[GME]

Yes, the first video looks like resonance or "chatter" or shaking as you call it.

I think that looks more like of a drive line issue. Steppers, drivers or the rack drives could cause all of them. I would suggest to look into those elements.

Thank you, David,

Thing is, it only does it at abrupt starts/stop, quick direction changes (like cutting holes with a pocketing toolpath), and when running at very slow speeds, e.g., 5 ipm on a Ref All. There is practically nothing when running at normal speeds. Not saying it's not the steppers, R&P drives or stepper drives, but if they have something wrong with them, wouldn't you expect some symptoms at 100-300 ipm?

The drives are Gecko 203Vs mounted to a huge heat sink which is exposed to ambient air, and with a fan in the control box. The 203Vs have a great reputation for reliability. I'm not sure what I would do to check the the Gecko drives.

FWIW, I was using all of these drive line parts on my previous machine, and all worked perfectly right up until I dismantled it last October. What supports bracing as a possible cause is the compete lack of resonance/chatter/vibration/shaking on my old 4 x 4, which used light weight 1530 extrusions (vs the 3030 full weight I'm using now). For the first machine, I built the base out of two 60" long x 24" deep cabinets. I made them out of torsion boxes for the sides, center divider and top. The torsion boxes, and the other cabinet elements were all made from plywood - no MDF anywhere. Each cabinet had 6 adjustable feet, and I made the feet out of carriage bolts and hockey pucks. It's a great combination, which helps dampen vibration. Nothing on the old machine moved (or vibrated, etc.) that wan't suppose to. So, while it makes sense to check the drive line, my recent experience with the running gear suggests it should all be A-OK.

Also, when I started this build I went over the R&P drives and looked for any looseness or play anywhere. None I could find. I removed the steppers from the drives and tuned them again to the Geckos. Frankly, i couldn't get much noise out of them through the full trim range. I adjusted to as quiet as I could and moved on. The steppers are the NEMA 34s CNCRP/Avid CNC sells. 960 oz-in, 7 amp and only 2mh inductance. Never a problem with them. I'm not sure what more to check, or how to check it.

And yes, I'm starting to wish I stayed with the tried and true and reused the cabinets. It would have saved the cost of all the leg parts, 80/20 milling charges, and the time I spent milling the extrusion ends on all the base parts. It was a considerable amount of time. As much to the point, staying with the cabinets might have prevented this problem from occurring. If I could bear the thought of breaking down the machine and going through another full assembly, I'd just change back over to the cabinets and call it a day. Depending upon how it goes, I may end up there anyway. I sure hope not.

Why break down the machine, you might ask? Because the process of lifting it would probablyl tweak the flat/square/level I worked so hard to achieve, and would require disassembly and reassembly anyway. There is a lot of weight, and it's just bolted together. Even a couple of thousands off would mean a teardown. Yikes!!! Just mentioning it makes my head hurt. :drowning:

Gary

[ger21]

Thing is, it only does it at abrupt starts/stop, quick direction changes (like cutting holes with a pocketing toolpath), and when running at very slow speeds,

Again, I think these are two different issues.

For the homing, have you tried adjusting the trim pots on the G203V?

[wmgeorge]

Thanks, Bill. Good idea.


I will try where CNCRP put the brace on the right Y side, but I believe the height was probably driven by the desire to have a place to mount the control boxes. I'm thinking putting it at the height of end braces (quite a bit lower) may yield better results. Also, I already have corner braces high up, so bracing down low may yield better results.

More to follow . . . .

Gary

Remember your Experimenting, and its hard for a perfectionist to do so but you must not decide ahead of time what will work and what won't. Your not aiming to fix just see what works better than the other. The fix will come when you add those permanent ones. Who said the bracing can't be as I suggested aluminum angle, attached to the outside with your "roll in nuts" made for the extrusion.

Bottom line since the exact same setup motors, drivers and I assume even the settings were running without issues on your old machine. So it has to be the bracing, or at least most of it anyway.

