It's PERFECT! - Calibration is Important
HTML-код
- Опубликовано: 9 окт 2024
- The custom fixtures are completely set up, and the code works great! But now the tools in the mill are having trouble producing the perfect pivot hole. Luckily, calibration will save the day (and the knives). In this episode, John and Angelo dive deep into tolerances, measuring, and the lengths they go to to make sure everything fits together perfectly.
This is the final part to our series on setting up our new custom fixtures, but we will continue to do "Fixture Friday" videos whenever we get a chance. Stay tuned!
------------------------
Our Products: www.grimsmokniv...
John Grimsmo Knives Instagram: / johngrimsmoknives
Erik Grimsmo Instagram: / erik_grimsmoknives
------------------------
Gear we use:
Canon M50
Zoom F1-LP Field Recorder
Rode VideoMicro Directional Recorder with Dead Cat
JOBY GorillaPod Compact Tripod with Ballhead
------------------------
Angelo Fisher Instagram: / afisherknives
Erin Kelly Instagram: / '
Skye: / skyemfg
Jo (pronounced Yo): / jovonrosen
Music from: www.epidemicso...
I went to Legoland and saw their multi piece molds. Those have got to be less than .0001 tolerance to eliminate the parting seams. And the tolerance stack between different size pieces is about .004 when I play with them. You're struggling in 2019 with metals. But how did they do it and address the ABS shrinkage that other companies still can't match, in 1977? With metal molds translating to plastic parts! Makes me appreciate how incredible Legos really are.
Thank you Grimsmo team for sharing your process. Your dedication to perfection is both a benchmark and an inspiration!
At 4:01 just to be pedantic, the dimension in that shot isn't basic, it's a normal variable dimension. Basic dimensions are shown by putting a box around the dimension and giving it no tolerance, since it's supposed to represent a theoretical perfect dimension. GD&T is a really powerful tool that's difficult for beginners to get up to speed with and I know that this video will probably be the first a lot of your viewers are hearing about GD&T, so I just want to make sure no one's being misinformed or confused. Keep putting out all the good content guys!
During a factory tour at Triumph, I found it interesting how they don't bother going for super tight tolerances on every part so everything fits. Instead, all parts have a relatively loose tolerance, and then the different parts (e.g. piston, engine block etc) are measured on the CMM, and a computer system matches which part should go with another based on which dimensions of each part best fits another.
I think Grimsmo does that with some pins he uses in his pivots? But it's for parts that are cheap to make and small since it is essentially a batch work technique which he wants to eliminate if it is substantially wasteful.
Love seeing Eric whacking away on the fitment and lockup working upstairs in the background :D
Use the probe for thermal comp. Write a macro that will apply external offsets in a “strategy” program that will account for machine and fixture growth. A lot of people overlook this, assuming that scales will take care of it. As your machine grows, you’ll see measurable deformation in places that linear scales can’t account for (spindle perpendicularity, for example)
John, you should mill the hole and the reacesse both at same op AKA 2nd op so you dont have the issue with concentricity!! also use Pin gauges!!
Hey John, here's a somewhat related probing suggestion, considering it is so intrinsic to your workflow. Have you ever messed with the speed and back off distance on your probe? put simply, you can change two parameters on how the machine probes, and how those balance can change the speed of your probing routine dramatically. The speed is the initial feedrate of the coarse probe hit, and the back off distance is how much the probe will retract for the second hit. So if you go faster feed, you'll need a larger back off distance, to account for the distance the machine over travelled before the probe triggered. At the shop where I work we just started playing with this, and took a 16 min probing routine (large castings, lots of centerline and A plane measurement) down to a 4 minute cycle. We tripled our probe feedrate, and doubled the back off distance, so the first to second hit time is down to less than half a second now. Your results may vary, but fun to play with.
that was awesome guys great job! good team effort as well!!
If you are trying to hold a positional accuacy of .0002" I suspect you will encounter problems just from what temp it is and that there will be some drift from the heat generated during cutting from thermal expansion. This could go one way or the other especially if your coolant temp is not maintained at an exact temp. 10deg change will move a 1" spacing .0001". Do you have a heater/cooler on your coolant to maintain it's temp?
Also keep in mind this can be very problematic if you are running ops on both pallets at the same time. Assuming the pallets are a foot apart and your temps rise 10deg during the operations of cutting one pallet will be almost .001" out of position from when you started from the iron table change. There would be even more error from the aluminum pallet but I didn't bother to add that in. Regardless it would cause you major problems. in trying to hold anything to below. +-.0005"
You win this years Ultimate Workshop Microphone award (-:
How about milling the smaller hole ubdersize first. And when you flip the handles. Then you would make it to size and ream or whatever that fancy super great qulity hole was named. And then do the outer edge?
I think they have 3 edges in one hole. The big one, the smallest one in the middle and the third one what have made problems and is bigger then the hole in the middle. So it is irrelevant from which side you start to mill.
Hi John. Very interesting but. I guess I am not getting it. If you have access to the pivot hole with the probe. Maybe You should consider leaving that hole a few thousands undersized on the 1st opp and then after You flip it over. Finish the pivot hole and the counter bore both on the 2nd opp. I have always had good success finishing the bore / counterbore on the same opp
Awesome video! I'm definitely learning a lot about fixtures and the value of probing parts
Documenting the problem is best so that future people know to check for it quickly.
“The more difficult it becomes because it just gets harder and harder and harder.” - Good for a t-shirt perhaps? ;)
It's so clean! Does the CNC clean itself somehow or do you all have to clean it?
With your current programming, how difficult would it be to add a 3rd or 4th pallet? If the program uses the probe to check if pallet 1 and 2 are there, it should be fairly simple to add a 3rd pallet and more or less copy paste code if you wanted to increase production right?
