[the_canuck]

Today I tried cutting Aluminum for the first time. Things went as expected and the machine cuts AL very well. I cut some passes with 1/4" 2 flute HSS end mill at 20 ipm turning the spindle 24k. Not ideal feeds and speeds but it worked just fine. I can cut passes up to 0.1" deep 100% cut without any chatter. Pushing harder I get minimal chatter. No coolant dry cut.

I also tried to tram in the gantry as I hadn't before other than rough measurements.

After about 4 hrs of loosen, shim, tighten, and measure on both x and y I am now within a max of +/-0.0008" of perfect tram measured 1 inch from center of rotation. It initially measured 0.004" before I started to shim. Not bad for a router with a gantry.

My edge technology dual indicator tram tool came in very handy.

Andrew

[the_canuck]

Waiting for my XHC wireless pendant to arrive. This should allow me easier part zero setting as my keyboard is a bit out of reach.

I've decided to give the tram process another try. I'm just not satisfied with the results from the first attempt.

Andrew

[the_canuck]

So rechecking tram last night I got a slightly better result. I'm now at +/- 0.0005" max. My theory is that it took a bit of time for the metal 'relax' in response to the shims.

Andrew

[the_canuck]

Here's a facebook link to an album of my first wing panel test cut in a piece of scrap.

http://www.facebook.com/media/set/?s...3321575&type=1

I fixed the y-axis set screw loose issue by replacing the original set screw with a longer one and drilling out the shaft to stop the backlash caused by a small setscrew setting in the larger key-way. The x-axis pulleys have 2 set screws and so far are staying tight.

x-axis servo gain still set too high causing some hunting when moving at slow speeds.

Andrew

[mactec54]

the_canuck

Try the Integration Gain setting, Set it to ( 1 ) or ( 2 ) you may have it to high

The Main Gain 60 to 80 should be good on your machine

[the_canuck]

After tweaking the Servo parameters again last night I'm getting closer but still getting some stationary buzzing (can feel it and only hear it sometimes) when they are just holding position.

X-axis and X-axis' (Running 2 servos for x-axis)
Main Gain: 70
Speed Gain: 33
Integration Gain: 12
Torque Filter: 127
Max Acceleration: 8
Max Speed: 20
Command Input Mode: Pulse/Dir
On Pos. Range: 127
Gear Num: 500
Servo Mode: Position Servo

Y-axis
Main Gain: 50
Speed Gain: 27
Integration Gain: 6
rest of the parameters are the same as x-axis servos.

Z-axis
Main Gain: 67
Speed Gain: 38
Integration Gain: 24
rest of the parameters the same as x-axis servos.

The X-axis and the Y-axis are using a 2:1 belt reduction drive with a 1" pinion and rack setup.
The Z-axis is a C1 Ball Screw with 12.5 turn's per inch.

The mass of the total gantry is between 200 - 250 lbs.
Y-axis mass is approx. 70 - 80 lbs.
The Z-axis mass is approx. 30 - 40 lbs

I have linear bearings on all axis that move freely when rack and pinion is released.

I'm having the most trouble with the X and Y axis servo tuning. Z-axis is behaving itself but still has a bit of stationary buzzing (can feel it but not hear it)

All axis work well at slower speeds. Higher speeds I was getting jerky movements until I halved the integration gain and backed down the main gain about 10%.

Andrew

[mactec54]

the_canuck

Your Integration Gains are to high, none should be above 6, start at 1 on all of the drives

It's quite normal to be able to feel & hear some kind of servo noise when stationary

Z-axis
Main Gain: 67
Speed Gain: 38
Integration Gain: 24 ( Set to 1 )

X-axis and X-axis' (Running 2 servos for x-axis)
Main Gain: 70
Speed Gain: 33
Integration Gain: 12 ( Set to 1 )

Y-axis
Main Gain: 50
Speed Gain: 27
Integration Gain: 6 ( Set to 1 )

Set these to 1 will make it more stable , you can play with this number but you should not be above 6


What settings are you using in Mach3

[the_canuck]

I'm running LinuxCNC not Mach 3 so the numbers will be different. LinuxCNC setup using StepConf. My latency Base period Max. Jitter is 16988ns. Min base period: 26488ns. Max step rate: 37752 Hz.