[GME]

Again, I think these are two different issues.

For the homing, have you tried adjusting the trim pots on the G203V?

Thank you, Gerry. Yes, I adjusted them before I reinstalled them in the R&P drives, which was just before I installed the steppers/stepper drives on the machine. I can do it again at homing speed, though. Do you recommend that I leave the stepper drives connected and use the vibration (resonance) to guide my adjustment, i.e., adjust the trim pots until I get the vibration/resonance at the lowest possible level?

Gary

[ger21]

Gecko recommends adjusting them at a certain RPM (check the manual). And it should be done on the machine.

[GME]

Remember your Experimenting, and its hard for a perfectionist to do so but you must not decide ahead of time what will work and what won't. Your not aiming to fix just see what works better than the other. The fix will come when you add those permanent ones. Who said the bracing can't be as I suggested aluminum angle, attached to the outside with your "roll in nuts" made for the extrusion.

Bottom line since the exact same setup motors, drivers and I assume even the settings were running without issues on your old machine. So it has to be the bracing, or at least most of it anyway.

Thanks, Bill. I'm just thinking through the various possibilities and using what seems the more likely prospect as a starting point. I'm going to use some over length pieces of 1530 extrusion in place of 2x4s and bolt them to the frame with some 1/4 thick aluminum angle. I should get a better connection than I would with clamps. I plan to try a variety of placements to see what seems to work best. I've already designed mounting plates for a more permanent fix, once I figure out what works best. I found some t-bolts that will make the process a lot easier than using roll-in t-nuts. It will make more sense when I photo document what I try.

Gary

[GME]

Gecko recommends adjusting them at a certain RPM (check the manual). And it should be done on the machine.

Thanks again, Gerry. Yes, I used the recommended RPM when I adjusted the steppers before. I'll do it on the machine this time. I appreciate your help.

Gary

[wmgeorge]

Deleted

[GME]

But you completely missed the point...

Did I, Bill? I thought the point was to experiment, i.e., try a number of bracing combinations to see if any make a difference, and if they do, select the one that makes the most difference. Then, take the best of what I find through experimentation and turn it into a permanent fix. That is what I am doing. There are any number of ways to go about it, but the bottom line is the same. What works best. Of course, I am searching for a permanent fix, but the experimental process is just using temporary means to hopefully lead to a good, permanent end. The fact that I am using aluminum angle to temporary fasten braces to the machine should be of no consequence, at least in my mind. The 3" lengths of angle are nothing more than a substitute for clamps.

Don't read too much into the fact that I have drawn up some mounting plates for a permanent fix. Whatever the permanent fix turns out to be, and I don't know what it will look like, I will have to fasten it all together by some means. My plates contemplate something like the base under your machine. If it turns out that I have to go with diagonals, rather than horizontals, or something else, I'll just have to modify the mounting plate design. No big deal. I'm getting pretty good at CAD, but extra practice is good for my development. At the furthest outside extreme, it might turn out that the best solution is fully sheathing all four sides with 1/4" (or thicker) metal. Who knows at this point. Everything is on the table.

Help me out here, Bill. What is the point I'm missing???

Gary

[GME]

I'm red faced with embarrassment having to admit something.. When I previously adjusted my steppers, I blew the velocity setting. The Gecko literature recommends adjusting the drives at 1/2 revolution/second. When crunching the numbers to translate 1/2 rev/sec, into IPM, I carelessly managed to plug in 2 revs/sec, 4 times too fast. No wonder I wasn't hearing much noise variation when I was adjusting previously. I believe it safe to assume that the drives are way the heck out of adjustment.

Good call, Gerry and DavidA1234. And to David, I apologize if I seemed to be pushing back on the idea of rechecking the drive line. Actually, I was going follow your suggestion, but I was pushing back against the idea nevertheless. I forgot that mistakes can and do happen. Hence, the red face.

I am going ahead with the bracing before readjusting the Geckos. If there is any non-resonance noise, I don't want it playing its tune while I'm adjusting the drives. I want the least amount of competing noise.

