Any tips for making my own MT2 test bar?

[fvdbergh]

Hi everyone,

(Living in South Africa, it is not possible to "just buy" an MT2 test bar. Local tooling suppliers admitted that they have them in their catalogues, but that they do not actually import them. I could buy one from India, but if they ship it as a parcel, it will take 3 months to reach me. I might be able to have one couriered, but then the courier charges would be double the price of the bar ... so it seemed like a good idea to make the bar instead.)

I took a stab at making an MT2 test bar over the weekend, but I am not satisfied with the outcome. Here is what I tried:
1. I took some EN8 steel (comparable to 1040 steel), heated to about 930 C, quenched it in water, and tempered it at 540 C. If everything went well, that should have given me a hardness of around 30 HRc. The idea was to make the steel less gummy, but still relatively easy to turn, and end up with a slightly more durable end product.
2. I re-machined my centres with the compound at 30* to recover from the scale after the heat treatment.
3. I set up the part between centres, and spent an embarrassing amount of time getting the tailstock adjusted to turn a cylinder on the main part of the bar (the MT2 taper part was at the headstock end, driven by a small dog).
4. I tried a lot of things to obtain a good finish on the cylinder part of the bar. Using a CCMT0904 coated steel-grade insert with a small depth of cut (~0.25 mm or 10 thou) at 920 RPM (30 mm diameter stock) produced a reasonably shiny finish, but with slight chatter right at the tailstock end (with a live centre). I switched to a CCGT insert, which seemed to avoid the chatter, but the finish turned out to be rather dull. Eventually I just filed and sanded/polished the cylinder to taste, which inevitably resulted in minor variations in the diameter (about 2 to 4 micron, or 1 to 2 tenths).
5. For the tapered part, I roughed with carbide tooling, and finished with HSS. Eventually I tried taking whisper cuts of about 1 thou on the diameter. The finish was not great, and I ended up filing the taper too.

Checking the cylindrical part on a surface plate with a 2-micron indicator confirmed that the diameter was good within about 6 micron (3 tenths) along its entire length. Not ideal, but that is what I get for filing/sanding.

The real bummer was the taper. I set the cylindrical section of the bar up on a v-block, and placed an end-stop on the end of the bar (to prevent axial movement on the surface plate). Picking an arbitrary point about halfway down the MT2 taper, I measured a whopping 40 micron (1.5 thou) TIR.

So it looks like I will be starting from scratch again, but I want to resolve my issues with surface finish first. Should I
a) Go with the CCMT steel insert, and take a decent enough cut to produce the finish I want (I could pick up a smaller nose radius insert),
b) Figure out how to use HSS properly in this set-up. I know the usual advice regarding a nose radius, honing, etc, but I always get an inconsistent finish (random "scratches" that are deeper than the surrounding material) --- how deep a cut should I be taking to produce a decent-but-accurate surface ? (I know it will not be shiny, but it should be smooth).

And one final question: should I rather use a solid centre in the tailstock? What RPM range is realistic for a carbide-tipped solid centre?

-Frans

 

[4GSR]

I would suggest finding a piece of "drill rod" it's already ground on the OD and is fairly straight in length. Don't know about OD's available to you, might be metric. Find something around 25mm OD and about 300mm in length? Put between centers and cut your No. 2 MT on one end, making sure it's concentric with the already ground OD. You could also find some turned, ground and polished material in about 25mm or 1" OD, or even 20mm or 3/4" OD. As for getting a better surface finish, apply some cutting oil to the surface as you are cutting this should help. The rest will have to be polished with emery cloth. Yeah, EN8 is not very "free machining" grade or possess the additives to give better machining characteristics needed for a good finish.

 

[fvdbergh]

I would suggest finding a piece of "drill rod" it's already ground on the OD and is fairly straight in length.

Hmmm... I actually have a length of 30 mm ground rod, Silver Steel if I recall. But I am a bit hesitant to start cutting it. Then again, I do not really know what I am saving it for (I do know that it cost about 8x what the EN8 did).

