PDA

View Full Version : Machining troubles (i dont know how to be less ambigous)



Fasttrack
11-26-2006, 10:37 PM
I've been working on this mini-bike project. I wanted to have a two speed manual tranny in it and i ran into some trouble in the process. I need to drill four holes about the bore of a gear hub and ... i'm not sure what to call the other piece ... anyway the four holes have to match up no matter how the two pieces are turned (as long as they're turned 90* of course). The first side of the part with out a name and the gear hub that will mesh with it came out perfect. I can turn 90* and have all the holes still line up because they are all the same distance from the center.

Well i thought i was doing everything right but the other side didnt turn out as well. When i turned the hub 180*, two of the holes were about a half a hole off. Have you guys got any tips for this type of procedure? (i hope these pics will explain this a little better)

http://i108.photobucket.com/albums/n22/fasttrack237/mini-bike/P1010952.jpg

http://i108.photobucket.com/albums/n22/fasttrack237/mini-bike/P1010956.jpg

http://i108.photobucket.com/albums/n22/fasttrack237/mini-bike/P1010957.jpg

http://i108.photobucket.com/albums/n22/fasttrack237/mini-bike/P1010955.jpg

Luckily this is the threaded side so i can make a new hub and threaded piece pretty easily.

Fasttrack
11-26-2006, 10:39 PM
A few more assorted pics relating to the project...

http://i108.photobucket.com/albums/n22/fasttrack237/mini-bike/P1010959.jpg

Thats my centrifugal clutch broken apart and my rear sprocket/hub/brake drum. The sprocket pops on and off the hub so i can change it out for bigger or smaller.

http://i108.photobucket.com/albums/n22/fasttrack237/mini-bike/P1010958.jpg

Thats the clutch assembly back together.

BadDog
11-27-2006, 12:26 AM
I don't know what the "right" way might be, but I'll take a stab at it.

I would probably fixture them all together on the shaft and locked relative to each other. Then mount on rotary table or otherwise find a way to accurately index at 90*. You could use a gear and rigid index finger and rotate on the shaft. Rotating on the common shaft guarantees the same radius, and the gear/index guarantees 90* orientation.

Paul Alciatore
11-27-2006, 01:11 AM
The trick is to use some form of indexing. A rotary table, a dividing head, or one of the less expensive spin indexes It will put them dead on 90 degree increments. Of course, you will need a means of mounting your parts on the index device of your choice. Chuck, collets, clamps, etc. and centering is important. First you center the rotary table below the spindle. Then you offset it by the desired radius. Finally, you must center the part on the rotary table.

BadDog
11-27-2006, 01:20 AM
Just realized, that stack-up would be a bit much to drill in a single stack, so amend that as required...

Fasttrack
11-27-2006, 07:22 AM
I wish i had a rotary table or spin indexer... :D

GKman
11-27-2006, 07:44 AM
"not sure what to call the other piece "

A spool?:cool:

Your Old Dog
11-27-2006, 08:31 AM
From the list of "Stupid ideas I haven't tried yet" I offer the following:

get a small birthday candle and light it. Hold the item you want to replicate at a 45 degree angle and play the candle flame on it so that soot is deposited on it.

Take a piece of Scotch tape or clear packing tape and position it over the sooted area and carefully lay it down. Wipe across the tape surface with your fingers so that the tape picks up most of the soot. Now lift the tape up and you have a positive image of the bolt pattern. Carefully position it over the gear you want to drill then center punch by eye over the middle of each hole.

Is it as accurate as the methods described above? Hell no! Is it accurate enought for a go cart? Yeah, I think so. When you are done you will have made what is called a smoke print.

If you want to make a negative image just put another piece of tape to the first piece sticky side to stickey side and pull apart. Enough soot will stick to the second piece of tape to produce a negative image.

Gun engravers put a very thin whiteing solution (similar to "WhiteOut" for typist) on the metal and then coat it with a trace of Damar Varnish to make it sticky, this makes the soot stick and then they can remove the tape for even more accuracy. Spray it with artist fixitif as used by charcol artist and it's permanent until you hit it with denatured alcohol.

