someone who doesn't already have one.

On the other hand, cameras are getting more and more pixels all the time. It may be you can use a cheap camera of sufficiently high pixel count with a macro mode (sometimes indicated by a flower on the mode selection).

[Gunslinger1911]

Hi DrB,

There is info (LASC ?) on how to use a reloading press, with a formula for diameter of ball, amount of weight and dia of dent (I measured with calipers) to calculate Bn - I tried this while waiting for my Lee tester.

I think I used a BB and 20# weight (x4 for press advantage).

This method and the USB microscope would be pretty slick. Oh crap - I talked myself into doing it !!!

[DrB]

Sounds reasonable to me. Lots of great info on the lasc site.

Here's the lasc tester experiment I mentioned earlier but didn't link to. I'll bet that with a better optical device (camera + macro lens or otherwise) the lee result standard deviance would have been much less.
http://www.lasc.us/Shay-BHN-Tester-Experiment.htm

Resources are limited everywhere gunslinger, if you find something that works (or doesn't), please share your experience with the rest of us! I'm interested in the idea of a usb microscope... Here's some of the celestron offerings... http://www.google.com/#sclient=psy&hl=en&source=hp&q=celestron%20microscope&aq=0&aqi=&aql=&oq=&pbx=1&bav=on.2,or.r_gc.r_pw.&fp=6b34e10cda1a477b&biw=1920&bih=978&pf=p&pdl=3000

One thought though... Is it a transmission microscope or a reflecting microscope (not sure of my terminology)? Anyway, if it uses a stage and slides with bottom illumination I would expect you may have some grief getting enough light on the surface of the bullet to get a good image (but maybe you can rig something)?

If you have a pretty high pixel camera already you may want to try taking a picture of your caliper jaws and see if you can get adequate resolution before spending any money.

[markinalpine]

"USB Microscope
For students or professionals, this Celestron® handheld digital microscope that captures still images and videos is a terrific computer accessory. Plug the 2.0 USB cable into your computer's port and take advantage of the built-in, upgraded 2mp digital camera (Up to 5mp with interpolation.), 10X to 40X and 150X magnification, and LED illumination. The 4-oz, 4-1/4" x 1-1/4" dia scope may be handheld, or slipped into its metal stand for steadier viewing. CD-ROM software includes a measurement function. Minimum requirements UVC Plug & Play with Windows XP, Vista, 7 and MAC 10.4.9 and later, CD/DVD drive, open USB port. Very cool.

14474 CELESTRON HANDHELD DIGITAL MICROSCOPE $59.95 EACH (was $63.95)

http://www.sciplus.com/category.cfm/subsection/4/category/43

The link above is to their catalog page.

From the photo, this item appears to be a reflecting type, with LED illumination built in.

Mark :coffeecom

[DrB]

Sounds like it could work... I wouldn't trust the measurement function until you test it against a standard like the bullet diameter, but that's no big deal at all.

Gunslinger, when can we expect pictures? :)

:kidding: (Sort of). :)

[Gunslinger1911]

I'll do a little surfing for the best price - hope to order this weekend.

(the science nerd in me loves new gadgets !!)

[DrB]

Amen, brother.

There's no science without observation & measurement.

[Capt Klutzz]

Sounds like it has possibilities. I think I would find some sort of standardized hole gauge or some thing to use as a comparative dimension instead of the bullet dia.

I recently bought the lee tester to check some ingots that i smelted from a scrap yard score against some HB alloy that I got from roto.
The purchase was dual pupose 1.) seed stock for the pot I am planning to buy as well as a calibration check on the new to me Lee tester.

I agrre that the scope leaves alot to be desired but can be usefully with some patience. I was lucky to have access to a better quality microscope that is part of a harness testing kit I use at work to to the same test on structural steel.

I didnt realize tho that the scale was metric so then I had to do the conversion but when all was said and done the results were identical and acceptable.

The best advise I could offer would be to figure a way to make a small stand for the scope to help steady it. File or sand the indent area of the test piece in such a way as to minimise the dark "grain" lines and then orient the gaduations on the scale at a right angle to the grain lines so that the measurments are easier to see.

[Doc Highwall]

One mistake I see people making is checking hardness too close to an edge or too close to another indent, this will give you a false reading that is lower.

[DrB]

One mistake I see people making is checking hardness too close to an edge or too close to another indent, this will give you a false reading that is lower.

:)

Yes, if you overlap a prior indent that would I expect give a larger indentation for the second indent (the first indent should be fine when measured on a radius arc as I did above), so your hardness will be lower.

The tail end indentation is an example of this, as I think you are alluding to, DH.

The important thing here was that the measurements were repeatable and consistent with Lee microscope measurements when done using a camera and digital software, but your catch & caution are appreciated.

Thanks Doc.

Best regards,
Caleb

[DrB]

Sounds like it has possibilities. I think I would find some sort of standardized hole gauge or some thing to use as a comparative dimension instead of the bullet dia.

Why? If you've mic'ed the bullet diameter, why would you want to use anything else? If you're measuring hardness of a bullet, it will always be available, in the picture, the indentation subtends the bullet diameter, and the distance from the camera to the bullet vs. the indentation will always be within a half caliber of being the same distance (which is a good thing).



I agrre that the scope leaves alot to be desired but can be usefully with some patience. I was lucky to have access to a better quality microscope that is part of a harness testing kit I use at work to to the same test on structural steel.


Yep -- I agree that the Lee can be used. I just find it a PITA for my eyes/lighting. If you have an alternative measuring device, then you're all set.



The best advise I could offer would be to figure a way to make a small stand for the scope to help steady it. File or sand the indent area of the test piece in such a way as to minimise the dark "grain" lines and then orient the gaduations on the scale at a right angle to the grain lines so that the measurments are easier to see.


Yeah, aside from the scope stand I tried every variation other than thinly painting the metal for better contrast.

I really didn't have an interest in fiddling with a scope stand as it seemed to me (that for my eyes, at least) the problem was as much one of contrast as ergonomics.

Holding up a camera and snapping a picture works great for me, and then I've got a record for that bullet/alloy. I also can work comfortably looking at the image on a large screen. A comparator would be even handier, I suspect.

I think if I was in a bigger hurry, I would go to the cabine tree hardness tester, as discussed on another thread. While I don't think it would be as inherently precise as the Lee with softer alloys, I think it would do just fine, and would be quicker than this approach. (If you already have a camera that will work though, this approach with a lee is less than half the cost of a cabine tree).

Best regards,
DrB