ID grinding hard chrome.

[Terry Keeley]

Yes you're right. I missed that latest post 5mm tidbit, thought he only had the ability to take 1/8" shank tooling.

The arbors have a taper that mates to the spindle ID, the ones in the pic have a 5mm nose for mounting wheels with 5mm holes, really odd although the CBN wheel I got with it from NSK has the 5mm hole. The flanges I made adapt the 5mm nose to 1/4".

I didn't show the collet for 1/8" shaft mounted points but it mounts to the spindle with the same taper.

Off to the local hobby shop to p/u some bearings...

 

[michiganbuck]

Hard chrome beats up a diamond wheel but for a production job sometimes the use of a diamond wheel can be best.

note: I didn't read all the posts so don't know how much the take, surface finish required,or what size or tolerances.

 

[MCritchley]

I wonder if the lathes spindle is suspect? What is the axial repeatability of the spindle? Taper grinding with a spindle that jumps could be giving you issues.

Chuck up a precision ball and see if the ball moves in the axial direction from a standstill to running with a tenths indicator. Does the spindle go back to the same position each time? High end lathe manufactures publish this "camming" value on the inspection sheet.

 

[eKretz]

Minor details, I get tripped up by them all the time...

Yeah, doing this in person would go much faster. Trying to troubleshoot these types of things long distance can be a pain. Especially when one, say, skims through quickly and misses important details...
Those adapters being 5mm in the straight section and going into the larger cross sectional area at the tapered fit should really be stout enough to do the job. Of course a carbide one would be even better... After you get it all back together Terry, do a deflection test for that portion of the spindle that slides too.

 

[michiganbuck]

Taper grinding?

If the thread is about unwanted Taper grinding .. sometimes that can be the way the wheel is wearing due to the infeed to the part.

Infeed to the part done inside the bore can keep the wheel from becoming tapered the wrong way wheel. Sometimes you might take 0005 outside and .0007 inside(or what).

 

[michiganbuck]

Perhaps he needs to check that the grind is flat/straight to the taper he wants, with having a belly, convex, concave then the error might be undesirable wheel wear.

The infeed may be tapering the wheel.
Perhaps the infeed is only at the position outside, before entering the part, constant like that would taper the wheel.

Going into a dead-end bore just makes the problem worse for undesirable wheel wear..

He might check to see the wheel is not gaining taper as he uses it.

QY:[but they cut only about 1/2 thou and want a re-dress.]
Is that due to wheel taper or loading up?

Vied sounds like chatter increases on the way out. perhaps a little slower travel and increased coolamt.

 

[eKretz]

QY:[but they cut only about 1/2 thou and want a re-dress.]
Is that due to wheel taper or loading up?

Vied sounds like chatter increases on the way out. perhaps a little slower travel and increased coolamt.

I asked that question earlier, don't think he ever answered, just posted the video.

 

[Terry Keeley]

I got new bearings installed (R/C car guys use these sizes) and the spindle was much better, maybe a tenth deflection with moderate pressure.

I also reworked my setup and got a better dress, maybe not as sharp since my contact angle with the wheel is 0* but it's "truer".

Using the standard 32A60-KVBE I tried grinding dry and it worked very well! The part stayed cool and I was able to take out a thou or so before needing a re-dress. No screeching, digging in etc., it just cut and left a good finish.

I then dressed the wheel and tried it with flood coolant (Ruslick Ultracut Pro CF at 10%) and it was back to it's old tricks, seems to be mostly on the out stroke:

I'd like to run coolant if possible to carry the dust away but obviously I got the wrong chit. My toolmaker buddy says they used some green colored water soluble stuff where he worked, I've also heard Sunnen MB-30 being used. Any other ideas?

 

[EmanuelGoldstein]

I then dressed the wheel and tried it with flood coolant (Ruslick Ultracut Pro CF at 10%) and it was back to it's old tricks, seems to be mostly on the out stroke:

Used that green stuff before, forget the name but remember the smell Grindy coolant generally isn't that slippery, in your case you might try straight water to wash the swarf out. Just squirt it in with a spray bottle ... something is not wonderful with your process, in the way it's grabbing on the out stroke, but at this point you just want to get something that works (reliably). You can always tune it up after you get close.

Sounds like new bearings helped at least. One step for man ...

