Needs Pictures: 0
Results 1 to 15 of 21
-
18th Jun 2021, 08:58 AM #1Diamond Member
- Join Date
- Aug 2019
- Location
- Revesby - Sydney Australia
- Age
- 56
- Posts
- 1,183
Repairing a chuck/replacing a chuck/register frustrations!
This is a bit of a saga.
My sad old ex-school Hercus has a sad little Pratt chuck. Its jaws aren't parallel – splaying out toward the tailstock (from youth grabbing little things very tightly?), and its about 4thou out of true.
Last week, I decided to try improving it.
First, chuck off the backplate, and clock its front. It was wobbling by nearly 100 microns!
This chuck and plate must have been swapped from another lathe spindle?
Easy, I think. Re-machine the "spigot" to fit the chuck, and we should run true.
A pleasureable 10 minutes to remove 2mm, remount the chuck, remeasure.
Hmmm. Body & face about 60 microns out.
Second, machine the jaws. Find a ring to expand the jaws out into, put pressure on, and bore a little off them with the sharpest tool I can find - an internal threading insert:
IMG_1572.jpg
5 or 6 passes, at a fraction of thou each time. About 1.5thou total D.O.C. until the front inside of the jaws got a little round machined in their middle.
I even decided to re-cut the little deep grooves, for the next time I'm boring out washers.
My 16IR was a little too pointy for that. Had to borrow a different internal thread tool:
IMG_1573.jpg
Now, vacuum out the swarf, and mount a 40mm precision round for testing, and...
still nearly 100 microns, or 4 thou out
I try a different round – a 16mm ejector pin – it comes in at nearly 2 thou.
I try tapping the high side, and notice light under the front of the jaws.
Not parallel? What the?
Remove the round, clock the freshly machined groove along the middle of one jaw as best I can
(it is hard to get good access in there).
Inconsistent results.
I fiddle around with the jaws a bit. They have always been loose in the scroll, but today I notice there is also slight wobble in the T slots. Maybe machining them while clamping outside won't work, because the jaws are rotating in the T slots?
Scratch my head a bit. Decide to re-machine the front while grabbing something deep in the chuck. Find 20mm round slug, put it deep inside, past the front of the chuck body, and re-do the boring.
Another 2 thou D.O.C., and I have jaws with a big step in them. Think about milling, Dremeling, hand filing, but decide to surface grind:
IMG_1574.jpg
Tricky to hold and shim for the level/taper, but end result looks something like:
IMG_1575.jpg
So, mount them back, expecting perfection, and... still nearly 2 thou out of round.
Plus, there is still a slight step in the back of the jaws when grabbing small round stock.
Another day, work out a way to machine that out in the chuck. I find a little steel sleeve, loosely grab it at the front of the jaws, bore extra clearance so it doesn't foul on my "boring" tool, and gently take the last of the step out:
IMG_1576.jpg
which seems to not line up with the previously machined front of the jaws, so I also grab my 20mm round inside the chuck, and finish the front of the jaws again.
Clamp small precision round, and check:
IMG_1577.jpg
Yay! At last some improvement.
Test with larger round stock:
IMG_1578.jpg
Aurrrrrrgh!!!!!!!!!!!! Worse than before.
After two days of careful machining and grinding, I have a chuck which is better at some radii than others, and has a total of over 6 thou removed (so this chuck will never grab very small rod again unless I mill/grind the bevels on the jaws).
The scroll is too badly damaged – time to cut my losses.
Toolmaker takes pity on me, and gives me an unused – still in box – 125mm TOS 3 jaw. Onto part 2...Last edited by nigelpearson; 18th Jun 2021 at 09:03 AM. Reason: missing attachments
-
18th Jun 2021, 10:32 AM #2Diamond Member
- Join Date
- Aug 2019
- Location
- Revesby - Sydney Australia
- Age
- 56
- Posts
- 1,183
Part 2 of the saga - new backplate
I have been keeping some old, side of the road throw out, barbell weights for this purpose (pipeclay's suggestion from a few years ago).
5lb weight is just thick enough. Mount on sad 4.5" Pratt chuck, grabbing inside the barbell hole:
IMG_1579.jpg
and gently machine a little step:
IMG_1580.jpg IMG_1581.jpg
then grab that in the chuck, and use threaded rod to lock it onto the chuck. Machine the face with a positive clearance insert:
IMG_1582.jpg
(because the negative rake TNMG was rubbing and creating sparks). Lots of fine dust. Vacuum helps.