Gary

[GME]

Good news; bad news.

First, the bad news.

My experiment with adding side bracing did not produce the result I hoped for. I moved the braces all over the place. The best results were down low at or near where I have the cross braces in the front and rear. Pretty much what I expected. If I had to attach a number to the improvement, I'd guess it about 25-40% better. I ran my gantry squaring routine again so I had a baseline to compare to. Definitely less movement on quick turns and round pocket milling, but there was still movement. I ran the gantry squaring routine with the wrench on the crossmember. There was less vibration. Not perfect, but better.

As Gerry pointed out adding rigidity and mass are the ways to conquer the issue. On the plus side, a permanent installation with corner bracing should add more rigidity, and that should give even more improvement. Adding 2" of spoilboard should also help (added mass).





Please pardon the mess. The 2 pictures above depict the temporary braces in place. What you can't see are the pieces of aluminum angle used to hold the braces in place. The angle is bolted to the legs and bolted to the underside of the oversize 1530 extrusion. 2 bolts in the leg; 2 bolts in the extrusion. The way I connect the the brace seems pretty rigid, it can't be nearly as rigid as it would be if everything was held together with a gusset plate mounted at the corners.



Here is a picture of the gusset plate I designed in Aspire. It is 4.5" x 7.5" It's rotated 90 degree from how it will mount. The edge with 5 hole will be oriented horizontally. Six holes will mount it to the legs, and 6 holes will mount to the brace. I'll be cutting down the 1530 extrusions I experimented with, so the extrusion will fit between the legs. Although the picture depicts the gusset in wood, it will actually be made from 1/4" thick aluminum.

I suspect that this is probably a lot small then Gerry contemplated when he suggested I add MDF corner bracing. He didn't suggest a size, so I'm just guessing. Anyway, I expect aluminum to be pretty stiff. I hope it will do the job. If I don't like the results, I can make it bigger. This little addon gets expensive in a hurry, though. T-studs run about a buck apiece. Roll in t-nuts run $1.50, plus BHCSs. Since I will be using the gussets on all 4 sides (8 gussets in total), I'll need 96 of whatever I use. And all the hardware is exclusive of the cost of the aluminum. Yikes! I sure hope it ends up to be worth it.

Now, the good news.

1. I got my steppers in tune. I did all 4 while I was at it. They were all a bit out, but not drastically so. Getting the formula right for 1/2 revolution/second helped. I know there are different ways of doing it, but I decided to change my homing approach speed to what I needed (24.451 IPM), moved the axis so I would have plenty of room for movement, and performed a Ref All. It worked great, although having my head close to an open controller box with the fans running was interesting. Especially so, because I wear hearing aids. I was still able to hear the sounds coming from the steppers. I just took my time and slowly dialed them in.

2. I ran a regular homing routine after tuning with the wrench sitting on the extrusion - just like was shown in the video. Even with the added bracing and steppers tuned, I was still getting resonance/vibration, close to as bad as shown in the video. What to do; what to do? I decided to play with my homing speed to see if slight changes would make a difference. By golly, it worked. By increasing my homing approach speed from 5 IPM to 10 IPM, the problem went away completely. Zip, nada, nothing, gone. Fortunately, 10 IPM isn't an unreasonably high homing speed, so I'm good. From what little I understand about resonance, I'm guessing 5 IPM was setting up a resonate frequency and changing to 10 IPM got me beyond it. Please pardon me, if I'm not describing it correctly, but hopefully the general idea makes sense. Gerry and DavidA were spot-on with their diagnosis.


If I had it to do all over again, I would consider filling all the extrusion cavities with epoxy granite. It would add both rigidity and mass. It's a time honored method for smaller metal mills, and it's even employed by machine manufactures. Of course, the additional weight would also make assembly of an extrusion-based CNC more difficult. In any event, I'm not about to tear the machine apart and start over, but this may give others food for thought.

Any suggestions or comments?

