My indicator says that the part is centered but when I start to drill the hole I can visually see that it's not centered.
Does small angle approximation really matter that much? The indicator is slightly off to the side thanks to shitty mounting system, it should still center right?
>>961539
>my mounts are not centred
>it doesnt drill straight
>>961560
Do you actually know anything about machining and geometry or are you just fucking around?
>>961567
nothing at all.
but it's a basic checklist that someone with absolutely no experience with machining can follow and understand.
if it's drilling off centre and the tool you are using to measure allignment says it is then there is a problem with either the instrument or your ability to use said instrument.
it could also be a problem like how the tooling and object is held. a worn bearing for example will cause either to slap around. check for play.
>>961539
An indicator is not a magical device that makes things centered.
It's a measuring device that tells you the distance between two points.
You've got it between point X and the outside of the workpiece. Assuming the workpiece is perfectly circular, then if it rotates in the chuck and the indicator doesn't move, that means that at every angle the distance between point X and the outside of the piece is the same, and that means it's perfectly centered _in_the_chuck_.
You could mount arse of the indicator anywhere in the entire universe that's not the chuck, and it would measure the same thing and give you the same end-result.
I'm not a machinist, but AIUI, your next job is to mount the indicator to the chuck, and align the mount the same way you aligned the workpiece.
drill could be wandering?
if you have the room mount the indicator in a drill chuck and check again
center drill?
how are you checking runout?
rotating the the part? the dial indicator?
did you check straightness up and down?
>>961539
you are spinning the indicator right, only ask as i saw someone mount piece, zero indicator then spin chuck, then proceeded to drill hole. (at first i thought he wants to rotate the work piece at some point, but no just drill, dude must have been stuck in "lathe mode" or something)
I'm a machinist (pic related)
If your dial is not on a flat surface, or completely perpendicular to the piece then there is no chance of you getting a clear reading. Why do you think machinists have angle plates, granite blocks, and machinist's squares? Because precision is key, and the surface is a part of that.
>>961613
Edit: if you don't wanna deal with your indicator and have basic machinists skills, do some simple lay out on the end of your piece.
Scratch half way up the piece and half way left/right and find your center.
Once you have your center, use a prick punch, then a center drill, then work your way up to your finish diameter.
>>961613
OOI, why does this matter?
If you're only interested in how the distance changes, why does the dial have to be perpendicular?
If the dial is not perpendicular, the variation in its reading will be the same, just scaled by the sin of the error from perpendicular, no?
If you are swinging the indicator then you are centered, at that distance from the spindle.
Looks like you have a Sherline there, so no quill. The vertical column must be truly perpendicular to the x/y ways or when you change height (or distance from tool tip to spindle) you are changing your position in relation to x/y.
In the pic the box is the head, the distance to table 65mm. With 1 degree of misalignment you get over 1mm error in position.
Put a precision square on the machine table and run your Z up and down it with an indicator to align. If the head is adjustable on its mount, then sweep (tram) the table to align the head once the z axis is aligned.
>>961539
>My indicator says that the part is centered but when I start to drill the hole I can visually see that it's not centered.
OP: do yourself a favor and get a coax indicator (pic related)
Yes you can get by without one--but it makes centering things under a spindle literally a 30-second task. It takes you longer to take the thing out of its case and put it back in, than it does to get your part centered *perfectly*. I just used regular idicators for a long time, feeling that spending THIS much for a gauge to do ONE task was silly,,,, but holeee fucking shit, there was times I spent ten minutes trying to center a part using regular gauges, and never got it perfect. With this thing it takes literally under 30 seconds to get the part centered perfectly.
Tip: oil the sliding parts of the body, and also put a bit of oil on the workpiece under the feeler end, where it slides around the part. This will keep the feeler end from chattering.
Also we note: the dial on these things is graduated in [somethings] but it doesn't give an actual linear measurement--it just gives an indication of relative movement.
>>961603
I'm rotating the spindle.
I trammed the mill previously but I can check again, this machine is adjustable in every direction.
>>961606
Yes I rotate the spindle, the indicator is clamped to a center drill in the spindle.