Remove rod to finish off the centre:
IMG_1583.jpg
Now for the threading. Instead of a bigger lathe, use an old face plate, using the dodgy-est mounting system I can cobble up:
IMG_1584.jpg
Bore out to 1.33", with a little step to remind me about the register:
IMG_1585.jpg
Now, the change gears. Turn this into that:
IMG_1586.jpg IMG_1588.jpg
using two 32tooth gears. Note that the 80tooth is hard to remove - I had to use a bit of brass to lock a tooth:
IMG_1587.jpg
Note also that the Hercus chart says to use the 80tooth in the middle, but I don't have any axle/spacer hardware to do that, so the 72 had to go in there.
Now for the tricky part. My job is mounted hard on a face plate – no gap. A face plate with an 8tpi 1.5" thread on it. So, I align the cutting tool exactly on the spindle thread.
Tool in the valley (root) on the back of the spindle means it matches the peak (crest) on the front of the spindle, which lines up with the valleys in the faceplate thread:
IMG_1589.jpg
and seventeen careful passes later, here is the thread. If you look carefully, you can see the transition half way down where some metal has been pushed into the face plate's thread:
IMG_1590.jpg
Thread is a little loose, but as long as I do a good job on the register, it won't matter, right?
Machine a spacer so I can reverse mount the new face plate:
IMG_1591.jpg IMG_1592.jpg
and carefully bore the register. It was hard to see in there without some focussed extra lighting:
IMG_1593.jpg
I decide here that I will keep the thickness/weight, and "history" of the barbell casting. So I machine a little of the outer face, but leave as much of the raised lettering as possible.
Mount the correct way. Tight for the last half turn, rubbing a little at the very end of the spindle thread (will try to make a photo). Smooth off the outside edge:
IMG_1594.jpg
(which abraded one corner of a TNMG. Lucky it has 6!)
Now for the chuck register ("spigot"). Take about 1/16" off, because I don't have a lot of thickness
IMG_1595.jpg
and finish up in there with a new TNMG's corner. Take my time, and I get a good friction fit – the faceplate can hold the weight of the chuck!
Finally, the bolt holes. M8, so I use 19/64 drill to locate on the drill press:
IMG_1601.jpg
(using electrosteam's milling clamps in an odd way)
then clamp and spot using an odd taper drill:
IMG_1602.jpg
and drill out at 1/8", then 8mm:
IMG_1605.jpg
Chuck came with bolts, but they won't clear lathe's belt tensioning handle, so I countersink for countersunk HT bolts (set screws?):
IMG_1606.jpg
Sadly, one of the holes is in an inconvenient place. Had to Dremel a flat before spot drilling:
IMG_1608.jpg
End result:
IMG_1609.jpg IMG_1610.jpg
But, mount it for testing, and 'tis not as true as I hoped:
IMG_1611.jpg IMG_1612.jpg
So, after 1 week, I'm at a loss.
What is causing it? Spindle shoulder? Thrust bearing? (unlikely) Nigel's bad machining?
Looks like some more careful dial indicating, and maybe some precision grinding, is in my near future
-
18th Jun 2021, 10:33 AM #3Most Valued Member
- Join Date
- Nov 2007
- Location
- melbourne australia
- Posts
- 3,228
I got a bit lost in all of that. I've never heard of anyone machining chuck jaws true- always grinding. Your variable runout is probably due to you measuring against stock of varying diameters. The runout of a 3 jaw varies with the position of the scroll. If you grind them true at a given diameter, they will only be dead-on at that diameter.
Also, please decide on a unit of measurement and stick to it! Either thou or microns (pref the metric one).
Sorry I can't be of more help.Chris
-
18th Jun 2021, 04:50 PM #4Most Valued Member
- Join Date
- Jun 2007
- Location
- sydney ( st marys )
- Age
- 64
- Posts
- 4,887
How deep is the recess for the register?
Is the back of the backing plate running out ?
Did you cut the register at the same time you cut the thread?
-
18th Jun 2021, 04:53 PM #5Most Valued Member
- Join Date
- Jul 2016
- Location
- Melbourne
- Age
- 35
- Posts
- 1,522
First thing to do is check your spindle for axial play, end float and run out. Have to trust each component and that means inspect each component.
Sent from my SM-G973F using Tapatalk
-
18th Jun 2021, 06:13 PM #6
Hi Nigel, Guys,
If the original chuck is as bad as you seem to suggest, then do as I do and keep it for rough and rusty pieces to clean them up for subsequent machining.
I think that you have already twigged that grinding, with a tool post grinder, is preferred to trying to turn the jaws by cutting them. You will probably find that there is some play in the spindle as well. Whilst the chuck end may not show any play the tail end could and this will allow the chuck to describe a wobbling motion.
Also machining the new backplate with your bad chuck may cause the backplate to have a machined in wobble. I would have done the bore and threads first then machined the plate when its screwed onto the spindle, then you can guarantee that it will be square to the spindle.Best Regards:
Baron J.
