Regrinding a 5C Speed-Collet Chuck (Hardinge-Sjogren)

[@Sanny987]

Getting an old worn out chuck back to less than 1 thou of runout is extremely impressive. You should feel proud.

[@josephford8884]

I'm with sanny987. You used carefully calibrated methods to bring the TRO to half a thousandth. Bravo!

[@Trainwreck1123]

James, I have been watching your channel since your tool post grinder series so it's been a while. I have to say you have had an amazing improvement in your filming, editing, and content delivery over the years. Your videos are so polished and your content is top notch.

[@RambozoClown]

You could mount just the back portion of the chuck, and test the rear collet support area for concentricity. I would really look over that split in the chuck for anything out of the ordinary. Some 300 pound gorilla might have tightened the bolts enough to lift those areas. I have seen that on surface grinding jobs that have threaded holes.

[@steveggca]

Hi James as a former machine tool maintenance guy , nice work! as others have suggested perhaps a know good quaility 5c collet is the next step. A 3/4" hardinge collet from their online store is $75 lyndex is also high quality and about half that. at the end of the day you do get what you pay for. the ultimate test for a spindle (runout and alignment) is a solid test bar . there is obviously a limit to how good an old collet closer and a low cost hobby lathe can be There is such a thing as " close enough"

[@Crusher9mil]

One suggestion, since you corrected a considerable amount of runout by re-clocking the backing plate I'd suggest putting some punch marks on the two pieces for future reference. You don't won't to put it back together wrong way round again in the future.

[@Crusher9mil]

PS. Excellent work...

[@rpavlik1]

I barely took a breath the whole time but you appear to have taken approximately the same approach and I'm glad it worked! It's kinda crazy that you can do better in your shop than the result of factory plus years of wear. Pretty empowering even if terrifying

[@brandonshudy789]

That looks like a great candidate for a milled slot and a spanner wrench for install and removal! I’m inspired but your willingness to make all your tool’s perfect.

[@Andrew_Fernie]

Or even just a drilled hole or three

[@sierraspecialtyauto7049]

Dang, James. The more you improve this chuck the more I wish I'd bought it from Dale when I had the chance, and done a D1-3 to D1-4 conversion to match my small lathe.

[@Awegner176]

Had the same exact problem with the same chuck. Sold it, priced accordingly, to somone with the patience to bring it back to life. Ended up just going with a collet handwheel and spindle bore adapter. Great video!

[@AndirHon]

I've never attempted anything like this. I've also never done a lick of machining in my life unless you count sanding a wood block in a drill. I'm just here because I find this entertaining.

[@mikemarriam]

Hardinge (Elmira, NY) guarantees their collets to .0002 TIR. They're a little pricy but it might be worth getting one and seeing if it would be worth investing in them instead of trying something else. Harding collets are one piece. Most imports are two. That may be the difference.

[@rorylong314]

I thought of that too. They’re $51 for standard metric and imperial sizes. Well worth it

[@JFirn86Q]

Only the Hardinge "Special Accuracy" are 0.0002" TIR (about $85 each!). Hardinge Standard 5C collets are 0.001" TIR! So if he is getting under 1 thou runout now, it is definitely the collet doing that being it is an import collet.

[@mikemarriam]

​@@JFirn86QI only buy the special ones. With my mill they are effectively perfect.

[@JFirn86Q]

@@mikemarriam Same, just making people aware that there is a difference.

[@chauvinemmons]

Guessing from the measurements he's getting over the pins I would say the collets are not at issue in this situation I'm assuming the collets must have come with this Chuck meaning they're old meaning they're high quality unlike the China trash we are being force-fed. Even quality brands have their junk made in China.

[@squelchstuff]

Nice fault finding and rectification James. Having speculated on the adapter mating faces being affected by torque previously, I am pleased to know I was wrong. Thanks for the journey.

[@jimdean7335]

I’m out of breath just watching this. Your persistence has paid off. Thanks for sharing.