Gary

[davida1234]

By increasing my homing approach speed from 5 IPM to 10 IPM, the problem went away completely.

Gary, good to hear you got the resonance out for homing.

However, IMO, the problem is not solved yet, you just went around it. Every time the machine slows down to 5 ipm while cutting (just about every curve), you might get this resonance which might affect the cut quality.

I would suggest to really try to find out where this is coming from. Is the mesh between pinion and rack too loose? Any play with the pulleys?

As to the brace, just get a sheet of 1/4" BB ply or 1/8" aluminum and cover the whole side frame with it. That would be a real brace. The brackets you are suggesting will be limited in effectiveness.

[wmgeorge]

You did not bolt the braces directly to the uprights but held underneath by aluminum angle?

[GME]

Gary, good to hear you got the resonance out for homing.

However, IMO, the problem is not solved yet, you just went around it. Every time the machine slows down to 5 ipm while cutting (just about every curve), you might get this resonance which might affect the cut quality.

I would suggest to really try to find out where this is coming from. Is the mesh between pinion and rack too loose? Any play with the pulleys?

As to the brace, just get a sheet of 1/4" BB ply or 1/8" aluminum and cover the whole side frame with it. That would be a real brace. The brackets you are suggesting will be limited in effectiveness.

Thank you, David.

The pinion is firmly meshed in the rack. Although I didn't mention it before, I tried loosening and tightening the tensioning screw (mostly tightened) to no effect. Nema 34s with this R&P system have been known to try to jump cogs on acceleration. I tighten mine enough to keep that from happening. It probably means more wear, it is what is. I'm not seeing any evidence meshing problems going on. No appreciable wear on the pinion. Well lubed.

When I had the R&P drive off the machine, before I installed on this one, I check for play with the steppers and belts removed. I couldn't feel any. I just dropped the drives (belts on this time). All seems fine.

I completely get your point, and your suggestions are clearly on point. I just don't know what's left to check. I suppose I could try changing out the drive spindle (the drive and pinion gears are cut out in one piece), but that's more than a $100 bill on a questionable "maybe." I'm not thrilled at the prospect.

I really hate to sheath the sides, but maybe that's the best way to do it. I don't have any 1/4" BB on hand, but I have some 1/2". The pieces are 30"X60", which is nearly the perfect size. Any reason why 1/2" wouldn't work? Would you anticipate a problem with the bottom edge being unsupported? I can always bolt an extrusion to the plywood, but won't bother unless benefit would exceed the cost. It will be PIA on the right side. I have my cable track on that side, which means having to take it loose (with all the wiring) and all that the process will entail. Oh well . . . .

I would really rather sheath in aluminum, but I'd need to buy a 48" X 144" sheet. Expensive stuff - about $500 in these parts.. It kills me to have .080" laying around and not be able to use it. Too thin, though. BB ply it is.

Disappointing to hear that my gusset arrangement won't provide much benefit. FWIW, it's similar to the Fineline Saturn 2 base, which has apparently been working okay. Lots of complaints about the Saturn2, but none about the base. Bill George has been happy with his setup. Of course, the Saturn is much heavier, and the base is steel, not aluminum. I already ordered the aluminum. Fortunately, it wasn't too expensive. I'm confident I will find another use for it.

Thank you so much for your help.

Gary

[davida1234]

As a planar member, 0.08 would work but it would buckle under stress. You would have to support it throughout.

No problems with the 1/2" ply.

Isn't the Saturn base welded? It is more elaborate get the stiffness of a weld connection with bolts.

[GME]

You did not bolt the braces directly to the uprights but held underneath by aluminum angle?

I oriented the angle so one outside face was against the leg (upright) and the other outside horizontal face was facing up. I bolted the angle to the legs and then bolted the angle to the underside of the brace. Seemed to me like a more solid/rigid connection than clamps. Plus, with clamps, I would have to take great care to keep from crushing the extrusion. Too tight, and there is a crush issue. Too loose, and the brace can move. I reasoned that the angle was much more forgiving. Also, I wasn't looking for a perfect fix. I was just testing around to see if there was any improvement, and if there was, where I saw the most improvement. I also reasoned that a permanent fix, like with the gussets, would give me at least the same amount of improvement, and almost certainly, much more. 12 bolts with a gusset has to be way more rigid than 4 bolts through a piece of angle - at least that's the way I viewed it.