Sounds more involved then it is. I can smoke print and transfer your holes in maybe 1 1/2 minutes tops. This technique is accurate enough to lay a fingerprint.

A.K. Boomer
11-27-2006, 09:06 AM
I would use my DRO, it has bolt hole calculator, if i couldnt do that for some reason then I would use my rotary table, if I didnt have a DRO with bolt hole calculator then I would just use my DRO and do the math --- a bolt hole pattern of 4 is about as simple as it gets,,, If I didnt have a rotary table or a dro then i would use the handle scales, there is no reason that you should be a half hole off unless youve made a drastic mistake,,, One; center the part and zero your scales or handles, two; take off in the direction on the x axis that you want to go for the radius, three; do the same in the opposite direction and then return to center, repeat all these proceedures for the y axis and you should be done,,, Big fat note; If your working on an old mill and have a ton of slop in your leads you can blow a simple operation like this by forgetting which way to keep the handles loaded, I cheated when i got my mill and didnt even want to mess with learing how to machine without a dro --- it takes all the guess work out of it.

Evan
11-27-2006, 09:36 AM
For marking out 3 and 4 bolt patterns use the lathe to hold the work and index from the chuck jaws to the bed with a stick or other appropriate stop. Scribe the work using a pointy tool and run the cross slide in/out to make a mark. Once the radial lines are scribed then scribe the bolt circle by precisely placing the tool at the correct radius and rotating the part. Lathes are very good at this. :)

Carld
11-27-2006, 10:08 AM
It looks like what you are building is what's called a dog clutch. I assume that there will be pins in the sliding part and the pins will engage one or the other gear when the center spool is slid one way or the other. The holes that the pins will enter in ether gear will have to be larger than the pin.

Now, take your calipers and measure across the holes from #1 to #2, from #2 to #3, from #3 to #4, from #4 to #1. Those measurements should be the same but for your purpose can be +/- .001 maybe .002. You should take measurements across opposite holes as well and from each hole to the edge of the bore. This will establish if you have a bolt pattern that is even and centered on the axis. When you find a set of holes that are as near perfect as posible then use that one as a templet. Slide the part to be marked on a shaft and slide the templet on top of it. Now clamp them so they can't move and take a transfer punch and prick the four spots through the four holes. Now you have transfered the holes to a new part. Set it up in your mill and using a wiggler, line up the spindle on each prick mark and drill them out. Continue the above process untill all the holes are done. Don't forget to oversize the right holes for the pins to slide into.

The problem with these type shifters is that the pins will loosen and fall out over time. Also, it is tempting to try shifting on the go and that will cause problems. It would be better to machine bosses on one part and recesses on the other part to engage with each other.

BobWarfield
11-27-2006, 10:10 AM
I wish i had a rotary table or spin indexer... :D

You've got the gears, one is bound to have the teeth such that you could use them to index the 90 degree positions. You can make a fixture that works with the gear to index with. Take a spare gear and you have a piece of fixturing you may well use multiple times. You just need to contrive a way to center your workpiece on that gear and hold it in place while the gear rotates and is indexed with some sort of pawl.

Best,

BW

Fasttrack
11-27-2006, 04:47 PM
Thanks for all of the great ideas! I hadn't thought about using one of the gears for indexing.

I like Evan's idea, but how would i index for a four-bolt hole pattern on a three-jaw chuck?

"The problem with these type shifters is that the pins will loosen and fall out over time. Also, it is tempting to try shifting on the go and that will cause problems. It would be better to machine bosses on one part and recesses on the other part to engage with each other."

That occured to me the other day after i realized my holes didnt line up on the one side. I thought - hmm why didn't i just make a straight boss and a straight recessed area... i guess i just didnt think of it in time. I'd like to salvage what i have here if possible. I guess i could still cut some bosses in it but it would look hoaky with the four holes drilled in it too.

My idea originally called from springs between the ring and the pins so that the whole "spool" would slide to contact one of the gear hubs. I had intended to cut a little cone and chamfer on the spool and hub, respectively. This way i was hoping to get the rpm of the shaft and the gear to match fairly close before attempting to slide the pins into place.

