Re: GME's New 80/20 CNC Build - My Design
Quote:
Originally Posted by
wmgeorge
Thats worse case, the calibration Certificate that came with it, was more detailed and concise. Makes you wonder what the cheaper ones are at?
Agreed, it's the worst case. I suspect that the error increases the wider you open the jaw, but I have no empirical evidence to support my suspicion. Interestingly, +/- .001 seems to be common, even with the cheaper ones. But then, I don't believe everything I read in advertising, and the cheap ones don't come with calibration certs - at least not the ones I've seen.
I don't think I mentioned it, but I bought a Vista CNC pendant. I got the P2-S. Pretty nice, and way better than the Shuttle Pro I had been using. It was a snap to set up.
Gary
Re: GME's New 80/20 CNC Build - My Design
Yes I have one of the Vista CNC pendants also and I love it. Mine is USB, I wish now it was WiFi but I use it all the time. Looking forward to seeing your machine making something Gary. I am stalling off a Honey Doo job
and hope to get back to running the CNC.
Re: GME's New 80/20 CNC Build - My Design
Quote:
Originally Posted by
wmgeorge
Yes I have one of the Vista CNC pendants also and I love it. Mine is USB, I wish now it was WiFi but I use it all the time. Looking forward to seeing your machine making something Gary. I am stalling off a Honey Doo job
and hope to get back to running the CNC.
I wish I had a wireless too. I know what you mean about honey do's. I have a long list of them.
I got my new spindle cable installed. I used a ferrite core at the VFD end, in my quest to fend off noise. According to what I've read, I did it wrong the first time around. I've read that you don't include the ground wire in the windings around the core. Got it corrected this time.
I ran my spindle today. Seems quieter than before, but it's been a long while since a ran it, so I can't be sure. I don't have a means of measuring noise, which is a handicap. I have a TV close to my CNC. I sometimes plug my computer into it for a larger screen. Anyway, I was able to position the spindle to within about 18 inches of the TV and looked for noise it. Nothing but its normal good picture. Should I see noise in the TV, if I was putting any out? This is the sort of thing you know about, and I don't.
I'm setting up to do gantry squaring. I'll take some photos for my report out. After that, I'll do my spoilboard. Once the spoilboard is done, it's time to start cutting. I'm working on toolpaths for mounting the spoilboard and cutting dadoes for the t-track I'm planning to use. I have to inventory my supply of t-track. I don't think I have enough 4 foot lengths.
Gary
Re: GME's New 80/20 CNC Build - My Design
Gary I have both a Oscilloscope and RF Signal strength meter which is broad banded. But I have used an AM/FM radio in the past to check for noise when I was active in Ham radio. Did you have problems with the machine you had running before?
Re: GME's New 80/20 CNC Build - My Design
Quote:
Originally Posted by
wmgeorge
Gary I have both a Oscilloscope and RF Signal strength meter which is broad banded. But I have used an AM/FM radio in the past to check for noise when I was active in Ham radio. Did you have problems with the machine you had running before?
I didn't try a radio, but I'll give it a shot. Thank you for the suggestion.
No, I didn't have an noise-related problems before, so I didn't expect any this time around. The only changes I made were replacing the cable between the spindle & VFD, grounding the cable at the spindle (back when, I added a ground under the spindle cap), and changing the way I did my windings on the ferrite torroid core. All of the changes should serve to reduce noise, so I didn't expect to have a problem now. About the only additional thing I could add would be a noise filter on the incoming 240V line. They aren't particularly expensive, and not at all complicated to install, but I don't plan to go the time and expensive, if I don't need it.
Thanks again for the suggestion.
Gary
Re: GME's New 80/20 CNC Build - My Design
Quote:
Originally Posted by
qtvietnam
Thank you, the article is very helpful
You are welcome. I am glad to be of service.
More to come.
Gary
1 Attachment(s)
Re: GME's New 80/20 CNC Build - My Design
Quote:
Originally Posted by
wmgeorge
I have used an AM/FM radio in the past to check for noise when I was active in Ham radio.
