[wmgeorge]

If you do an online Search you will find disagreements even among the professionals. Of which way is right. Arguments all over about grounding at one end or both and etc. Some from the manufactures, some from the wire makers and just general confusion. Are you having issues with your system?? Otherwise if its working just move on to something else.

I have been in Ham radio or amateur radio, building radios and antennas, licensed after passing a test since 1977. I was taught in the USAF how to do cable shielding and wiring on air craft in 1962, good enough for them.

[GME]

If you do an online Search you will find disagreements even among the professionals. Of which way is right. Arguments all over about grounding at one end or both and etc. Some from the manufactures, some from the wire makers and just general confusion. Are you having issues with your system?? Otherwise if its working just move on to something else.

Good afternoon, Bill,

No, I didn't have a problem in the past, and I previously did not ground the shield at the spindle end. Nevertheless, I elected to ground the shield this time around. If I get a noise problem, it's easy enough to clip the shielding back and tape it over. I don't anticipate a problem.

As I think I mentioned in an earlier post, I previously wrapped the cable around a ferrite core inside the VFD box, which isn't something you see discussed much. I picked the suggestion up in a post, quite by accident, when reading about someone having problems with noise. Knowing that it was unlikely to cause problems, I decided to give it a shot. Did the core prevent noise in a cable grounded at only one end? I have no idea. I'll be using again, though. I don't think it can hurt anything, and as long as I have it, I might as well use it.

I don't want to leave the impression that I'm obsessing over the shield grounding. I'm not. If I'm obsessive about anything, it's noise in general. It's because I've read so much about people who had problems with it. That was/is the last thing I want to deal with. By all accounts it can be a bear to diagnose, and I don't pretend to be a good diagnostician. Anyway, It was just something I wanted to take care of as long as I was replacing the cable anyway. I went with the Igus CF6, because it is more flexible than the Igus cable I had been using. I've also engaged in the discussion for general reference for others who might be interested. I used to be one of those lurkers who read a lot and contributed little. Given the number of folks who have viewed this thread, but remained silent, there are lot of them out there. Not criticizing, but recognizing how it is. Even though I am not an expert in electrical devices, I can share what has worked for me. If all this helps just one person, I'm happy. At the end of the day, my cable replacement was but one step in the build process. Nothing more.

I acknowledge and agree with your comments about the pros being all over the place on how grounding should all be done, and how confusing it can be. Makes you think that whatever you do, it will all work out. Unfortunately, we know from all the people with noise problems that not all of the solutions are necessarily the best, or even work. Faced with conflicting choices you are left with having to pick one that seems reasonable and try it. That's what I've done. So far, so good.

I'll be posting some pics of my plug and the ferrite core I used a little later on, in case anyone is interested. I think I mentioned before that my connector is made for a larger cable, so I had to make some modifications to get what I wanted. I really didn't want to deal with opening up the spindle again and changing connectors - and spending even more money.

Next up: I'll be adjusting my sensors and sensor flags. Not a big deal, but part of the process. IMO, sensor flag mounting was a weakness in the Saturn 2 design, because they had virtually no adjustability. Since I am using extrusions, it was easy to make the flags adjustable. It is going to a little tricky, though. The flags sit under the gantry interface plates when the sensor activates. Getting under the interface plates to adjust them may prove interesting. We'll see.

Gary

[GME]

Spindle connector.

I'll start with a simple plumbing fitting that made possible grounding my VFD cable shield inside the spindle connector. Here's a picture:



Yup. It's a simple copper tube strap. I started off by putting it in my drill press vise (not one of my precision mill vises) and tightened the vise down until the strap flattened out. This isn't soft copper, but it flattened out without much effort. I then took one of my bodywork hammers and flattened out that slight raised portion shown in the center of the bend. Once flat, I used emery paper to get off all traces of oxidation. Using the connector clamp as a guide, I cut the strap to length. I used one hole, enlarged it slightly and drilled a 2nd hole. Deburred and it was ready to mount.

Here is a photo of the connector I used. Sorry about the quality. My camera is an older Canon Power Shot. It takes decent photos generally, but doesn't focus on closeups.



You can just make out the copper strip showing around the edges of the cable clamp.



Here's a shot from a different angle. If look hard, you can see just a little bit of shielding along the top of the connector. I stripped the covering off to the length I needed. I peeled back the shielding, and snipped off all but about 1.5" or so. I took what was left of the shielding and carefully folded it back on the outer cable sheath. I didn't separate the shielding strands any more than necessary to fold the shielding back. I wanted to keep it as tight as possible around the sheath. Not difficult, but it took a few careful minutes. I disassembled the connector (this one comes apart in several pieces - see the screws at about the midpoint), soldered the conductors, and put it back together. Checked the connections visually and with a meter. Done!.