X-axis:
Motor Steps per Rev.: 2000
Driver Microstepping: 1.0
Pulley Teeth (Motor: LeadScrew) 1.0:1.0
Leadscrew Pitch: 0.6366197 rev/in
Max Velocity: 25.0 in/s
Max Acceleration: 50.0 in/s^2

Y-axis:
Everything the same except for
Max Acceleration: 60.0 in/s^2

Z-axis:
Everything the same except for
Leadscrew Pitch: 12.5 rev/in
Max Velocity: 1.2 in/s
Max Acceleration: 5.0 in/s^2

[mactec54]

the_canuck

I have not run LinuxCNC so can't help any with the control settings, Max step rate:37752Hz, Your Drives are independent of the control so those setting I have given you will be correct, I run a U100 or a Smooth stepper, The U100 I find very good, for a simple set up you get 100Khz step rate with this, & a much better smoother running machine, the Smooth Stepper has a much higher step rate

[the_canuck]

I took a break from servo tuning tonight and spent some time on Linux updates and getting the wifi card working with 2 antennas.

I then fought through getting my new pendant working: LinuxCNC Documentation Wiki: Using A XHC-HB04 Wireless MPG Pendant

Just before 10pm I finally got it working, well at least the major buttons. I still have a few others to get working.

Andrew

[nlancaster]

the canuck that is a good looking machine. May I ask where you go that wireless pendant. looks great and would be awesome to have one.

[devmonkey]

the_canuck, you have given me the inspiration to finish my machine. This is a steel/aluminium/hwin/dmm build as well, 1500mmx1500mm, although with ballscrews rather than racks. I'm about to wire up all the dmm kit (exactly the same package you bought in 2010) and have a question, is it really required to use the resistors and diodes to parallel the supplies?

[mactec54]

devmonkey

a question, is it really required to use the resistors and diodes to parallel the supplies?

No it is not needed, I have installed a lot of these systems, & had Zero problems ,with the power supplies

One thing you must do though is to set each power supply to the same DC output number 48V, They can be set as high as 52V but they all must be the same number

You have to set them before you parallel/join them together, Here is one I have just done

[the_canuck]

Ya, I agree they are not needed if Vout is closely matched but if not matched switching power supplies will fight with each other and can cause weird problems on the output or during start up. I used to design and build switching power supplies and we often ran them in parallel but with a basic combiner circuit (resistors and diodes) or more elaborate combiners if needed after test. These supplies seem pretty tolerant probably because they have some built in components on the output.

My solution is overkill which goes with my approach to this whole project. :idea:

Made some progress on the table. I found a local source for cast ground AL plate. I am using some AL scrap as machined spacers on the cross beams. I plan on machining them flat and then attaching the ground plate on top and tapping it on a grid pattern. The pretty ground plate cost is an incentive to make sure I use a spoil board when I think I'm going to machine into the table. :nono:

Andrew

[the_canuck]

Progress on my table. Lots of cutting AL the last 2 days. The spacers are 7075 scrap. I machined them flat and then cut the 3/4 cast plate to fit around the table legs. There's a small gap around the edges on purpose. It will allow flood coolant somewhere to go. After the table is done, I'll add a catch basin below the main table so I can recycle coolant.

Andrew

[ger21]

The pretty ground plate cost is an incentive to make sure I use a spoil board when I think I'm going to machine into the table.

Machining into the table will happen when you least expect it.

[the_canuck]

Vacation time. Off to China. I'll be back at the end of the month.

你在10月見！

Andrew

[Davi37ggs]

You have to set them before you parallel/join them together, Here is one I have just done

[mactec54]

Vacation time. Off to China. I'll be back at the end of the month.

你在10月見！

Andrew

Have a great Vacation

[the_canuck]

Wanted to give you a heads up on a cool event happening this Saturday in Loveland. See website for location info.



Website: NoCo Maker Faire | October 5, Rocky Mountain Center for Innovation & Technology

It’s a gathering of scientists, engineers, hobbyists, artists, small businesses etc. that are coming together to show off the stuff they make. Please check out the website for a list of companies, government agencies, etc. that will be there. Along with the booths and demo’s there will also be workshops, speakers, hands on learning opportunities etc. If you have kids interested in science or are a science geek yourself it’s a can’t miss event.

I was asked by Karen C. one of the coordinators to come and speak on the big stage about my CNC router I completed this summer. http://www.cnczone.com/forums/cnc_wo...l_monster.html With my trip to China I didn’t have enough time to prepare any demo stuff or speech and had to turn down the opportunity. Besides my CNC machine is too big and much too heavy (est. over 2000lbs) to drag up to Loveland. I did promise to promote the event so that’s what I’m trying to do. Next year I promised to have a booth and demo of some sort.

Andrew