Bill, I gave the radio a shot. On AM band, I got quite a bit of noise. I tried a couple of talk radio stations. With the spindle running at 18,000 RPM, I went from being able to clearly hearing the voices, to just being able to make out what they were saying. I then switched to FM. It was okay, Maybe a tiny bit of noise, but hardly discernible. The radio was 2-3 feet from the spindle during the test.
The machine itself appears to be working fine. No noise showing in my control computer. My TV is still not showing signs of noise.
How does your experience with Ham radio to compare to my observations?
Right now, I'm a little concerned about the noise, because I'm about to work on squaring my gantry. The technique I'm using requires running the axes over a large are and running the spindle to drill holes. If something is misbehaving, there is no way to know for sure. I'd hate to go to all the work squaring the gantry only to find out later that I was dropping or gaining steps and all the work was wasted.
To add a brief recap, I have an HY 2.2KW VFD, which is in it's own box and physically separated from my control box; spindle is grounded inside the cap; ground wire connects to VFD and runs from there to the star ground; Igus CF6 spindle cable with shield grounded at both ends - at the connector metal cable clamp at the spindle end and in an Icotek EMC Shield Clamp at the VFD end ; 3 hot spindle cable wires wrapped around ferrite torroid core - 8 turns each - between the Icotek shield clamp and the VFD. Here is a picture of the Icotek shield clamp, which is one that often shows up in mactec54's posts on grounding shielding in the VFD box:
https://www.cnczone.com/forums/attac...d=423756&stc=1
Gary
Re: GME's New 80/20 CNC Build - My Design
Bill, I gave the radio a shot. On AM band, I got quite a bit of noise. I tried a couple of talk radio stations. With the spindle running at 18,000 RPM, I went from being able to clearly hearing the voices, to just being able to make out what they were saying. I then switched to FM. It was okay, Maybe a tiny bit of noise, but hardly discernible. The radio was 2-3 feet from the spindle during the test.
The machine itself appears to be working fine. No noise showing in my control computer. My TV is still not showing signs of noise.
How does your experience with Ham radio to compare to my observations?
Just tune your AM to a station free area and turn on the spindle, if you hear a lot of static and don't when its off then you have RF leaking somewhere. Your TV is digital and may not be affected by the RF. I don't use those terminals, all mine are soldered... the shields that is. The power wires are crimped terminals. My VFD is in a metal enclosure with metal door grounded and the air intake vent is metal and grounded. Everything is grounded. I also have a better quality VFD, Hitachi and it was the second one I have used without issues. No line filters, never needed.
I would wait and run your machine before doing anymore, it might be fine.
Ham radio..... when we were trying to get distant stations like overseas Japan, Italy and etc the noise in the background noise would have to be very low. The power companies often did not give a rats behind if they had poor HV connections like on the 13.8 Kv primary. So I would take a AM radio and a sledge hammer and drive around until the radio had so much static you knew you were near a bad power pole wire connection. Walk over to the suspected pole and give it a wack with your sledge hammer, if it got worse then I would call it in. They were required by FCC and law Not to transmit interference. It raised hell with the AM and TV signals but nobody but us Hams knew where it was coming from.
4 Attachment(s)
Re: GME's New 80/20 CNC Build - My Design
I'm overdue for an update. Here it is. I squared my gantry.
https://www.cnczone.com/forums/attac...d=423944&stc=1
https://www.cnczone.com/forums/attac...d=423946&stc=1
Since I don't have my spoilboard mounted yet, I used two scrap pieces of plywood for my squaring routine. Yes, there are 3 pieces in the photo, but the one in the center is to facilitate using my Woodpecker Bar Gauge measuring device (stainless rods held together with the red blocks shown sitting at an angle in the photos).