>>961659
Thanks I'll try this. I trammed to the table but didn't think of using a square.
>>961701
Machine is too small to use these. I need like minature versions of tools.
Ok so I did this and got a difference of 5 thou over 3 inches of z travel which is like 0.01 degree off from square. I know it's not really 5 thou because of the angle of the tip, but the hole i am drilling is only 4 mm deep so i don't think it really matters.
And fuck these chinese or russian made holders, they cock the indicator at a slant when you tighten them. Even the dovetail screw that came with the indicator is oversized and crooked.
>>961770
How far off center were you when you went to drill?
If you want to get creative, and you need the hole in the center of your part, hold the part in the head of the mill and your drill in the table. You can eyeball the centering quite well and once it get it spot drilled the drill will tend to stay centered better on a rotating part.
I guess you don't have a lathe?
That's a 7mm dia part and my indicator says that shit was centered.
I've drilled better holes by eye.
I have a lathe but my non adjustable tailstock is out of alignment so it's useless for drilling.
k stop crying cosine error, its not going to have anything to do with this
it has more to do with transfer/comparative measurements.
for our sake, we will be using it to see run out so the actual measurement is arbitrary
you can also adjust the angle of the needle
and your holder is fine, you could stand to use it better
>>961738
>I'm rotating the spindle.
k, that rotary table rotates right?
rotate the part keeping the dial still
also,
>>961770
do this, but do this to your part in the mill
do it on 4 sides, up and down
rotate the dial in the holder around the part at different heights
do a sweep of the table
looking at the sherline mill, its no doubt you are not getting perpendicular holes the way that thing is laid out.
make sure all that shit is tight
also, make sure you are taking small pecks and use coolant, any excess pressure and it looks like it would load the whole (ghetto ram and turret)
>>961849
>I have a lathe but my non adjustable tailstock is out of alignment so
fix that shit
you have a mill and a dial
>>962037
What will rotating the part do? I know the runout on the part is one or two thou, this is within the lathe specs.
>>962078
it will check for other runout
how far away from in the z were you when you indicated from the spindle? like in the first pic?
and then all the way down with the center drill?
>>962115
Like 2.5 inches.
I'm going to re square everything tomorrow.
I don't know what to do. Because the axis for squaring travel in the x is so far back, any reading I take from there shows almost no change while the reading at the front shows a huge change. Now this wouldn't be a problem if the machine wasn't floppy as shit, any kind of force applied to the head causes the reading at the front to swing by like 10 thou and repeatability goes to shit.
>>962037
>Shit talks the accuracy of his mill
>suggests fixing tailstock with it
Well I squared everything up as best as I could and this is how the test cut came out. It isn't perfect, the ridge is just barely enough to catch a finger nail on. Maybe one thou deviation over 2 inches, but I think that is within the specification of this machine and good enough for what I am doing. Tomorrow I'm going to do the y axis test cut.
I have a sherline CNC. You have to square everything that's the key.
After its square she will run like a dream. Pic related
spend the cash and get proper tooling
>>962531
What kind of anti backlash are you using? Because the standard nut system has massive play.
Ok make sense of this for me. The circle was scribed by the head of the indicator, the dot is the start of center drilling. The circle is concentric, I checked by spinning the chuck on the lathe. The indicator was clamped to the center drill.........
>>963096
I trammed the fucking thing and the x y deviation is 1 thou over 2-3 inches.
Even if the z was tilted, the indicator is simply an extension along the tilted axis, removing the indicator and going down the axis doesn't change x y placement.
>>963053
You do know that 3 jaw chucks are inherently inaccurate? The jaws are on a spiral so each jaw will tighten at a slightly different position. No issue as long as you don't take the workpiece out and put it back in again.
>>963262
Yeah yeah, that has nothing to do with the magic nonsense going on here. I made a new pin that fits in a collet for my good indicator holder. Now I got the indicator measuring off center pretty much.