[@pauldorman]

If you were confident in your headstock alignment, I wonder if you could improve the result with an aluminium lap on a mandrel in the tool holder. The lap could be accurately made between centres, which could also be used once the mandrel is transferred to the tool holder to ensure accurate alignment. Once positioned and the cross slide locked down, move the saddle out of the way, install the collet chuck, apply some lapping compound, and you're off! The only complex task would be the transfer, but I think you could measure any misalignment with a test indicator or two to show deflection when you clamp it down.

[@sblack48]

I would never have guessed that a little pencil grinder would be rigid enough to give such a good finish. Well done

[@boomermatic6035]

Awesome video on troubleshooting a problem and diminishing returns, this should be a required video for everyone, as the concepts can be applied to any field.

[@coreyb4073]

youre regrind looks promising, I believe you are on the correct track as far as the next step, first i would suspect the collet, after that the back journal. you will find with the back journal you can lap it to deburr it, there is usually quite a bit of damage back there so the most important thing is to get the high spots down. dont forget, the .001 your seeing at 1/2" off the face would be 4x that on the back journal. But i would suspect the collet above all else

[@samir5740]

21:47 boy, that was satisfying, that needle was nailed in place, soooo good to see it not moving. I saw somewhere that if you mach two machined surfaces to each other, like that collet is matched like a dream to that chuck, it will be a bag of cats in any another chuck.

[@shannonsears3496]

Great job with the pencil grinder. I have been using them in a V block for years, mostly to true the face of diamond and borazon wheels. A shop i worked for used a pencil grinder on a cnc grinder for a very high speed head. Collet on pencil grinder not running true was the biggest problem.

[@hersch_tool]

Very methodical and logical measurement process. I learned something, that's for sure. Thanks James!

[@Not_My_Name5912]

It would be interesting to see what the difference is with a quality collet. good job as usual👍

[@Daniel_en_el_trabajo]

I want to see that, or mayby other metod of holding that can have better accuracy

[@powerbuilder0510]

spin up a solid bar between centers that otherwise is identical to a collet but without splits that wedge it in the chuck. then since it wont collapse and shrink into the chuck any error in the chuck mounting surfaces will be directly visible in solid collet once its pulled up/in and mounted. the solid collet will act like a gauge pin or other such things, but be instead of parallel or triangle shaped the end of it mounts as a 5c collet.

[@g.tucker8682]

I thought you software guys always number three things as: 0, 1, 2 As always, highly educational. Thank you

[@gbspikyfish]

I tried that, but I couldn't see where I'd made zero punch marks ;)

[@theradarguy]

Good job. My first dovetail was not too long ago when I bought an older but higher end lathe than I had at the time and it was missing the compound and the cross slide lead screw. $200+ of cast iron and a lot of patients.. Thanks for the video.

[@grantsaviers6557]

Several comments - nice effort. I would suggest getting a reference 5C collet. Hardinge makes special super accuracy collets, they have 4 slots and the best TIR they offer. Not cheap and often used on tool grinders (I have a set for a Cuttermaster). Then a straight and round pin is needed to check TIR at the nose and as far out as the pin length permits. I've found that drill rod and drill blanks are neither round or straight enough when in the tenths world. Dowel pins I've bought from McM are straight, round, and are relatively cheap. (it kinda makes sense - bent dowel pins don't work). A set of these then can be used to measure 5C and R8 collets. A set by 1/8" steps to 3/4" won't break the bank, I bought the longest offered. Subsequently, I threw out a lot of R8 collets and some 5C. An R8 MSC brand set won, but I don't have any of the best brand name R8's. Then with your surface grinder you can tune up v-blocks to match and have the V bottom parallel with a dowel pin. It's fun to do that.

[@arthurmoore9488]

Thanks for the video. I'm restoring an old lathe and need to fix the MT3 taper on the spindle. This is perfect.

[@Jesse_The_Enchanter]

I’m not sure if someone else mentioned this or not… When I installed a collet chuck like this for a shop I used to work at, the chuck was adjustable like a 4-jaw. Brand knew the runout was .005”. I adjusted it to .0002”. It took a bit of time. You have to loosen the bolts in front and adjust it in like a 4-jaw and tighten them.

