[Caprirs]

Well, here is where I reveal how much of a machinist I ain't. I need to remove four broken bolts and one broken pipe thread fitting from a turbocharger. What started out as a project to make the car more fun turned into a royal pain. These will not be removed with Easy Outs or left hand cutters as the threads have been baked and seized together through a zillion heat cycles. All the threads broke while I was trying to "unthread" them.

I have a CNC VMC and plenty of carbide endmills. My question is: how do you center up over the broken threads? I cannot install thread inserts if I get into the original threads in the housing so I need to be sure that I'm really on center of the holes. What is the best technique/method for finding the center? Do I just eyeball it as close as possible? Once the bulk material is out, I know I can gently run a tap though to clean the remaining pieces out but I'm paranoid about damaging the threads.

Thanks,
Mike

http://www.mc-machine.com/images/rsturbo/6.jpg
http://www.mc-machine.com/images/rsturbo/7.jpg
http://www.mc-machine.com/images/rsturbo/8.jpg

[Geof]

All the holes appear to have a chamfer which is almost certain to be concentric with the thread. Take a fine file and carefully clean the flat surface so you have a nice contrast between the dirty chamfer and clean surface. Measure the approximate diameter of the chamfer and then take a scriber point or something sharp and bend it offset so that when it is mounted in your spindle it sweeps a circle a tiny bit bigger than the chamfer. Using something like a jewellers loupe now just line the spindle up with the chamfer using the pointer and rotating it by hand. The best way to do this is to work on each axis separately; get as close as possible on one then go to the other then back to the first. With a bit of luck depending on how close your scriber sweep is to the chamfer diameter and the magnification of your loupe you can get to within better than 0.005".

[widgitmaster]

That's easy, just put the mating part back on and dial in those holes!

[HuFlungDung]

You might be able to center up off the countersink around the holes where the stud is broken below the surface.

Be aware that a properly centered hole in the middle of a broken off screw thread does not 'look centered' because the helix of the thread is coming up one side and tends to make you think the hole is 'over here' when it really is not.

Honestly, I would not like to take the approach of drilling the stud out completely. I would drill a small hole right through, using a small drill press because I want to feel when the drill enters the clear space between the stud and the bottom of the original hole.

Spray some good penetrating fluid down the hole. Maybe let is sit overnight. Then try the easy out method. It may still not work.

Ultimately, I would drill maybe a 3/16 or 1/4" hole and use a small welding rod and stick weld down the hole, maybe 1/4" deep or so. I use a hi strength rod, UTP65, for this, available in 1/16" dia. Just weld around in the hole, and get it good and hot. The weld will shrink and pull the stud with it and break the rust bond. Continue to peck weld and build up the stud until it is as high as the top of the hole. Then, weld a little piece of keystock to it, put the vise grips on it and wiggle it out.

This method will leave the original threads intact.

[Caprirs]

Glad you guys chimed in. Thanks for the advice. I have little experience at this sort of work. Typically, I'm doing production machining from chunks of bar stock where I can scrap a few getting things dialed in and get to establish my own datums. (http://evlgt85.com/gallery/MC_Machine_Samples) That approach clearly doesn't apply on this.

The mating piece does not exactly have "nice" holes to work from either but that's a good idea. I think the mating piece is laser/plasma cut so the holes are pretty rough. It's also shaped such that the holes are difficult to access when it's in place.

I'm fairly sure an easy out will do nothing here. I've used them before and the way these bolts broke leads me to think they are pretty heat corroded into the turbo housing. The bolt material is relatively soft from being annealed by the exhaust temps the turbo sees. I think the easy out will just sorta gouge away at the inside of the hole I drill out be unable to break loose the threads.

I only have a wire feed welder, no stick. If I take it some where, I'll go straight to someone with EDM. I can also buy a new turbo unit for a few hundred dollars. But this was working fine before I decided to "improve" it and I hate junking stuff that is salvageable.

One of the bolts did come out cleanly so I can measure the original depth of the holes and also know the length of the bolt. That should allow me to go the correct depth without risk of going too deep. My concern is just knowing I'm on center. From there, I can patiently peck and circular interpolate. I have some carbide drills and plenty of small carbide endmills. I'm thinking Geof's approach might work if I can eyeball it in. I do have a 10X loupe. Also have the tiny tip for my indicator that might work instead of the scribe. I'm going to attack this over the weekend so any further help is appreciated.

[awright]

A backyard machine shop technique I used recently worked perfectly on a piece of brass pipe corroded into a cast iron steam radiator, but the procedure is labor intensive. Use a hand electric drill or a drill press to drill a small hole reasonably centered. Perfect centering is not desirable, as you shall see below. Then drill the hole out with progressively larger drills, shifting to 1/64th " increments in drill size as you approach the threads. If you have complete letter and fractional drill sets you can get less than 1/64th" increments for some steps. Clean the hole well and inspect carefully for signs of the threads showing up after each increment of drilling. Stop drilling as soon as you can see any sign of threads.