I created several toolpaths in Vectric Aspire for the routine. The routines call for cutting 1/4" holes in a square pattern, 45" on each side. In the second photo, you will see a metal dowel pin in the 4th of 4 holes. I started by creating the first toolpath starting from the left, and subsequent toolpaths moving to the right, and maintaining the 45"/side square. That allows me to create a set of 4 holes, measure, make adjustments, and then rerun a new set of holes and keep track the holes I was working with. After cutting 4 holes, I inserted 1/4" metal dowel pins in each one. I had a good, tight fit, that required a hammer to tap in. The holes were cut to 3/8" depth.
As far as measuring is concerned, I checked the two cross distances with the bar gauge to determine which diagonal was shorter. I locked in the gauge to the shorter length. Then, I placed the gauge across the other diagonal and observed the gap between one end and the dowel. I used feeler gauges to measure the amount of the gap and give me the start figure for adjustment.
Here's how it works. Don't ask about the underlying math. I cannot explain it. It's been too many years since I studied geometry and trigonometry. I only know that it works.
1. Determine the size of the square you want to use and make up toolpaths for it. The larger the square the better. You get better accuracy. I would have gone with 48" X 48", but for the fact that put my machine together with a setback in the front to allow me to do dovetail work off the front. So, 48" x 48" wasn't in the cards for me.
2. As described above, cut the holes, insert dowel pins, and determine the difference in the distance between the diagonals.
3. Now the math.
a. Start by taking the difference between the diagonals and dividing it by 2 (half the error).
b. Divide the result in a. above by the length of the side of the square.
c. Multiply the result in b. above by the length of the gantry as measured by the center-to-center distance between the gantry mounting points.
This gives you the amount you need to adjust your gantry.
Here's and example.
Given: My gantry is 74" in total gross length. The center-to-center distance between the mounting points is 69.5". This is also the center-to-center distance between my Y axis linear rails. 69.5" is the distance necessary for this example. The error observed in the dowel-to-dowel cross measures is .030". Each side of the square pattern is 45".
Math: .030" dowel-to-dowel cross measure/2 = .015". (.015"/45") X 68.5" = 0.00033333333 X 68.5 = 0.022833333". So, with rounding, the amount to adjust the gantry is 0.023".
Why do you adjust the error to account for the gantry length? Take 2 intersecting line angled 5 degrees. 1" from the apex the two lines will be X distance apart. At 48" from the apex, the two lines will be Y distance apart, and Y distance apart. The Y distance apart is multiples of the X distance. So, you adjust the measured error to extend it to the length pertinent to where you will make the actual adjustments.
Here is a picture:
https://www.cnczone.com/forums/attac...d=423954&stc=1
So, how do you know which direction to move the gantry? Well, it depends upon which side of the gantry you want to move relative to the other side. Let's assume that when you measure the diagonals, the short measure is from the left front to the right rear and the long measure is from the right front to the left rear. That means that to adjust to square, you either need to move the left side of the gantry toward the front, or move the right side of the gantry towards the rear. If the long measure from the left front to the right rear, then adjustment is reversed, i.e., you either move the left side of the gantry to the rear, or move the right side toward the front. Another option is to split the difference, and move each side. Moving both side might be an option in cases of limited gantry adjustability.
How did I make the adjustments on my machine? Glad you asked.
1. I loosened all of the gantry mounting screws to finger tight, except for the rearmost outside screw on the side I wanted to be the fixed, pivot point. I left that screw just a little more than finger tight. It needed to be able to allow the gantry to rotate, but not slip around. (Note: having adjustability is a blessing and a curse. A blessing, because without adjustability you would be left with an out-of-square gantry. A curse, because having adjustability means there is want amounts to a degree of "slop", which can mean getting unwanted movement at the pivot point. Unwanted movement translates into an unsuccessful adjustment.)
I placed a dial indicator on the linear rail corresponding to the side I am adjusting, and placed the end of the indicator against the plate I am moving. Here are 2 photos of the setup:
https://www.cnczone.com/forums/attac...d=423948&stc=1https://www.cnczone.com/forums/attac...d=423950&stc=1
The lower plate is fixed, i.e., mount to the linear rail blocks. The plate above it is what moves relative to the lower plate, so my dial indicator zeros off it.
https://www.cnczone.com/forums/attac...d=423952&stc=1
Then, I used a common F-style clamp at a diagonal to move the gantry in the desired direction - in this case, toward the rear. The dial indicator measures the amount moved.