[@kevind1865]

You're likely at the runout limit of an old 5C collet. Even a brand new Hardinge 5c collet in a master spindle will be .0002-.0003" runout. I'd call .0007" with an old collet pretty dang good for a reground spindle. Well done.

[@TheAyrCaveShop]

Well done James... Takes a lot of patience for that detailed analyzing...definitely paid off. Enjoyed....now its ready for action... ATB.....

[@urbanawoodproject3123]

Great analysis and result here. One thing that I think could have been interesting would be to measure the runout using a few different sizes of gauge pins. In theory, the larger gauge pin (just one or two thou larger I would think) should run truer since it is forcing the collet to contact the taper further out near the end where you've got the grind perfect. Whereas a smaller gauge pin would probably show a much worse runout.

[@Clough42]

Phil (Almost Machining) suggested this to me as well. It would be interesting to try.

[@miguelfernandez2325]

Every time I watch a video from you I learn something. Fantastic

[@benmartin9648]

I just got one of those today and will be trying to bring it back to life. Thank you for the videos!

[@mcorrade]

good call. I can't tell you how any times I went just a bit to far and blew it. nicely done

[@danarrington2224]

It's worth noting here that gage pins are not always perfectly round. Some pin classes are intentionally tri-lobed. I would suggest that you get a dowel pin and test the runout to see if you get the same results. I do think you taught a very important lesson with this video. Always take notes when you are disassembling something so that you can put it back together in the exact same way. It's especially easy to do these days since everybody has a smart phone with a camera in their pocket.

[@centurialinc]

We had tons of issues with our R8 spindle and collets for our Grizzly cnc mill conversion. You have to keep going! We where never able to eliminate .001" run out. We did everything you did and more. And just never figured it out! Try lapping the collets. We tried that with no improvement. Buy a precision collet. We did that with no improvement. Try to re grind the collet. We tried that with no improvement. It was just impossible to find the error. in the end we had to try to cancel out the run out by hand for each tool. We also found the bearings in the head had some runout as well and replaced them with no improvement. So you have to keep going. Find a way to get down to .0001" run out for all of us and show us how you did it.

[@stephenjohnson6841]

Another great project/video James! Very well done as usually! Thanks again!

[@edsmachine93]

Nice work. I liked your process of checking. Thanks for sharing the process. Have a great day.

[@GBWM_CNC]

I literally love watching people chasing microns! Or the tiny little inchies

[@larryschweitzer4904]

Late to the party BUT I've struggled with a similar issue and resorted to mounting an air die grinder. I got closer but never did get it perfect. Lots of time in it. Maybe I'll go back to it later.

[@DeserdiVerimas]

I thought that for sure you'd be using the toolpost grinder you made in order to grind in the taper!

[@outsidescrewball]

Enjoyed…great discussion and methodically sleuthing to chase the issue

[@DonaldRyan-w3o]

Nice job on the chuck. As others have said a high quality collet would make a big difference in what you are seeing for run out. Also why don't you drill a hole in the outer dia. of the adapter to fit a .500 dia. bar to tighten and loosen the chuck on the spindle instead of the strap clamp?

[@TheUncleRuckus]

It may not be perfect but it's a hell of a lot better than it was, great job James! 👍👍

[@garysgarage101]

Very nice. I agree, unless you go with high precision everything, you’ve achieved a great result. Excellent content as usual.

[@rickdenney5772]

Others have mentioned the need for a known-good collet, now that you are down in the range of where the import collet errors may be dominant. If the taper measures concentrically, then there's nothing else to do with the chuck for now. Of particular concern is that the collets are only contacting the chuck at their split points. That tells me that the collets are slightly oval in a way that tightening is not taking up.

[@mertonsilliker3686]

Plan your work, work your plan. Good job.

[@billshiff2060]

The remaining error close to the chuck has to be the collet/material but farther out is the register inside. Try a long indicator probe to measure that runout deep inside before doing anything. Nice to see inside one of those. We have one but I don't particularly like it. I like the back of the headstock locking lever type. But the very best is the Royal air powered collet closer. I like those because I like to change parts on the fly without stopping the machine.