A quick word about this connector. The cable clamp part of it came painted and with convex ridges designed to hold the cable. I was after better contact are, so I filed off the ridges on both halves of the claim and continued filing until I removed all of the paint in the contact area. The soldering pins look like they were sized for something like 12-4 awg wire, while the cable clamp looks designed for something closer to 10-4 awg. It was tough finding a suitable connector.

In my last post, and in an earlier one, I mentioned using a ferrite core on the VFD end of the spindle cable. It's also called a toroid ferrite core, given its shape. I don't know much about these things, except that reduce noise. I also learned thay can be made from one of two materials. One is supposed to cover higher frequencies better than the other. I just used the one suggested. Here is a photo of the core I used:



Gary

[nlancaster]

wow, that is the bulkiest spindle connection I have ever seen.

AVE comes to mind on those pictures.

[wmgeorge]

wow, that is the bulkiest spindle connection I have ever seen.

AVE comes to mind on those pictures.

I just ran the heavy duty shielded VFD cable to a oil/water tight gasketed J box mounted directly to my moving Z carriage. Used a smaller cable from there to my OEM spindle connector. The motor is grounded with the clamp that holds the spindle (its all aluminum) and its grounded via a ground wire back to that J box and its connections. The spindle holder supplies over 18 sq inches of grounding surface to the motor frame. Simple and it works.

Added: Perhaps because I did not use a Chinese VFD has something to do with low noise and no issues. It cost me another $100 or so more to get my Hitachi, same one I used on my Emco 3 phase metal lathe without any issues.

[GME]

wow, that is the bulkiest spindle connection I have ever seen.

AVE comes to mind on those pictures.

Yeah, it's pretty big and ugly. Interestingly, it's similar in appearance to one of the Amphenol industrial connectors. Probably a knockoff.

Actually, it is described as a 20mm aviation connector.

I like the looks of the plastic ones a lot better, but don't like the fact they are plastic. Big & ugly, but it's all metal and works.

Gary

[GME]

I had some more time with my machine. I did some adjusting of the Pepperl & Fuchs proximity sensors and targets. I admit to some surprise using 1/8" steel targets. The sensors barely overlap the targets when they trip. Those who use the P&F sensor probably aren't at all surprised. This is my first time using them. On my last machine, I used large (18mm), cheap, Chinese sensors, and I was using them in a head-on orientation. I used the large ones, because they had a larger sensing distance.

What follows is a picture of the sensor and target on Z axis. It shows the location when it tripped. The vast majority of the sensor is still exposed.



Gary

[GME]

I verified the steps per and measured the backlash on the Avid CNC Ballscrew Z axis. I found one thing strange. I went to the Avid site and used their XML file selector for Mach3 to find out the steps per they preloaded for the Z axis. It was 5080 ipm. I set Z at 5080 in Mach4 and used a dial indicator and Mach4's steps per wizard. Surprise! When I commanded a .75" move, i got exactly 50% of that, or .375". Huh??? Naturally, the wizard told me to double my steps per, so I did, 10160. Set up my dial indicator again and commanded .75". ABSOLUTELY PERFECT!!!!:banana:

Next, I set up my dial indicator again and moved Z downward until it depressed the indicator plunger about .75". I rezeroed the indicator and commanded a .5" upward move. The indicator read exactly .5" of movement. Any backlash shows up on direction changes. NO MEASURABLE BACKLASH!!!! :wee:

If anyone is wondering, I started with the Z axis, because it is the easiest to check and isn't dependent on a rack and pinion to set up. X is the next easiest, because I am only dealing with one stepper. Y will take much more time, because I am dealing with a Master & slave, and cannot Ref All Home until I loosen all of the gantry screws. If I tried to Ref All without loosening the gantry, either the gantry would either have to be dead on perfectly aligned, or referencing all would put it into a bind - maybe a very bad bind. I don't want any binds and the stress it would cause. Better to loosen the gantry, reference all, tighten the gantry back down and move on to getting Y mechanically square.

I'll post a photo of the the dial indicator setup for the Z axis when I'm next in shop. No magic to it, though. Just mount the indicator with the plunger facing up under a movable part of the axis and measure away.

I have to say that I'm really impressed with the quality of the Avid CNC Z axis. Well thought out. Little known fact: Avid CNC manufactures the Z and other parts in their manufacturing facility in California. I just learned it today in a conversation with the folks at Avid. I knew their main facility was located up the road from me in North Bend, WA, but I didn't know they had manufacturing facility. I just assumed they had companies building parts for them. I gather they are going to shoot some videos downstream showcasing their California facility.