I dialed the part in so that the indicator swing was under half a thou, then I accidentally pushed the indicator instead of rotating the spindle. Now this is where it gets weird; the indicator was still in contact with the part, but it was swinging all over the place now instead of being 'centered'. I gave the indicator another push in the opposite direction and it began to show a centered reading again when I rotated the spindle.
oh my god what the fuck do you not understand about triangles
what the fuck else do you even think is happening then? there is like 2 things it could be your indicator is fucked or your z is tilted
you indicate all your fucking shit at one height
when you move down on your tilted axis
because its not straight line 90*, it will not stay same center point
its the only fucking axis moving
>>963408
ok tell us how you use your indicator step by step
>>963435
You're missing the part about moving down the tilted axis. What you are describing is not possible unless the indicator somehow swung around vertically while clamped to the tilted z axis.
Most of your problem is the machine is not rigid. It does not look like it will stay square if you do happen to square it up. Tool pressure alone will move that machine out of square.Some lathe chucks are adjustable..that one is not .Boring soft jaws will bring the chuck into center. Picture is Devlieg that I run at work..it was built in the 70's .
Fuuuuuuuuuuck, this stupid chinese shit is telling me this is centered. I checked at three different heights on the part.
>>963517
Yeah maybe I should buy the fixed column for it....or just chuck all this shit in the trash and get a job at mcdonalds.
>>963548
Forgot pic.
>>963548
Anon, it's taken four days of 16 people telling you what was wrong to come to this conclusion. McDonald's is probably a little out of your league.
>>963589
It's not a conclusion, it's only removing a number of unknowns to maybe get to the bottom of why this is happening because as I showed here >>963489 a tilted axis would not do this.
>>963550
This might also be my fault, I did not check if the dead center was inserted properly.
Doing a bunch of tests now; I'm going to center the part with the indicator, raise the head to the highest point and with a drill chuck and bit check the centering and also bring the head all the way down and with the dead center check the centering.
I'm afraid that the actual head or ways might be warped.
>bahh my indicator my mill, my holder, my ways, my head
9/10 times its the operator not the equipment
I've found a possible solution, an interesting way to DIY a centering scope. With something like this or a camera based one I can see how the center moves as I move the head up and down.
https://www.youtube.com/watch?v=F5txhXUeVkM
You guys were right, I checked the tram again and it's off. I swear I had this thing dialed in.
The way of removing tools on this mill is to hit the drawbar with a mallet after you back it out a bit and that must be fucking everything up.
>>961770
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
I re did this and adjusted the axis until I had side travel to one thou over the whole up and down z axis. Then I tested the table to make sure, I brought the indicator into contact with the right side of the table and turned the spindle to go to the left side and the indicator was floating off the left side of the bed by like 10 thou!
FML none of this makes any sense.
I know you guys don't care anymore but this is true diy and I'm going to be posting until I figure this out.
After the last series of tests showed that the table was not square with the z when side to side deviation is minimized, I started thinking and realized that there could only be two explanations. The table itself had to be warped or that the spindle was out of alignment from the z axis, something which I thought was Impossible on this machine.
Testing the table showed that it is indeed warped, by about 5 thou over the whole 12 inch length, but that isn't enough to create the 25 thou gap I was seeing when the indicator was spun from right to left.
>>965109
That only left the alignment of the spindle to the z. But how could there be misalignment? Sherline themselves state that the head on their mill and lathe are aligned by milled key ways. That would have to mean that the actual base that travels on the ways is misaligned. I then remembered from some research that sherline uses tapered plastic gibs, could the gib have worn down? I loosened the gib and moved it up and down, there was no change in the head. That's when I noticed the pictured, this was described in documentation as a backlash adjustment method. I undid the two little hex screws and the entire head just flopped to the side by a good 20 thou.
Wow just fucking wow, the only thing that puts pressure on the gib, the only thing that keeps the spindle in alignment under cutting load is this thing that pulls the lead screw nut to the side and binds the lead screw. Did I also mention that the end of the leadscrew is free floating!
>>965117
Lesson learned, never take anything at face value. I wasted a week because I thought the head was keyed in alignment with the z axis.
Use your shitty mill to make a less shitty mill. Repeat.