[@clintchapman4319]

Awesome Bud! Way to stick with it!

[@AmateurRedneckWorkshop]

I liked your trouble shooting steps very thorough. Thanks for the video keep on keeping on.

[@ronhaefner7833]

"Call-it" done. Well done.

[@MoraFermi]

At this point, grinding the collet in might be the better option. Great work!

[@PBVader]

Impressive, axial runout could be futzed with in the backing plate, but then it would be back to the grind. Much better than truing a 3 jaw chuck.

[@remoman]

You can build up material inside the taper by getting it chromed up to 1mm thick if you want. That how spindles are repaired and reground.

[@joemcgarry1106]

James, I would have liked to see you use your shop made tool post grinder, with a Norton 46 grit amumiun oxide wheel. have not had much luck using pencil grinders for prescion work, the bearings are too spongy. The job came out really good in the end, due to your skill set. 👍

[@ADBBuild]

I don't think he can mount a small enough wheel on his tool post grinder. It's really only good for outer diameters.

[@joemcgarry1106]

@@ADBBuild I went back and viewed his build video. You are right, as he as no provision to use smaller mounted stones. It looks like he cold adapt an er-16 collet holder to the spindle. Maybe he will look into that project in the future. 👍

[@AlmostMachining]

Great job James!

[@Jake-zc3fk]

Interesting. I’m curious what happened to your tool post grinder build and why you are not using it here? That tool post grinder looked like it turned out nice!

[@amundsen575]

great job, get a new premium collet and test, check the back seating surface , have it built up with spray weld or use moglice epoxy or press a sleeve in and regrind as this chuck is no longer in a production environment im sure you you can get better as spindle bearings show very little runout

[@e.scottdaugherty8291]

I concur, to quote Adam Savage "You're just chasing Zero's", after a bit of this. The results invert with time to progress." {I probably mangled that in my memory, after a time chasing zeroes, you don't see much progress on a LOT {in cost too}, mostly it's time. I have confidence in you gettin' it all to nothingness. 😁

[@chrisstephens6673]

I made the mistake of buying a cheap import 5C collet holder, thinking how bad could it be. To cut a long story short, after making a bump-tru back plate and a whole of other fiddling, I drilled the chuck body to intersect with the inner alignment surface, inserted some hard bronze pieces and bored it true. For a while I got virtually zero at the collet face and at 150mm from it too, but as I say only for a while and only with one of my genuine collets, the import ones can be "variable".

[@dlstanf2]

I remember watching @MuellerNick, years ago, bring a taper back into spec by scraping. The video is still on KZbin.

[@mjwint]

Nice methodology in determining root cause. I was a bit disappointed, look for the 3D printer grit extractor for the bring job :).

[@matsbengtsson3756]

The chuck looks perfect. I bet the remaining error is in the collet. Try a few different collets.

[@melgross]

Remember that two different collets of different sizes has the exact same errors, so it’s not likely to be the collet.

[@gangleweed]

I came across this phenomena many years ago with the run out in a 3 jaw chuck on a dividing head and the guy trying to get a mandrel to run true by resorting to hitting the back end of the chuck jaw with a copper punch in the vain attempt to get it to move a couple of thou..........without asking permission I re-machined the backplate register to give it a .005" clear and the rest is simple to get the part to run true.......this trick will work for any 3 jaw chuck when you just must have the part it's holding to run true.

[@sferg9582]

I think you have gone as far as a 5C will let you, given the straight diameter at the threaded end of the collet, and there being no way of correcting the adapter diameter to bring it concentric. Good job on the ID grinding, came out nice.

[@airgunningyup]

nice work , troubleshooting chucks can be very time consuming.

[@DIAMOVEVO]

Thank you 💞

[@raysimon1368]

I use to use a cutter grinder spindle to grind in tapers never did i ever get runout less than .002 with a import 5c collet i purchased a set of hardinge high precision 5c collets i thinkyou got it better than you should have with import collets how many times are you going to need it any better but i understand why you do because but is it worth 2500 dollars for a set of hardened and ground precision collets i think you got a great collet chuck now i would bet the bearings in the spindle of the lathe will not hold 0008 for long unless you have already replaced them too you do great work always a pleasure to watch your videos from a 50 year retired machinist

[@tonypratt1989]

James, please get a lever DTI, the use of a dial gauge in this case is painful!