 

[eKretz]

Traverse speed is still too fast IMO. If you're sitting at 120 rpm on the lathe you should be travelling at about 7.5 inches per minute to be at 1/16" per rev. It looks like you're moving considerably faster than that to me. Like 3" every 4 seconds or so, which would be around 45 ipm. At that traverse rate if you have a ¼" wide wheel you're basically cutting square threads with the wheel (covering more than the full width of the wheel with every revolution). If you can't slow the stroke down, speed up the spindle. Start at 300 and see how that goes. You could use a little more FPM anyway.

 

[michiganbuck]

Traverse speed is still too fast IMO. If you're sitting at 120 rpm on the lathe you should be travelling at about 7.5 inches per minute to be at 1/16" per rev. It looks like you're moving considerably faster than that to me. Like 3" every 4 seconds or so, which would be around 45 ipm. At that traverse rate if you have a ¼" wide wheel you're basically cutting square threads with the wheel (covering more than the full width of the wheel with every revolution). If you can't slow the stroke down, speed up the spindle. Start at 300 and see how that goes. You could use a little more FPM anyway.

I addition to slowing your Traverse you might also slow your part RPM to perhaps 50, and with finding elimination of chatter increase it by small increments trying to find the sweet spot.

I have found that for some OD grinding jobs running the part way slower than normal can sometimes be what works.

 

[EmanuelGoldstein]

You made me go look .... usually they do that "your browser is too old" stuff and I throw a rock at the screen but ... does seem kind of fast, epecially work speed, but I generally don't grind 1" holes so will shut up.

But the traverse ... can't tell easily because the part is in a fixture ? But I always drive the wheel at least 25% past the edges of the hole, in both directions. The wheel does wear tapered, and you have to drive past the edges a ways to get to a straight section. Otherwise you get a lumpy-ass bore so when you try to take the last few tenths out for finish, it's too messed up and you can't do it. With all those transfer ports you're going to have a bad enough time trying to measure that thing, if it isn't straight you'll be screwed. In fact, I'd start dressing for finish and measuring a bunch farther out than usual, just because of that.

 

[MCritchley]

Any thoughts on the interrupted cut creating a harmonic?
I watched an old grinder hand fill gaps with JB weld so there was 100% contact during similar interrupted grinding.
He was really good and would resort to filling the gaps when all else failed.

I hear some really weird harmonics going on, have you considered loading the lathes spindle with a stick? Pre load the spindle so it doesn’t jump around.

And to what others have said, slow it down. Rev’s and feed, change the harmonic.

 

[eKretz]

You made me go look .... usually they do that "your browser is too old" stuff and I throw a rock at the screen but ... does seem kind of fast, epecially work speed, but I generally don't grind 1" holes so will shut up.

But the traverse ... can't tell easily because the part is in a fixture ? But I always drive the wheel at least 25% past the edges of the hole, in both directions. The wheel does wear tapered, and you have to drive past the edges a ways to get to a straight section. Otherwise you get a lumpy-ass bore so when you try to take the last few tenths out for finish, it's too messed up and you can't do it. With all those transfer ports you're going to have a bad enough time trying to measure that thing, if it isn't straight you'll be screwed. In fact, I'd start dressing for finish and measuring a bunch farther out than usual, just because of that.

It's hard to judge work speed by the video if that's what you mean - frame rates don't match up with RPM correctly so there's a sort of stroboscope effect - they never look right. I'm just going by his mentioned RPM, diameter and what I like to run for FPM. He's in the slow-to-acceptable end of the range for me on work FPM. Buck's idea about bringing that FPM range up slowly to find the sweet spot is probably a really good one with the borderline rigidity. I have had things work out better with slower work speeds occasionally too - and even sometimes grinding with the rotation of the work instead of against.

Gotta get that traverse down though so that the work is only passing by ¼ the width of the wheel per revolution of the part. Max. Less might be better in this case due to the lack of rigidity. Don't need to see the part inside the fixture to judge that - unless I miss my guess, the wheel is ~¼" wide and I'm judging that 3" stroke in about 4 seconds by the length compared to the diameter which he mentioned as being ~1". It is important to be able to quantify that feed per rev. It has just occurred to me EG that you may have meant the length of traverse.