The more i think about the more i want to do the boss/recess idea.

Well thanks again guys!

Wirecutter
11-27-2006, 05:29 PM
I would use my DRO, it has bolt hole calculator, if i couldnt do that for some reason then I would use my rotary table, if I didnt have a DRO with bolt hole calculator then I would just use my DRO and do the math --- a bolt hole pattern of 4 is about as simple as it gets
I'd have to say the same. I have a centering indicator that allows me to get the mill spindle aligned to the center of a hole, too. Once that's dialed in, I use the DRO's bolt hole function.

I paid about the same amount for the DRO as I did for the (obviously not new) mill, but I'd do it again. That thing is just great for getting exactly where you want to be, and quickly.

-Mark

topct
11-27-2006, 05:40 PM
Fastrack,

I would have used an existing transmission.

I know it would not be as much fun, but accuratly duplicating the shaft centers on some plates to hold it together and figuring out some kind of clutch would still be quite an exersize.

And you would not be wasting your time trying to reinvent the wheel.

jdunmyer
11-27-2006, 06:20 PM
Personally, I'd drill the holes on the coordinates, even if I didn't have a DRO. As others have pointed out, a 4-hole circle is dead-simple.

For those of you who don't have a bolt-circle function on their DROs (like me), Machinery's Handbook has tables of coordinates. Look in the Jig Borer chapter, it's near the front in my 22nd Edition. As long as you're paying attention, the bolt holes will line up perfectly every time. The tables go up to 28 holes.

RPM
11-27-2006, 07:18 PM
When I was making my own replacement dials for my old Atlas lathe, I used the face plate as a pretty accurate indexing plate. you can lay out 90 degree lines using the existing axes for bolt holes, or you can divide up the circumference using a trip of paper taped to the edge. Didn't believe I or the lathe could do it, and it was so easy and so (for me) accurate I was amazed. Many bull gears are also pre drilled for indexing, my one is already divided into 60, with a special detente pin to lock in the setting. The bigger the circle you are using for this, the more accurate your spacing is going to be. If you'll excuse the terrible pun, there's no real need to re-invent the wheel :-) Good luck!
Richard in Los Angeles

RPM
11-27-2006, 07:37 PM
Dear Fastrack
Sorry, but with my 'brilliant' advice was only worth something assuming you had a lathe? Doesn't everybody? Only a mill?
Richard in Los Angeles

wierdscience
11-27-2006, 07:51 PM
Bob has it,use the gear itself to index the holes.The one pictured counts as 36 teeth,36/4+9,every 9th space you have a hole.

Fasttrack
11-27-2006, 09:02 PM
I think the gear idea is the way i'll be going!

"Personally, I'd drill the holes on the coordinates, even if I didn't have a DRO. As others have pointed out, a 4-hole circle is dead-simple."

Thats exactly why i picked four holes instead of five or three. I figured, since it is four holes spaced 90* apart, and the y axis and x axis move perpindicular to each other, it would be super easy. Like i said, the first one went off fine with out a hitch. The second one - well it wasnt so hot. I think maybe i didn't have my spacing from the center the same for each hole. I can't figure out how i screwed up the 90* otherwise because it stayed in the vise the whole time.

What i did was put a little divit in the center of the hub/spool to be drilled. I then used a v-block to hold it in the vise, used a transfer punch in my drill chuck (which has a surprising amount of run-out in it ... :( ) and then moved the carriage the radius i wanted to drill away from the center. After drilling this hole, i returned to the center and moved the carriage the opposite direction the same radius. Then, i returned to the center again and moved out, this time in the y direction, the radius i wanted to drill. Needless to say it took awhile with all the changing from bit to transfer punch.

I think maybe that the run-out in the chuck had something to do with it. My spindle wasn't always centered on the work depending upon how my chuck was oriented. That would throw the distance from the center that my holes were. I think thats where my problem was.

Fasttrack
11-27-2006, 09:05 PM
Fastrack,

I would have used an existing transmission.

I know it would not be as much fun, but accuratly duplicating the shaft centers on some plates to hold it together and figuring out some kind of clutch would still be quite an exersize.