Amazing video John..... man those are tight tolerances...
John, how many knifes you have to sell in a month to stay on a profit side after all expenses?
Is using aluminum for your fixture plates causing any thermal problems?
Hi John,
Another great Probing video.
I am a ROOKIE when it comes to probing so please don't be insulted when I ask you this calibration question as I have never heard you mention it. Does you probe always stay in the spindle in the same orientation? If you remove it and don't reinstall it in the same orientation (i.e) it can be reinstalled 180 degrees out which I was told could result in probing accuracy issues once the probe has been calculated. I am sure you guys can verify if this correct or not. I placed a mark on the front of my probe to make sure it always goes back in the same orientation. I am confident you would not have done this. You guys are the Renishaw Probing Team. Many thanks for sharing all your knowledge.
John
Yes a probe is placed in the spindle in the same orietation every time
The vast majority of industrial CNC machines have "spindle orientation" capability, which means they can rotate the spindle to a specific _orientation_ on demand. This is done because the tool holders most of these machines use have cutouts on them that absolutely *must* line up with protrusions on the spindle (called "dogs", I believe) in order for the tool holder to load properly. Some spindles and tool holders have a special shape instead of cutouts and dogs.
Since the tools are held in the tool changer in a specific orientation, and the spindle always goes to the same orientation when performing a tool change, the probe is, indeed, always loaded at the same _angle_ or _orientation_ .
When putting in a tool manually, the operator must align the tool holder cutouts with the dogs on the spindle, so tools are always loaded in the same orientation when loaded manually as well.
Oh, and regarding the possibility of being 180° off, the dogs on some spindles are sized differently, making it impossible to load them _backward_ . That's not always the case though. For instance, the dogs on the BT30 spindle on my machine are the same size, making 180° rotation possible. If I remove a tool, I have to remember how it was oriented before I put it back into my tool changer if orientation is important. Personally, I load my probe with the battery compartment facing the front of the machine.
If the heads tolerance is that tight, couldn't the hole through the handle be allowed to have some room to move? Meaning the head is keeping the pivot in place, the hole would be redundant. Not sure if you know what I mean, or would want to do it that way, but it's certainly easier. Just a newbie trying to learn.
John, that's an awesome inside sharing! Two palets == 2x problems. Maybe the solution could be one bigger pallet? Please entertain this thought :)
Oh, forgot to mention. Your vids are superseded only by SpaceX... that's a compliment :)
Oh, another idea... mill pivot holes on the inside of the handle a little undersized, and the bearing cavity really does not care about precision. So if you finish the hole perfectly from the top after you flip the handle, together with the pivot head cavity - that probably should improve the accuracy as well.
Why don't you bore the pivot hole and the counter-bore in the same operation?
try milling from first op with a saw end mill enter on center contour mill then exit on center
These videos are epic
If you really need a close tolerance on a bore. I'd use a carbide reammer.
I wound't trust the interpolation, unless you have a closed-loop machine of course.
So...... like where can i order a pen???
22:52 Ha! Balls are touching...
Why not machine the outside first since the bearings can have some gap?
Keep at it 👍
23:55
Why the probe does not rotate?
Rotation will eliminate runout
That's something I'd like to ask Renishaw as well. It seems like an "obvious" solution to the runout issue. It's such a simple solution I can't imagine they didn't consider it, which makes me wonder if they studied it and realized that it causes more problems than it solves.
I have two guesses why they don't orient the spindle...
1. Dialing out the runout isn't really all that hard so it wouldn't help all that much anyway. I wrote my own probing macros, and they do orient the spindle, but I also dialed-out the runout anyway because I couldn't live with myself if I didn't.
2. Not all machines will let you orient the spindle to any angle (for instance, I believe that capability is an option on Haas machines). Maybe leaving spindle orientation out of the equation makes their system uniform across all machines, which would be beneficial in some ways.
Oh, wait, did you mean rotate continously, or rotate to the appropriate orientation and then stop?
@@Hirudin continuously until probe touches item to get value
@@Hirudin apropriate orientation - I don't know. Could work, but...
@@Hirudin angle issue for spindles - agree. SO continuous rotation will do the job
A dimension in mm .. SHOCK HORROR
10:15 giggity!!!
How do you Now if checking is correct when there is ASME you dont hace the Tools Check everything the appropriate way
I dont Wang to be mad just my saying
Yeah metrology.
Wow you have no idea how to calibrate a probe. Deflection offset values? Ring gauge, tool length gauge
Why did you switch from solidworks to fusion 360? I tried 360 and I just feel it gimped compared to solidworks and solidcam combo. Is this a cost thing?
And why tf are you measuring pocket knifes to the tenth! Put five thou tolerance on everything and let er rip
Time for a cmm.
The probe ,if he wanted , can probe finished dims and print them out just the same as a cmm
@@gusmcgussy3299 But that doesn't tell you anything besides that the machine repeats, you can't measure with the tool you make with. You could have a ball screw with .003" slack in it and still get "good" measurements.
@@bcbloc02 than why does boeing accept our printouts sent wirelessly from post routine probing.... most is true position within .001 ans hole sizes .0002+-
@@bcbloc02 i dont disagree with you... just confused at this point
@@bcbloc02 glass scales make it so there could be 3 inches backlash.... linear movement is measured
Pls stop saying slop. The correct term is clearance
Bad fixture design.
Dont you realize those machines only repeat 2 tenths... lmao... your chasing ghosts lmao.... besides... 1 thou or a cpl tenths isnt going to make that pin sit like that...sorry
12 planes of movement? Your shop is bigger than it looks.