Rookie question about accuracy, starting

[Bob T]

Hi All.

I have owned a 4x8 standard table with a pc router for several months and am teaching myself with vectric vcarve videos and trial and error with some success.

I have had trouble with fitting parts together. (mortise and tenon, rabbets, box lids) On the screen, these fit, but to find out the extent of my error, I made a dovetail to join 2 short lengths of oak. these were made from the same 2' board, screwed to the table with 4 screws, and not cut all the way through the stock so that I would separate the pieces cutting out the back on my planer (to minimize chance of movement).

After several trials, I found that the receiving piece needed to be offset outward .055" to be large enough for the joint to work. While I expect some slack needs to be built in, this seems extreme considering the accuracy the machine is capable of.

My OriginCal micrometer measures my bit's shaft at .2499 and the cutting edges small by about 5/1000s.

I appreciate your comments and help.

BT

[gc3]

ck machine moves true distance 48"...96" slop in z axis...bit deflection can play a part in this also

[Bob T]

Thanks Gene. Is slop in z axis loose bolts?

[Kyle Stapleton]

You should not measure the bit shaft, You should cut a slot in some kind of hard stock (maple, ect.) and measure the slot. Bits are made by starting with a rod and sharping the flutes into them, so the odds of them being the same size as the shaft are not good.
I have had good luck with telling the bot that my bits are .245"

[chiloquinruss]

I know it will sound silly but I have had a few senior moments soooo' make sure that you have set the right path, I.E. INSIDE, ON, OUTSIDE, cause it do' make a' difference! Russ

[Bob T]

tried to check for play, did not find anything loose except a long hex bolt that seems to steady the router in it's sleeve. Cut a test 4" square, measured 4.035x4.07. Cut a groove with a 1/4" bit, measured .2435. Looked in set up threads for tips in recalibrating, found none. any ideas?

thanks for help!

BT

[gc3]

attach tape measure to table use 60 deg vbit zero at 1" and move machine to 95" same along 48" side 1-47...there will be distance errors this sounds like one of the issues to me

[bleeth]

Well you learned all things marked .25" aren't! You didn't say what you were cutting or what your strategy was. A 4" square shouldn't be off by over a 1/16 or even 1/32. But, cutting strategy is important, particularly with small bits. (Gene mentioned bit deflection first!) To get fine accuracy when your material is thicker a good strategy is do do your first passes in the climb direction and then finish with conventional. Smaller parts can't be cut too fast either.
Highest accuracy will come from your machine only when the time it takes to fully adjust the mill is spent first. Squaring, leveling, ensuring z is plumb, tightening, etc. is critical.
You can add bits pretty easily in your software database. You can edit the existing .25 bit tool parameters and save it with a new name to reflect the actual bit. How to do that exactly depends on what software you are using.

[Bob T]

thanks for the suggestions.

I will try the tape measure. Then I will probably be asking you what I can do about it . . .

My tests have been run in scrap red oak, .8 thick. I see it on my toolpath menu, but have not played with conventional vs. climb. I don't know that I have a strategy.

When we moved it into my shop several months ago, we leveled it, but I have not checked it since.

How much influence can bit deflection have? I usually cut through 3/4" in 3 or 4 passes.

I certainly want to take the time to properly adjust my machine. But I don't know enough about it yet to know what is wrong . . . .

Thanks, will work with your info. BT

[blackhawk]

Are you sure that your unit values are correct? A 4" square should not be .07" oversize. Bit deflection would not account for that much error.

Are you sure that your material is not moving on you while cutting?

Is this an older machine? Your racks and pinions may need to be replaced.