At this point, it's difficult to peck the remnant of the stud loose from the threads with a sharp prick punch because it still resists collapsing inward due to the full circles remaining in the threads. To make it collapse easier, use a triangular needle file to file two longitudinal grooves in the thicker side of the remnant until the thread barely shows at the bottom of each groove. The two grooves should be about 30 to 40 degrees apart on the thick side of the remnant.

Now use your sharp prick punch to knock the section between the two notches inward from the threads. The strength of the thicker metal left due to imperfect centering of the drilled hole is actually beneficial here because the piece will be strong enough to come loose as a single piece, rather than just the top thread breaking loose. Last time I did this (a week ago) I was able to grab the section knocked inward with needle nose pliers and pull the whole remnant out with a twist.

Depending upon your finesse, this can leave a thread essentially undamaged except for a small divot at the point you prick punched. Run a tap down the hole to clean out the corrosion and clean up any damage done. Strength of the threads should be unaffected unless you were really clumsy.

Good luck.

awright

[dertsap]

easy out would work if you use penetrating oil and heat ,ive managed to pull apart a lot of stuff like that with a torch and wd-40 ,
just let it expand with the heat and pour the wd to it ,youll be surprised how well that stuff will get in there

[Caprirs]

Well some success tonight. I carefully cut some aluminum soft jaws to hold the turbo. Centered as close as possible over a broken bolt. Center drilled and then drilled with a carbide drill. Since I had a hole now, I tried an easy-out. Not even close to budging. In fact, the top of the bolt buldged out enough to make the cast iron chip a little.

So I plunged an endmill down the drilled hole and circular interpolated a little bit bigger each pass. I kept shifting the center of the hole until it just started to expose the minor diameter of the bolt threads. The old bolt still would not come free. On the last pass (taking only .003" more), the remaining "coil" of threads finally broke loose. Yay. And then a millisecond later, that coil wrapped around my endmill and snapped it. Oops.

Ah well, I have more endmills and the part setup in the machine. Just have to center over the remaining holes. Thanks for the help. This place is awesome.

Mike

[greg b]

mike i remove broken threads almost weekly on all sorts of housings but unfortunately you need to know how to weld.providing the bolts have broken of near flush i weld the end of the broken stud with either a mig or tig welder and build it up about 3 mm then i hit the weld rather hard with a hammer to stretch the thread and loosen the rust then i weld a bolt or nut on to the stud and simply wind it out ,this also works on broken ezy outs that have a bad habbit of braking of especially in small threads .you may need to use a small amount of heat on stubborn bolts
hope this helps greg b

[hoss64]

greg b is right. I work in a tool and die shop, and we do the same with broken bolts.

[Konrad]

Easy out don't really work on small threads, because there is not enough "meat"...you are making the threads tighter, [wedging]

I have lots of luck by drilling out "dead center" to tap drill size and then unwinding left over coil.
But....most of them time..... a new trick I found out....drilling out in reverse, grind a drill in such a way hat it works in reverse, when reaching the bottom of bolt, the whole coil comes out!
On nasty looking bolts....loosen rust first by 2 or 3 times bolt expansion and contraction, using just the right torch for giving that bolt a quick heat, almost red, then cool, also using penetrating oil, ....a couple of times.
Konrad

[Caprirs]

Ha! I got the rest of the bolts out! And didn't break any more endmills. I centerdrilled, drilled, and lastly helically interpolated so the endmill wasn't trying to grab the remnant coil along the full length of the endmill. This worked although kinda slow. After each drilling and before the endmill, I tried using my easy-out. No go each time and I feared breaking the easy-out making more work for myself. Very patient I am when I have no clue what I'm doing. Fear of screwing up makes me go slow. At least I'm learning something in my old age.

Since I'm a dummy regarding this stuff, can you guys explain your pyromania obsession with heating the bolts? Is it just an excuse to use fire? (flame2) These are sitting in a turbocharger housing which sees daily temps of 1200*F for the past 22 years. The bolts are not truly rusted in. There is some other corrosion caused by the high temps and maybe something galvanic I assume? How does the high heat loosen the bolts?

Thanks for all the advice. I had to go with the tools I had and the method I was most comfortable with.

[HuFlungDung]

If you can heat a spot of steel redhot while the surrounding metal is still cool, the expansion of the heated area will be restrained in the direction of the surrounding metal, so it will upset in any direction where it is not restrained. This upsetting is a real (non-elastic) deformation of the metal. So, when the hot spot cools down, it will be smaller in some direction than it was to begin with.

In the case of a bolt where you drill it and weld inside the hole (or even if you could apply intense heat inside the hole it would help), the molten weld metal shrinks as it cools and pulls the shell of the bolt away from the base metal threads.

Hitting the hot bolt with a hammer will also force an upset, and the bolt will shrink as it cools. In my experience, drilling and welding the hole provides a deeper penetration of the upset effect and will effectively loosen more thread than just working on the top. However, the corrosion may be concentrated mostly at the top.

If the bolt is in an accessible boss, it will also sometimes work to weld a piece to the top of the broken stud, allow it to cool, and then apply high intensity heat around the boss, expanding the base metal away from the cooler bolt. However, heat conduction is quite rapid, and it is difficult to get enough differential between the base and the bolt to break the rust bond.

Oh, forgot... congrats on your success.