And you would not be wasting your time trying to reinvent the wheel.


If i actually needed a mini-bike for some reason - i would have made a serious attempt at finding an exsisting tranny. I've got my truck to work on and a pretty fast go-kart to have fun on so this is just one of those "lets see if i can do it" projects.

The clutch i'll be using is actually a centrifugal clutch salvaged from an old go-kart (and salvaged is the right word because it was all screwed up, but a bit of machining and creativity and i had something to mount on my lawnmower engine that will feed into the tranny. The release and engage time on a centrifugal clutch wont be as hot as a normal one so i'll have to put up with some soggy shifts as i take it out of gear and wait for the clutch to disengage to put it back in gear...)

wierdscience
11-27-2006, 09:23 PM
If your using the mill ,then setup a DTI and use it to positon the table instead of the dials.Far more accurate and it doesn't matter if you lash out the table or not before a move.

wmgeorge
11-27-2006, 09:41 PM
When I was making my own replacement dials for my old Atlas lathe, I used the face plate as a pretty accurate indexing plate. you can lay out 90 degree lines using the existing axes for bolt holes, or you can divide up the circumference using a trip of paper taped to the edge. Didn't believe I or the lathe could do it, and it was so easy and so (for me) accurate I was amazed. Many bull gears are also pre drilled for indexing, my one is already divided into 60, with a special detente pin to lock in the setting. The bigger the circle you are using for this, the more accurate your spacing is going to be. If you'll excuse the terrible pun, there's no real need to re-invent the wheel :-) Good luck!
Richard in Los Angeles

I did kind of the same thing. Used my CAD program to draw a circle the size of a CD-Rom, and then used the divide(?) function to automatically make degree lines/marks all the around the circle every 5 deg or so, added some text numbers, printed it out and glued to the CD, and my very own degree wheel.

Fasttrack
11-27-2006, 10:38 PM
Heh; uh what's a DTI? dial indicator...?

man i feel dumb lol

C. Tate
11-27-2006, 10:44 PM
Sounds like the part was not indicated properly before the pattern was drilled. Indicate each before drilling. If you were off center it can cause eccentric movement of the holes when you rotate.

wierdscience
11-27-2006, 10:57 PM
Heh; uh what's a DTI? dial indicator...?

man i feel dumb lol

Yup,infact I keep a couple 2" throw indicators around just for that purpose.

Jim Caudill
11-27-2006, 10:58 PM
Dial Test Indicator?

ulav8r
11-28-2006, 09:14 AM
A DTI is a dial test indicator, but is not the right tool for the job. I was going to say a dial travel indicator was the correct instrument but a quick look in the MSC catalog lists them as dial indicators, available with up to 5 inches of travel. A caliper could also be used if you can set up a fixed stop to measure from.

Fasttrack
11-28-2006, 11:10 AM
"Sounds like the part was not indicated properly before the pattern was drilled. Indicate each before drilling. If you were off center it can cause eccentric movement of the holes when you rotate."


Rotated? I didn't rotate anything - left it clamped in the vise the whole time...



My dial indicator only measures .5" of travel but has a .0001 resolution. I was wondering about using a magnetic base and some kind of clamp on the table to make two fixed points to measure with a caliper. Not sure how accurate this would be though. Any thoughts?

wierdscience
11-28-2006, 10:14 PM
"Sounds like the part was not indicated properly before the pattern was drilled. Indicate each before drilling. If you were off center it can cause eccentric movement of the holes when you rotate."


Rotated? I didn't rotate anything - left it clamped in the vise the whole time...



My dial indicator only measures .5" of travel but has a .0001 resolution. I was wondering about using a magnetic base and some kind of clamp on the table to make two fixed points to measure with a caliper. Not sure how accurate this would be though. Any thoughts?

ulav8r has it,make up a set of clamps for the mill table to hold your calipers.One clamp fixed to the table the other to the saddle.It will work like a primative DRO keeping track of table position by direct measurement.

Just make sure if the mill's travel is greater than the caliper's to remove the caliper before running the table down:D