You're welcome. HowEVER....just so you know...the ONLY reason I write these posts...is so you and others benefit. Knowledge that is not shared has no value.
TL/DR: Machines have operating envelopes and sometimes we don't get out of them all they can do because we don't understand how to move within that operating "space".
====================================================
Giving just attribution, the guy who runs cnccookbook.com, in my working with him to get his excellent F&S calculator to include the features necessary to characterize a machine like the cartilage-based sharks, once told me "I don't think you're pushing your machine anywhere near hard enough."
Well, he was partially right. That is, I was babying the machine....somewhat...maybe by 50%. But, the problem wasn't weak drive motors, or worn leadscrew backlash, or lousy controller capabilities. It was HDPE. But not "just" the HDPE.
In any system like a 3-axis CNC, all the three axes have to act very similarly in "compliance", that is deflection under load. You don't want X to be really soft while Y is rock solid and Z is some other value. They work best when they are close in compliance, rigidity, even primary resonant frequency. And the HDPE in the sharks are actually somewhat matched to the rest of it.
So, that becomes a problem when you say "to heck with this weak router, I want a SPINNNNNNDLE!!!!!" Well, a spindle weighs 12# and a router weighs some other amount, far less in my DeWalt D618 long ago. Young's Modulus says those different weights make the system move differently. C'mon...we've all seen this. Tramming these beasts is a nightmare because there's compliance at each joint in the system that has to be individually compensated.
Yikes!
People get dixie-billet clamps, and that solves the holding-tight problem, but adds to the weight sag.
You can shim up the spindle mount, but that just doesn't fix the z-axis slide rails also not being perfectly vertical.
You can then fix what holds those slide rails, but that is only the same in the middle of the gantry.
Yup, as you move from end to end, the sag in the gantry changes.
So sure, you can stiffen the gantry (yes, I have a kit for this
),
But that then transfers that torque down the gantry arms.
The first thing you notice (no, really...do this...you'll freak) is that if you watch the cross-wise drive plate underneath the bed while you tug at the top of the gantry ALONG the length of that gantry...you see an S shape.
Okay, fine...you can use the crane hoist and temporarily move the bed off, get underneath, and totally beef up that with very little weight...I'm on it.
Oh...but then, you're left with the plate that joins the top of those long linear bearings to that under-bed drive plate.
And you realize that ALL of the former deflection is STILL in that member.
Sure, the underside doesn't move
And sure, the gantry is stiff as heck
and for me, it's now pure metal all the way forward from those bearings...well, alloyed metal anyway
.
And yet, still, I'm f*cked by a 3/8" plate that really should have been 3/4" thick, all the way across, which is now the ONLY flexing element, and boy does it.
I've done every one of those things except that final fix.
The machine is now a BEAST, and moves with the same movement dynamics, because I've added less than a pound.
Material removal rates?
Okay, remember how I said I was babying the machine?
I decided fine...I'll do better! And I did...scary, sometimes failing, but stronger and harder and better, but well over double what I was getting before. Stupid deep cuts for this machine. Toolpaths that took an hour then took 20 minutes, and three depth-passes instead of five.
Which then allowed me to feasibly make those upgrades to the gantry system. 10 different precision components that ('cause I's poh!) reused the precision components already in use on the Shark. That's right....can't assemble the final unit until it's not only all cut, but the machine is so far disassembled there is NO going back. Yeah, well I'm an in-gin-err, and my sh!t works...which it did
.
And now? OMG....cuts an ORDER OF MAGNITUDE harder faster, and more accurate at those. That's right, I can now PUSH the bit....Yeah, I DARE you to push a full slot on a shark. I can now, with no bit diving, and no deflection....except in that one damned joint
.
Seriously...those cut parameters I recommended earlier yield 0.0275 cubic inches per minute. I've gotten (in a torture test) 0.866 cubic inches. Yes, nearly 32X
On a shark that electronic configuration-wise has ZERO changes. That's right, if you had the systems I have, or rough equivalents (chip BLOWER as in with a nozzle, not a vacuum, and a pulsed mister...maybe $30, two components and a power supply on eBay) you'd be able to do WAY more than anyone imagined using these machines.
Yes, other machines are way better, and I'd love to have a $150K machine, but for nearly all projects, ±0.005" is well within customer and product needs. And asking for more than that should be questioned, anyway.
So now, those toolpaths that took an hour that then took 20 minutes? Yeah, one maybe two passes, 8 minutes, and don't stand near the machine because the chips are HOT, and go flying down my shirt and burn me
....again
. If I want to watch, I use the web-cam
Sharks are cool. They're all capable of SOMETHING. Getting what you want out of any robot is all about convincing it to make it for you. Success requires this, and cannot be achieved without it.
I'm always envious of the 12,000# "Big Blue Beast" here in the 8,000 square foot metals-fabrication shop...that hasn't moved in three years. Why? Because a while back, it sh!t its brains, and manually keying in 2K lines of ladder code (ew!) is not really something I'm willing to do. After all, every single CNC-requiring job that's come in has been capably handled by the Shark. So, Pbbbbbbbtttt on you big boys.
As I was once told by another engineering mentor, "good tools can make a good engineer better. But even great tools can't make a bad engineer good."
And as I tell my clients, "my sh!t works."
Cheers, y'all. It's been bizarre listening to the #1 presidential debate during the typing. What a 4 year-old.