[@simonconvey9645]

You were wise to stop right there. I *think* that you can buy single collets from hardinge with gaurunteed specs.

[@seldendaniel8819]

Nice work, and inspirational.

[@Motocue912]

Vary nice job! I would call that good . I’ve been a machinist all my life and I build pool cues for a living. I have ground many chucks and you can only go so far before you start backpedaling. 😢

[@paulquentin9387]

You could go the swiss way, and get a morse taper reduction sleeve with a set of w12 or W20 collets to fit directly into the spindle and toss the 5c collet chuck...

[@paranoiia8]

Like always great video. But in another hand I was wondering why you dont make hole in that adapter for removing it(as you can just put inside pin/screwdriver and rotate it), so you would not need to use strap clamp. Or that would mess balance of adapter too much?

[@zachaliles]

Since the adapter plate doesn't need to be in a certain position I would have put it back on in a different position. That way when you finally part ways with that chuck, for whatever reason, it will really confuse whoever gets it after you.

[@mikeydk]

I have tried to drop a part on a finger. It only fell 10-20mm, and the part was not long and only about 50mm in diameter, but on the underside of my finger was the sled on the lathe, which didn't didn't budge... could feel that for a week

[@charlieromeo7663]

Nice job, James. Just curious why you would do this with an import collet? Would it be worthwhile to grind the collet somehow?

[@CaskStrength777]

If you wanted to get better life out of this, doing what you did, and then having the item hard chromed, and then mounting between centers and hard grinding a thin coat of the hard chrome with something like diamond wheels would give you a true surface, and on a material substrate that is extremely hard compared to the actual tool body. This would wear much slower than the tool steel body. This would also let you repair something even if it's worn out otherwise, you just grind the new plated surfaces. I do this kind of work daily, but not on tools.. excellent overview of dressing grinding and going through the errors and the glazing that happen during rotary grinding

[@stevebell1128]

Nice job. More work but for accurate common sizes possible to internally grind few collets in situ? Even more work.....but if lathe spindle bore big enough make a drawbar collet chuck? Could register in spindle bore and thread on outside to hold in place perhaps. Got drawbar collets mounted directly in spindle on 80+ year old leinen lathe and run perfect and zero overhang from spindle bearing so nice and rigid, much better than Crawford d1-6 collet chuck on other lather (but obv larger lathe so likely to be worse I guess)

[@chrislee7817]

Do you have any other collets to check against? Would be interesting to see if there is a difference. Nice to find some high quality 5C and see the difference.

[@richardcurtis556]

The. scientific method. strikes again. Nice work.

[@crichtonbruce4329]

Imports... I bought a Morse taper #3 to #2 adapter from an importer were the inside was visibly off center producing ridiculous runout. I got it exchanged and the new one was even worse! Must have been a bad batch from the factory. The supplier gave me a refund without question. Anyway, I also suggest you borrow or buy a quality collet for your testing.

[@erik....]

I ground my old milling machine taper a few years ago using a small tool grinder head, there's a short clip on my channel.

[@flikflak24]

Could be fun just for ship's and giggles and try doing some 5c shrink fit ( just to see if it was possible and how well it would work

[@mrwolsy3696]

James, nobody sells iso30 spindles for the RF45 style mill, could this work?

[@ChrisHarmon1]

Cant tell you how many times ive had my knuckes smashed into the ways when knocking a chuck loose with rubber mallet. You always try to be ready but somehow we never are.

[@WoutRikkerink]

Could it be the grease on the mating surfaces that creates the runout?

[@romualdaskuzborskis]

Now this is quality content! But, sir, please, oh please next time use tens indicstor, its helluva eyesore trying to look in to those 0-3 thou idscrepancies..