-
18th Jun 2021, 09:45 PM #7Diamond Member
- Join Date
- Aug 2019
- Location
- Revesby - Sydney Australia
- Age
- 56
- Posts
- 1,183
If your jaws are hardened, grinding is probably necessary (unless you have exotic boring bars with CBN or PCD inserts).
For soft jaws, or your average old steel like these, I guess it depends on how much metal you are removing. Best way would probably be to machine then grind?
Your variable runout is probably due to you measuring against stock of varying diameters. The runout of a 3 jaw varies with the position of the scroll. If you grind them true at a given diameter, they will only be dead-on at that diameter.
I have used good chucks that run out less then half a thou over their entire range.
Also, please decide on a unit of measurement and stick to it! Either thou or microns (pref the metric one).
When talking about the Hercus, with its imperial lead screw and thou-based slides, I usually talk thou.
However, I bent my imperial plunge DI, so have had to use the metric Mitutoyo recently. Hence some metric crept in
-
18th Jun 2021, 10:36 PM #8Diamond Member
- Join Date
- Aug 2019
- Location
- Revesby - Sydney Australia
- Age
- 56
- Posts
- 1,183
Nearly 1/4":
IMG_1615.jpgIMG_1614.jpg
Is the back of the backing plate running out ?
Did you cut the register at the same time you cut the thread?
IMG_1613.jpg
but that flat outer lip was machined at a different time to the middle.
Thread was done much earlier than the register.
The register was machined in several steps. Machined out to 1.46", then took off back plate and spacer, flipped over, tried on spindle, too tight, spacer back on, flip over and remount, machine a little larger, et c.
Given more time, there are things I could have done much more accurately;
- machine the thread tighter
- machine the spacer more accurately, or surface grind it
- mark alignment of the spindle, spacer, and back plate for better consistency
- borrow an internal mic, or telescoping gauges, to do the register
- use a live centre against the spindle to reduce spindle float
but, I thought the back plate was locating quite well on the spindle before I machined the chuck locating spigot/step/groove. I thought that was the only absolute requirement?
-
18th Jun 2021, 11:10 PM #9Diamond Member
- Join Date
- Aug 2019
- Location
- Revesby - Sydney Australia
- Age
- 56
- Posts
- 1,183
There is some axial play (unmeasured). Pushing up and down on the chuck is possible. When I was building the headstock, I could decrease it by clamping the spindle harder, but then I couldn't turn the spindle by hand!
For machining, light cuts should stay within the oil's float capacity. Heavier cuts might deflect enough for steel to rub on the iron headstock, but for the light machining of this back plate, I doubt it?
Haven't measured end float. A DI on the stepped shoulder (that the back plate screws up to) shows maybe 5 or 6 microns when rotating the spindle. Basically, within the limits of surface finish and the size of the ball on my lever DI. To measure end float properly, I would need to push it in with the tailstock, and then back with a clamp?
Run out on the flat part of the spindle was about 10 microns, so less than half a thou.
(It is probably now worse, from some light scratches that the back plate has left there. It is a tight fit, even with the smear of machine oil)
I did think of spindle float and run out, but decided that should average out in the final machining step?
(the spigot that the chuck sits in)
-
18th Jun 2021, 11:36 PM #10Most Valued Member
- Join Date
- Jun 2007
- Location
- sydney ( st marys )
- Age
- 64
- Posts
- 4,887
It would of been better to have finish machined the register and thread in the one setup, taking it off and on probably exaggerated the errors.
-
18th Jun 2021, 11:40 PM #11Diamond Member
- Join Date
- Aug 2019
- Location
- Revesby - Sydney Australia
- Age
- 56
- Posts
- 1,183
Indeed. It was the not gripping (not being parallel) that really needed fixing. Having less run out would have just been nice.
I think that you have already twigged that grinding, with a tool post grinder, is preferred to trying to turn the jaws by cutting them.
(basically, it is trading off a slightly inaccurate interrupted cut, against a nasty dust producing process that needs a flimsy little grinding wheel and shaft to fit inside the 20mm hole in a chuck?
You will probably find that there is some play in the spindle as well. Whilst the chuck end may not show any play the tail end could and this will allow the chuck to describe a wobbling motion.
Also machining the new backplate with your bad chuck may cause the backplate to have a machined in wobble. I would have done the bore and threads first then machined the plate when its screwed onto the spindle, then you can guarantee that it will be square to the spindle.
-
19th Jun 2021, 12:58 PM #12Most Valued Member
- Join Date
- Nov 2007
- Location
- melbourne australia
- Posts
- 3,228
Are you sure you don't have hardened jaws? Even my crappy old Chinese 3 jaw chuck has hardened jaws.