There is no way to bolt the braces directly to the legs without drilling holes through either the braces or legs. Nothing "quick and dirty" about that. It would have required perfectly aligning the holes with the slots, and with an overlong extrusion, it would have required much more time than I wanted to devote.

Make sense?

Gary

[wmgeorge]

Well the aluminum is 15mm wide and bolted to angle 76 mm long you can do the math for the surface area vs the 30mm side bolted to the 76 mm upright. About double the holding surface area and that’s what counts.
Clamping? Why would you bend the aluminum, you were going to use protective blocks, correct?
Since you have basically copied the CNCRP design and some of FLA design and neither one uses side panels, why would you need?? Both of those designs work, including your first machine.?

[GME]

Isn't the Saturn base welded? It is more elaborate get the stiffness of a weld connection with bolts.

Nope. The Saturn router frame is welded, but the base isn't.





Here are two photos Bill George contributed earlier in this thread. My gussets are similar, except I designed mine to have two rows of bolts on both the horizontal and vertical members. Bill said the horizontals were 1.5x1.5" steel tubing. That comports with my recollection from the time I briefly had one in my possession (it was irredeemably defective). I was going to use 1530 extrusions for the horizontals, or double the size, but in aluminum.

Gary

[GME]

Well the aluminum is 15mm wide and bolted to angle 76 mm long you can do the math for the surface area vs the 30mm side bolted to the 76 mm upright. About double the holding surface area and that’s what counts.
Clamping? Why would you bend the aluminum, you were going to use protective blocks, correct?
Since you have basically copied the CNCRP design and some of FLA design and neither one uses side panels, why would you need?? Both of those designs work, including your first machine.?

Sorry Bill. Your reference to metric suggests I may have confused you with the 8020 shorthand. 1530 is 1.5" x 3". It's the abbreviation for 8020's imperial sizes. The extrusion I used had a 1.5" x 3" profile times the length needed. The aluminum angle I used was 1.5" X 1.5", so it matched up with the bottom of the extrusion in the orientation I used.

Yes, I would use clamp blocks with C-clamps. It's just that it's easy to torque down too tight with well-oiled C clamps - especially when using 2 on each end. A single clamp on each end would leave pivot points, which wouldn't be good. I could have made a more rigid connection, if I had used a second piece of angle on top of the extrusion. I thought about it, but didn't want to go to the bother of cutting and drilling more angle. What I used I had on hand. I'm satisfied with the way I did it. Not the best way, but an adequate way for what I was trying to accomplish.

I didn't use legs on my first machine. I used two 60" long cabinets made with torsion boxes on the ends, center and top. They had 3/4" backs inset between the torsion boxes. You couldn't rack them if your life depended on it. The two bases were heavy (even though they used torsion boxes) and solid as a rock (as you would expect with torsion boxes). I used 6 hockey puck leveling feet on each cabinet (12 between the 2 cabinets). Base movement was never going to be an issue. The machine set perfectly flat on the base and couldn't move.

Side panels, especially using aluminum. is probably the five star fix. However, the cost of the aluminum needed to do the job is prohibitive. I don't like the idea of using plywood on an all metal machine, but I'm unwilling to part with the bucks for aluminum.

I already have the aluminum for the gussets on order. It will probably arrive on Monday. I am going to go head and cut them out and give them a shot. Nothing ventured, nothing gained. If I don't like the result, I have another project they will work on. They are way overkill for the other project, but they won't go to waste. I already have the BB ply, so I can change directions without additional cash outlay.

Gary

[davida1234]

Nope. The Saturn router frame is welded, but the base isn't.

Well, then go ahead and try the gussets, maybe it is enough.