[@tjuggernaut29]

Hi! Do you mind providing the model number for the NSK grinder? Thanks!

[@pauldorman]

I think it is model US$2000 😁

[@MF175mp]

I would assume there to be a ding in some of the mating parts of the chuck itself if there's runout. It's unlikely that the whole chuck would be bent in an accident. Another option is that someone just reground it to run true in their out of whack spindle. Did the body of the chuck run true?

[@MF175mp]

I would also assume the rear cylindrical fit might be running out some as well. You don't seem to have a DTI, especially a long lever version to check it?

[@MF175mp]

The error in the angle you first had came probably because your indicator wasn't exactly on center height while indicating the angle in. It will skew the angle a little bit that way.

[@Clough42]

It's possible, though I would expect the error to be in the other direction if it were high or low.

[@MF175mp]

I would turn a precise 5C shaft holding it in some other chuck, then without removing it chuck the 5C chuck in backwards. Then re-turn the mounting face and taper of the chuck. Because you can't regrind the cylindrical fit without making it loose and this will ensure it's all running true. Provided it's tight enough fit in the cylindrical part. Could be worth making there a short expanding mandrel in the place of the cylindrical fit, to cancel any movement from the fit. But very short to keep it solid, it doesn't need to expand more than a thou or so maximum.

[@Clough42]

@@MF175mp not a bad idea. Even if the taper isn't perfect, it'll be close enough that touching up the front taper again after the fact would remove any remaining imperfection.

[@Larry537]

What runout does the O.D. body look like?

[@rickmellor]

Proper procedure includes getting grinding dust in your face. Good job!

[@martinbeaumier7172]

Didn’t you make a grinder for your lathe a while back?!

[@jukkapekkaylitalo]

collets were riding on split points. Y might want to lighten those points on your cheap collets. So that the load point is divided more evenly through collet chuck. Your collets just going to grind and scrape your chuck again.

[@alevgsid]

Биения в районе 0.01 дюйма (0.23 мм) можно увидеть невооружённым взглядом. Для этого не нужен индикатор. Биения такого порядка означают, что у вас нет станка, а есть просто крутящаеся нечто.

[@sierraspecialtyauto7049]

McMaster-Carr sells pretty good collets at a pretty good price.

[@npalen]

James, wouldn't the OD of the stone be parallel to the compound travel regardless of which side it was dressed on?

[@roflchopter11]

Yes IF the grinding spindle axis is parallel with the compound. If not, the stone is actually cut into a cone. And the "projection" of that cone will only line up with the compound on the "side" of the stone that's dressed.

[@g.tucker8682]

That's what I was wondering - the stone only has one "side." Edit: I was wrong yet again, see below.

[@g.tucker8682]

​@@roflchopter11Ah, gotcha

[@roflchopter11]

@@g.tucker8682 very good. If one really cared, one would also need to dress the stone at the height of contact with the work (which may or may not be the centerline height of the lathe).

[@npalen]

You're right. Guess a person should misalign a bit in the right direction and then dress on the opposite side to get a slight taper making a stone slightly thinner towards the back. @@roflchopter11

[@Compasscard]

That NSK grinder costs the price of my tiny lathe...didn't expect it to be that expensive

[@JFirn86Q]

Even Hardinge's branded collets are TIR 0.001" ($51 each)... so those cheap import collets are never going to be better than 0.0005" at the best. If you are at or under 0.001" of runout that's most likely the collet at this point.

[@wls3118]

Another great video!

[@BirgitWallmuller]

The chuck needs an adapter backplate anyway, so how about a new backplate which is is slightly off center (by as much as the the back surface is on the current one), so the error there is canceled out. the taper in the front would have to be reground this way, but the collet would sit perfectly straight in the chuck. alternatively a new backplate, but the back surface is machined on the chuck (ie. the regrinding you mentioned, but as it would only be on a new back plate, it is on a disposable part and even on a failure a comparison with the current back plate should give better measurements how it should have been (to machine a third one with that off center error measured before). and if it goes well....

[@Clough42]

I've been wrestling with all of these possibilities, trying to decide which is best and if it's worth it.