Because it is difficult to ensure that the opposite side remains stationary, this is a measure, adjust, repeat process. Fortunately, the errors tend to get smaller with each succeeding attempt. I dialed mine in on the 2nd attempt. Okay, you saw what appears to be 4 attempts in the 2nd photo, didn't you? I didn't count the first one, because I had a problem with the hole size.
When I started the process, I intended to take a 1/4" endmill, drill the holes with it, and insert the dowels. The best laid plans. . . . Turns out the endmill cut too large a hole and the dowel wallowed around in it. Not good. On to Plan B. I rewrote to toolpaths to use a 1/8" bit to cut a pockect 1/4" diameter. That worked. I got the good tight fit I wanted. So, scratch the first set of holes. Set 2 got me close. Set three got me to within .001 over 45" and set four confirmed it.
There was some earlier discussion about VFD/Spindle noise, and my attempts to eliminate, or at least minimize, it. It's still there, but so far so good. It looks like it's not so bad as to cause any problems - at least for now. If a problem shows up, I'll add a line filter.
Next up:
Spoilboard
Gary
2 Attachment(s)
Re: GME's New 80/20 CNC Build - My Design
I have a little problem I use some help with. I've found that it get quite a bit of vibration with movement. It's greater on the y axis, much less on X and no issue on Z. The vibration is more pronounced with short movements, i.e., under an inch. I reduced my velocity on Y from 600 to 300 and that helped some. I haven't played with acceleration yet, but I'm going to.
All this led me to conclude that I should have bolted the CNC to my concrete floor. Doing so now is a problem. Here are a couple of photos to illustrate what I'm up against:
https://www.cnczone.com/forums/attac...d=423956&stc=1
This photo show the type of foot I'm using. The photo is deceiving, though. It looks like the foot sticks out beyond the extrusion, but it doesn't.
https://www.cnczone.com/forums/attac...d=423958&stc=1
This shot gives a clearer picture. The feet don't stick out.
While not necessarily clear, I have a threaded plate on the bottom of the extrusion. The foot has a 1/2" threaded part that swivels some and screws into the plate on the extrusion. A pretty standard machine foot.
If the feet stuck out, I'd just drill holes though them and into the floor, epoxy threaded rod into the concrete. Not and option. Not enough room to drill.
The obvious solution would be to just start over - dismantle everything, add bolts in the floor and reassemble. Not something I want to do.
I recall reading posts from Dave Faulkner about his 2 x 4 Saturn 2 (welded steel) slamming around pretty hard in operation and needing a substantial, or at least rigid, base to hold it. My machine is not as heavy as a Saturn 4 X 4, but it's nearly twice has heavy as the Avid/CNCRP Pro 4 X 4. I don't recall this being an issue with the Avid machines, but I've haven't gone on an extensive read trying find anything as yet. I feel like I'm overlooking something - probably something obvious. Suggestions???
Any ideas short of a dismantle/rebuild?
Thank you in advance.
Gary
1 Attachment(s)
Re: GME's New 80/20 CNC Build - My Design
Recall the discussion we had a while back when I suggested adding diagonal braces? I would start by adding those to the sides that did not have some installed.
You've got slots in those legs, make up some 2 sided L brackets to fit and holes to match those slots. Weld at the sides. So you've enclosed two sides of one leg, both L bottoms bolted to floor, holes with bolts to slots. Trying to get Fusion 360 to export a JPG, but here is a screen shot. Pretty simple just make 2 for each leg, Weld at the long upright corner.
BTW the best present I ever purchased for myself was a Bosch rotary hammer/drill, works great as a 1/2 drill motor and wonderful as a hammer drill. Makes drilling in concrete a breeze.