Gary

[ger21]

Gravity can "hide" Z axis backlash, especially with a low friction system.

[GME]

Gravity can "hide" Z axis backlash, especially with a low friction system.

Thanks, Gerry. Good point. Suggestions to get around the gravity issue? I can measure for backlash going up, reset and measure going down. Other than that, I'm stumped. FWIW, the Avid Z axis has what Avid termed a "one way brake," which is intended to prevent downward drift, when you shut down the machine.

Gary

[davida1234]

If the Z-slide is moving smoothly at this point and has no measurable
backlash (gravity aside), you could run it a bit and try to measure it then?

[GME]

If the Z-slide is moving smoothly at this point and has no measurable
backlash (gravity aside), you could run it a bit and try to measure it then?

Thank you, David. You suggestion makes perfect sense. On my first machine, I ran it for awhile, then rechecked everything, including backlash, gantry square, and even steps per, to make sure nothing had changed.

I'm very close to being up and running. The grounding clips to ground my VFD cable at the control side just arrived, so I guess I turn my attention to getting the spindle wiring finished. Not too much to it, so it shouldn't take long.

After that, I will be engaging my X and Y axis drives, get it homing and the gantry squared to the Y axis. Unless I am very lucky (I rarely am), squaring the gantry will take a considerable amount of time. The method I will be using involves drilling 4 holes at the corners of a 48" square, inserting metal dowels, taking relative measurements, and adjusting until the two cross distances are equal. I built a lot of adjustability into my design, which is a blessing and a curse. A blessing, because I can adjust as much as I need to. A curse, because it's really fussy trying to hold gantry interface plates in position while trying to adjust a 74" gantry beam. Trying to move the mounting plates (they sit on top of the interface plates) into the correct position can be tough. Friction fights you every step of the way. Been there done that, except my current design will have much more friction to contend with. At least I have a good idea of what to expect - assuming no big surprises. Things have good too smoothly to this point, so I'm overdue for a blindside.

Oh, and I have to get my spoilboard together. I really like your use of HDPE, but it's just too rich for my blood on a 4 X 4. I think I will be using 18mm Baltic birch for a base layer. I plan to cut dados into it to accept T-track from Orange Aluminum. (If you ever need T-track, they have the best prices around, and it's great quality). I'll then glue 3/4" MDF to the plywood and route out slots matching the channel in the T-track. The MDF will overlap the edges of the t-track. That way, you don't need to bother screwing the t-track down. I did something like it with my first machine, it worked great. The only problem I had was with t-track spacing. I had the tracks about 8" apart - way too much in practice. This time around, I plan to space them 3" or 4" apart. If my explanation is a bit fuzzy, it should make more sense when I report out on it in this build thread.

Awhile back, I read where someone suggested loosening the gantry fixing screws, making adjustments, and then running the gantry back and forth before tightening everything back down. At the time, the suggestion didn't make any sense to me. Now, as I face squaring, it makes perfect sense to at least try it. I believe the idea was to use the movement to let everything settle in before tightening the fixing screws. It took me awhile to get it, because the author didn't explain the idea behind the technique. The posts just described how to do it, and commenting that it was something the pros setting up big machines do. Funny, but the technique stuck with me, so even though I didn't at first grasp the why of it. I guess in the back of my mind I intuitively knew it might be useful.

Gary

[davida1234]

I had my gantry screws "finger tight". I used the clear plexi sheet method to check for squareness. It is somewhere in my thread, I believe in the 40's. Very easy to do.

[GME]

I had my gantry screws "finger tight". I used the clear plexi sheet method to check for squareness. It is somewhere in my thread, I believe in the 40's. Very easy to do.

Thank you, David. I recall reading your post about your plexi method when you posted, but for the life of me, I cannot find it. The Zone's server has been painfully slow the last couple of days, which make searching difficult. It's so slow, I keep timing out. Not sure what' going on, but I'm not having problems anywhere else on the web.

It would be great, if you could describe you technique again. When I read your post back when, the technique didn't stick.

Thanks again,

Gary

[GME]

What follows are a couple of photos of the setup I used to measure steps per and backlash for my Z axis. Note the steel piece clamped to the 80/20. It provides a good surface for the magnetic base for my dial indicator and keep it all rigid.





You can also clamp the setup to the fixed part of the Z axis and measure from the top. I've done it both ways in the past.