As for getting a better surface finish, apply some cutting oil to the surface as you are cutting this should help. The rest will have to be polished with emery cloth. Yeah, EN8 is not very "free machining" grade or possess the additives to give better machining characteristics needed for a good finish.

I know I can get EN19 (4140) which produced a beautiful finish before (with my current lathe and tooling), I just figured that it must be possible to obtain at least a usable finish on hardened EN8.

I also see some 12L14 from the local wholesalers. Perhaps that would be the most sensible thing to try next.

 

[Bob Korves]

You will have a very difficult time getting a good surface finish and an accurate finish size with tiny depth of finish cuts. I think we have all been there and done that. Try to get the finish you want, and the size you want at the same time. Here is an example: Rough cut to within ~.040" (.1mm) of your desired size. Measure the size of the work carefully. We will now take off the remaining material in just two cuts. Using a sharp, well ground HSS bit with a fairly small nose radius (~.020", .05mm). Take a carefully set cut of half the remaining depth to finish size (.020", .05mm cut) at a high enough cutting speed to achieve a good finish. Measure carefully after the cut and see how much material actually came off the metal. Let's say the amount actually removed was .045mm. That is .005mm less than the setting, and now you have .055mm to finish size. You cut less than desired last time by .005mm, so increase the final cut by .005mm, making the final cut .060mm (.055 + .005), with everything else the same as the last cut. You are much more likely to hit your size that way than by trying to sneak up on it, and will also get a much better finish and more even finished size. It also takes less time... Practice on similar scrap material with this technique until you are happy with the surface finish and trust your ability to hit your number before doing it on work that you care about. The whole 'magic' here is to make the last two cuts with the same parameters and to nearly the same depth and expect to get predictable results.

 

[Asm109]

Bob, you need to rework your metric to English conversions. .020" is NOT .05 mm .002" is close to .05 mm..

 

[Bob Korves]

Bob, you need to rework your metric to English conversions. .020" is NOT .05 mm .002" is close to .05 mm..

Oops, I will edit the post. Only off by a factor of ten...

Edit: Done! Make sure I got it right this time...

 

[fvdbergh]

Thanks for the suggestions, gentlemen!

I took another run at it this weekend, this time using much heavier cuts (0.5 mm off the diameter) with a steel-grade CCMT0904 insert. The finish improved dramatically, and since the diameter of the test bar was not critical, I could rescue my first attempted test bar by just moving the taper a little bit forward. I still had to polish it a smidge, but not nearly as much as the during the first attempt.

I still found it somewhat strange that the lathe cut straight (on the cylindrical section), but when I ran an indicator (mounted on my toolpost) down the length of the cylindrical section, I saw about 25 micron's worth of "taper". I guess this means that my tailstock adjustment compensated for the flex nearer to the centre of the test bar during the cut. (This was my first attempt at turning between centres, so maybe this is normal)

Anyhow, I used the test bar to check the spindle of my mill (an Optimum BF20), and found that the runout increased from 8 micron near the spindle nose (say about 20 mm from the nose), to about 30 micron at the end of my test bar (about 100 mm from the nose). I indexed my test bar 180 degrees in the taper, and repeated the measurement, with the same results. From my understanding, this test implies that
a) my test bar was a reasonable success, and
b) the angular runout of my mill's spindle is more than good enough.

If I run a 2 micron DTI directly on the inside of the spindle's MT2 taper, I can measure no runout at all. I guess this means that they ground the taper after the spindle was assembled.

So it looks like my spindle is good, but I still measure about a 60 micron runout on a ground bar held in my ER20 collet chuck/arbor, near the collet, and this increases to about 130 micron if I move about 25 mm further away from the collet. So now I have to test my ER20 collet chuck to see if this angular runout is in my collets, or in the ER20 collet chuck itself. I suppose that I could even make an ER20 test bar by cutting the ER20 collet geometry directly onto a test bar, but hopefully I can check the ER20 collet chuck indirectly using different collets and different diameter ground rods.

-Frans