Isn't that what we are talking about here? I mean, if you're grinding your chuck jaws, it's pretty much a given the chuck is old and worn.Chris
-
19th Jun 2021, 06:04 PM #13Diamond Member
- Join Date
- Aug 2019
- Location
- Revesby - Sydney Australia
- Age
- 56
- Posts
- 1,183
-
19th Jun 2021, 07:10 PM #14Diamond Member
- Join Date
- Aug 2019
- Location
- Revesby - Sydney Australia
- Age
- 56
- Posts
- 1,183
Some improvements
I took the new chuck off my back plate, and did some measurements;
- Chuck body grabbed and centred in a larger lathe's chuck. TRO at both ends about 0.001".
Rear lip (that fits over the spigot in the back plate) about 0.004" lateral wobble??? - 2-3" micrometer, front face of chuck, to rear lip (that fits over the spigot in the back plate) are parallel.
- Grabbed 16mm precision rod (ejector pin) in larger lathe's chuck,
then clamped down on that with new smaller chuck.
Inside and back of rear lip also about 0.004" run out.
This seemed quite wrong for a new chuck. I ignored that for a while, and concentrated on my back plate.
Thoroughly cleaned thread on spindle and in back plate, and oiled spindle thread. Unmounted and re-mounted back plate several times. To reduce any spindle float issues, I push the spindle into the thrust bearing with the tailstock, a live center, and a tube against the spindle.
Back plate's chuck mounting face has about 0.001" wobble:
IMG_1621.jpg
so I spend an hour machining tiny fractions of a thou off that, chasing perfection:
IMG_1622.jpg
Sadly, only got it down to about 1/3 of a thou (10 microns)
Started with a round, positive rake insert (RPMW) and got it to the point where it was just putting a little dust in the chip breaker of the insert, but then had to change to a pointier insert for the corner of the spigot. My only sharp inserts are TNMG160402.
Somehow, a pass with that was cutting nearly a thou of wobble again. No idea how. Maybe their negative rake pushed the back plate further around on the spindle thread?
Gently filed and stoned a few high spots in the back of the chuck:
IMG_1623.jpg
Even dilled the bolt holes out a little, from 8mm to 8.1mm, just in case the countersunk bolts are distorting something:
IMG_1624.jpg
Bolted chuck on, not super tight, and clocked the front face. Nearly a thou run out, but really tightening the bolt near the high side got it very accurate:
IMG_1625.jpg
Now, to test some round stock:
IMG_1629.jpgIMG_1628.jpgIMG_1627.jpgIMG_1626.jpg
Fairly mixed.
6mm carbide rod was best, then 50mm machined 4140, then 12mm ground something, and 16mm ejector pin was worst at nearly 4 thou. Interestingly, the 6mm carbide was short - held in front third of jaws. Will have to inspect jaw surfaces more closely?
So this might be a crappy little new chuck
(that I can't get it any better than the crappy little old chuck),
but that seems unlikely.
Will strip and thoroughly clean it next. It might be full of that thick brown goo that manufacturers use for rust prevention
-
19th Jun 2021, 07:58 PM #15
Hi Nigel, Guys,
Two or three thou run out is not unusual for even a new chuck.
I got a new expensive 125 mm three jaw chuck and a commercial Myford backplate. I faced it off so that it was as true as it could possibly be, then machined the register 2.5 mm deep exactly to the size given in the instruction sheet that came with the chuck. This made it a tight push fit into the chuck back. Spotted through for the mounting screws. Drilled and threaded M8.
After bolting everything up I found that I had to machine about a thou and a half off the register because I couldn't get the chuck to sit flat on the backplate. Total runout with a 20 mm diameter test bar was around 3 thou and got worse as I used larger diameter test bars.
I subsequently found that whilst trying to get the chuck flat on the backplate that I had raised a burr on the inside of the chuck register and it was that causing the chuck not to lay flat on the backplate. Trouble was it was almost impossible to see and I didn't realise it was there until I got a cut on my fingers !Best Regards:
Baron J.
Similar Threads
-
4 jaw chuck
By outoftown in forum THE HERCUS AREAReplies: 1Last Post: 12th Aug 2020, 11:40 PM -
Repairing Jacobs No. 34 chuck.
By nigelpearson in forum METALWORK GENERALReplies: 5Last Post: 17th Jun 2020, 05:13 PM -
Wax Chuck !
By BaronJ in forum METALWORK GENERALReplies: 10Last Post: 16th Sep 2018, 01:17 PM -
collets : to chuck or not to chuck
By steamingbill in forum METALWORK GENERALReplies: 4Last Post: 26th Jul 2014, 10:17 PM -
New 4 jaw chuck.
By Alaks McGREGORE in forum METALWORK GENERALReplies: 25Last Post: 8th Jun 2013, 03:28 PM