Gary

[GME]

I reused the R&P drives and rack and pinion drives from my first machine. So, the steps per should have remained the same, shouldn't they? Just for grins, I decided to recheck the steps per for the X axis anyway. My measurements came out just under 1 microstep different. How could that be, given that I used the same measuring technique I used before? I don't know. What follows is a description (with photos) of how I measure steps per. IMO, way more accurate than using a tape measure, which seems to be the most common technique I see.

I use 1 2 3 blocks to measure steps per. I suspect I have way more 1 2 3 blocks than most folks have laying around. Why? While I already had 4 tucked away, I bought 10 more specifically for setting steps per. Was it worth the cost? In my quest for accuracy, it was worth it to me. It may not be worth it to you.



Here's the initial set up. I clamped the dial indicator to to the router mount. I then laid out ten 1 2 3 blocks as shown. Note: As indicated above, I have more than 10, but I couldn't find the other 4 blocks. The more blocks the better, because the longer the measuring ranges, the more accurate the result.

Note that all of the blocks are laid flat, except the block closest to the dial indicator. The one on edge is to put the indicator touch point higher hand the rest of the blocks. I zero the indicator near the top of the 1 2 3 block that's on edge.



In this photo, you will note that I moved the block that was set on edge from the front to the back of the line. Needless to say, care must be taken when removing the the block, so you don't disturb the dial indicator or the blocks lined up.

I use Mach4, which has a steps per wizard. You input your current steps per, how far you want the axis to move, how fast you want the axis to move, and which axis you are dealing with. You then tell it to move and see how far the gantry actually moved. Dial indicator will tell you whether you moved too far (more than zero) or too little (less than zero). Input the correct actual movement in the wizard. There is a recalculate button, which, when you hit it, recalculates what your steps per should be. Mach3 will do mostly the same thing, although Mach3 will automatically update the steps per for you. Mach3 doesn't give you the option of setting the movement speed. In Mach4, you have to update manually.

With 10 123 blocks oriented as shown, the movement is 30 inches.




This photo shows the dial indicator after I commanded the axis move in the Steps Per wizard.

I could have set this up to move a longer distance using the same 10 123 blocks. To do that, I would have decided how much farther I wanted to move and clamped the appropriate 123 block in place. Then I would remove all of the blocks closest to the dial indicator back to the clamped one and place them at the other end of the line. I didn't do that, but it's an option.

I find that you have to run this, update your steps per, and rerun it. For some reason, it doesn't usually come out in a repeatable way the first time. The same was true when I was running Mach3. At the end of testing and retesting, I was within .0005" over 30". Note that while my indicator reads down to .0005" (resolution), the indicator itself is not accurate to that level of precision. That's why longer measurement distances are far better than shorter ones. Whatever small error an indicator has, when spread over a long distance, the error becomes insignificant.

Example: An indicator with .001" accuracy could be off .001/inch. Multiply the error by 48" inches (my cutting envelope), and the error becomes .048" at 48". That's more than 1/32" inch. Even for woodworking, that's more than many of us find acceptable. If you measure over 30", as I did, .001" accuracy translates to 0.00003333333/inch, or .0016" over 48" Pretty darned good. Longer is better.

Yes, I know. I went very basic for those of us who have been doing this stuff for awhile. However, I suspect that newcomers to CNC may appreciate basic.

Gary

[GME]

Last night, I engaged all of my R&P drives and took a test drive. It went without incident and what I hoped for. I had already moved everything by hand over every inch of travel - many times. No tight spots anywhere. After confirming every moved as it should. I engaged the R&P drives, loosened the gantry fixing screws and referenced all home. I didn't expect any problems, and was gratified to see that it homed as it should. Every small step that's successful feels like a big victory.

Gary

- - - Updated - - -

[davida1234]

Dan, the clear plexi method is on page 61.

https://www.cnczone.com/forums/diy-c...op-cnc-61.html

https://www.cnczone.com/forums/diy-c...op-cnc-61.html

- - - Updated - - -

[GME]

Dan, the clear plexi method is on page 61.

https://www.cnczone.com/forums/diy-c...op-cnc-61.html

https://www.cnczone.com/forums/diy-c...op-cnc-61.html

- - - Updated - - -

Thank you, David. I appreciate your going to the trouble of locating the posts and sending links.

Your method is certainly is fast and easy. For me, the problem is with relying on visual inspection to judge it square or not square. I've had parallax mess things up for me on any number of projects, Even thin plastic is enough for me to be off by several thousands. Also, just the thickness of the sharpie line is enough for me to miss square by several thousands. Right now, I'm shooting for as close to perfection as I can get it.

Gary

[wmgeorge]

This is what I use but not on most wood projects. 11HC-12-16R 12" Combination Square with Square and Center Head

- - - Updated - - -