Dead on with the rest of the 2nd paragraph.

I think the excessive stroke speed may explain some of the "catch and miss" sort of effect that we're hearing too. The wheel is hauling through the bore cutting full width and dulling the full width of the wheel out, then loads down until it all of a sudden sheds those dulled particles with an effective bang.

 

[Terry Keeley]

Thanks for all the help guys, I think I'm headed in the right direction.

As mentioned I tried the Norton 32A60-KVBE dry and it worked quite well but sucked with the Rustlick. I also tried the 19A46-K5VG and it was bad wet or dry. The 1/2" wide ceramic I cut from the 5SG46-JVS worked pretty good dry but would only take about 5 tenths per side before it loaded up, wet it was bad like the others.

I then tried the 170 grit vitrified CBN wheel after truing it with a moly round (thanks CarbideBob), it didn't work well dry but with the Rustlick did it ever cut, the coolant was coming out grey!

I should have my 100 grit CBN wheel from Diagrind this week and am hoping it'll work well for roughing, then I'll switch to a 90 grit white AO wheel for finishing.

I've got the traverse about as slow as it'll go, otherwise it starts to studder. The part is 1.700" long and I outfeed 1/2 the wheel width at each end, all are 1/4" wide except the ceramic which is 1/2". To "get that traverse down though so that the work is only passing by ¼ the width of the wheel per revolution of the part" I figure I need to speed up the part to about 400 rpm, is that right?

The interrupted cut over the ports can't be helping, I might look at filling it with Cerro metal, thanks for the idea.

 

[EmanuelGoldstein]

I should have my 100 grit CBN wheel from Diagrind this week and am hoping it'll work well for roughing, then I'll switch to a 90 grit white AO wheel for finishing.

Do you want a super-polished finish or are the grinding grooves helpful for oil retention, like honing does ?

 

[Mechanola]

Traverse speed is called feed. Way too fast. Grind dry and feed manually. Slow down everything, rotation of work, rotation of tool, and above all, feed. In return you could deepen the cut a little. Develop a feel for what you do. Still another improvement could be in lifting the tool off the work at the end of a pass, return to the enter position, use the stone in one direction only. Makes a difference.

 

[EmanuelGoldstein]

Traverse speed is called feed.

Not in the US ... or at least not by Heald or Bryant. Or Landis or Cincinnati or Norton, for that matter. Wheel moving into the work is feed, cross-wise motion of the wheel is traverse. Delay at the ends of the stroke is tarry (but i.d. grinders don't).

Still another improvement could be in lifting the tool off the work at the end of a pass

That is not how Heald and Bryant do it.

 

[michiganbuck]

qt: [Traverse speed is called feed.]

I don't know about calling it this or that but the traverses is definitely feeding the wheel into the part. The part rotation is also feeding the part into the wheel. The increments of increasing the wheel circle diameter to take stock is yet another feed.
At some point, wheel feed can stress the grinding action enough to test the rigidity of the machine and set-up. It can also test the wheel hardness and so release abrasive grits, perhaps in a not normal way. Feed can also stress the spindle power by reducing RPM.

This part, and the method of grinding into a taper to a dead-end can cause issues. Increasing the wheel circle before entering the part and traveling in can cause wheel breakdown that could cause tapering of the wheel, then on the way out of the bore the wheel could have less OD surface area and tend to grind more freely for a time until the wheel resumes a straight OD. The ideal situation is being able to OD feed at both ends of a bore and whit this part doing that is difficult.

Increasing the wheel circle diameter when inside the part may aid in retaining the shape of the wheel.

Equal amounts at both ends of the travel might be best.

All feeds going slower may reduce the feed that is most stressing the method.

Yes, I am thinking the grind is dead ended..perhaps it is not and there is some gettting off that part at bothe ehds.

 

[michiganbuck]

You might give a mandrel-mounted diamond wheel a try.
Diamond Wheels Mandrel Mounted 1/8 3/16 1/4 5/16 3/8 1/2 5/8 3/4 1 Inch Diameter, Internal Grinding

I see a 3/4 dia wheel with a 1/4 shank here... but I don't see the price.

RE:[My spindle will only take 1/8" shank mounted points or it has a 5mm arbor that I've made 1/4" flanges to fit.]

As ridged a spindle as you can fudge up is best.