[peteeng]

HiGary - The sound deadening material is not the go. This absorbs environmental sound, not structural vibration. Damping tape is heavy and viscous. So it increases the mass of the surface and if the surface does vibrate the materials internal friction dampens the motion. It behaves like leaf springs. As this type of spring moves they rub together and friction between the leaves works against the motion to damp it. Persistence will pay off and what is your gear ratio on the motor to the rack? Peter

[pippin88]

Putting steel rods in is a bad idea.

Adding weight (mass) is not the answer!

You need to add stiffness and/or damping.

Old heavy cast iron machines are good not because they weigh a lot. They are good because they are stiff (big sections) and have good damping from cast iron.

It is a common misunderstanding that adding mass cures resonance.

As Peter has pointed out the damping tapes work because they have a material that damps vibrations.

Sand is often recommended here. The movement of the grains against each other helps to dissipate the vibration (damping). I haven't seen the effect quantified in my reading.
(Keen to see a source if anyone has one)

(It's damping not dampening).

[peteeng]

Hi Pippin - Traditionally adding mass is the easiest way to reduce vibrations. It takes twice as much energy to get something to move if it's twice as heavy. Thin things vibrate easier then thick things, rectangles vibrate easier then triangles. Cantilevers vibrate very easily. Cast iron is not remarkably more damp then steel. Cast Iron is used in machinery because it's easy to cast and machine. Its extra dampness is a bonus. Being cast means the walls are thick or they are solid parts so they don't locally vibrate much and a machine needs mass to be stable especially if it moves fast so it doesn't dance. Sand as you explained correctly works the same way as the internal microstructure of cast iron and steel. The metal grains move and work against each other via friction and this absorbs movement energy slowing the vibration. Sand does the same thing. The damping tape consists of a mastic material with aluminium film on top. The Al constrains the top surface so as the bottom moves it shears the mastic. Internal friction then damps the motion.

https://en.wikipedia.org/wiki/Damping_capacity

https://en.m.wikipedia.org/wiki/Damping_ratio

Its tough to find a zeta chart with steel, CI and other materials that we use... I'll keep looking. Peter

[wmgeorge]

It was my understanding that all Steppers are prone to resonate at some frequency and adding or adjusting mass is just one way to move it. There are adjustments to make in the settings but as I pointed out before Gary is using the Exact same control system motors, drives and even down to the settings as his old one. He has changed the design so now its moved to a frequency that is noticeable. So he can either just ignore it or starting making changes so it moves to a less bothersome frequency. Adding mass via sand (very impractical now that its built) or weight by some other means. Gary I know not everyone has the same stockpile of junk (materials) use in experimenting but I just tossed it out there using some steel rod to try. Rebar is not expensive, its a bugger to cut but for just some trial and error? Remember in brainstorming there is no right or wrong

[GME]

Putting steel rods in is a bad idea.

Adding weight (mass) is not the answer!

You need to add stiffness and/or damping.

Old heavy cast iron machines are good not because they weigh a lot. They are good because they are stiff (big sections) and have good damping from cast iron.

It is a common misunderstanding that adding mass cures resonance.

As Peter has pointed out the damping tapes work because they have a material that damps vibrations.

Sand is often recommended here. The movement of the grains against each other helps to dissipate the vibration (damping). I haven't seen the effect quantified in my reading.
(Keen to see a source if anyone has one)

(It's damping not dampening).

I too have looked for scholarly articles on use of sand. Most all that I found had to do with soils, and mostly I could only read abstracts. They charge for the full article/treatise.

Epoxy granite seems popular when it comes to machinery. It reportedly has better damping properties than cast iron. Here's a link to a study: https://www.hyuckchung.com/uploads/2...grant_2014.pdf,

Epoxy granite appear to have supplanted cast iron in some cases, and been used to augment it others. Apparently, some manufactures of heavy equipment are using it, and even have propitiatory formulations.

I won't argue against damping tape. Before Peter mentioned it, I didn't know it existed. I am still reeling at the price, though. I haven't yet figured out how it should be applied to an extrusion so that it doesn't interfere with moving parts. I've seen some damping sheets used in automobiles and aircraft. Some are similar to the 3m product in appearance and advertise vibration damping, but appear used mostly for sound reduction.

Gary

[peteeng]

Hi gary - at 5ipm what is the rpm of the motor? Peter

[GME]

It was my understanding that all Steppers are prone to resonate at some frequency and adding or adjusting mass is just one way to move it. There are adjustments to make in the settings but as I pointed out before Gary is using the Exact same control system motors, drives and even down to the settings as his old one. He has changed the design so now its moved to a frequency that is noticeable. So he can either just ignore it or starting making changes so it moves to a less bothersome frequency. Adding mass via sand (very impractical now that its built) or weight by some other means. Gary I know not everyone has the same stockpile of junk (materials) use in experimenting but I just tossed it out there using some steel rod to try. Rebar is not expensive, its a bugger to cut but for just some trial and error? Remember in brainstorming there is no right or wrong

Yes, rebar is cheap enough. I'm way out of my depth. Would a load of rebar add mass and dampen, or would it just rattle around? My gut reaction, which isn't worth a damn in this context, tells me that unrestrained rebar would just rattle around. Too little contact area with the extrusion inner surface to provide much, if any, damping. I haven't crossed it of my list of possibles experiments, though.

Thank your for your willingness to help, Bill.

Gary

[wmgeorge]

To try it out, just wrap some rags or tape around the rebar.... simple. If it does what you want come up with a permanent fix.

[GME]

Hi gary - at 5ipm what is the rpm of the motor? Peter

Hello, Peter.

5.092806 rpm.

Gary

[peteeng]

Hi Gary and others with this sort of gantry - I have downloaded the 80x160mm section and created a model 2m long. How long is the Gantry Gary? I have run a modal analysis which determines the vibration modes of the gantry by itself. Structures have global and local vibrations. Global is when the entire element or part moves, local is when a small part of a part moves. Here they are with comments:

122htz global banana thru thin direction
218htz global up and down
261htz local at hole end
327htz second order banana back and fwds
362htz torsion global
428 local at hole in end
430 local at hole in end
446htz local at hole in end
450 local at hole in end

Gary if the holes in the ends are threaded screw a bolt into them to stop the small local vibs. There are literally hundreds of them associated with these holes. They are what's called a "free edge" and they are prone to vibration. Now I know the rpm vs speed I'll try to figure the possible driving frequencies. Peter your motors are 200step/rev?

We are not in mid-band resonance territory Gecko say this is 60-120rpm area. So the motor is doing 5.09rpm. This means that the mechanical steps are tapping the structure at 5.09*200/60= 16.97htz and the usteps are tapping at 5.09*2000/60= 170htz so it would seem we are in the territory. (EDIT corrected the maths)

if you disengage the rack and run at the 5ipm and the structure vibrates then the vibration path is thru the pivot. This would be a good clue. Can you check the 5.09rpm please...

https://www.geckodrive.com/support/s...stability.html

One thing I'll investigate was there were a few "rigid body motions" This means the whole part is shaking vs vibrating. A small technical difference. In this case have to neglect these as the gantry is connected to the structure so shaking depends on the entire structure not just the part as I have modelled. But rigid body motion (or shaking) may have something to do with your machine.

Gary what is distance between your columns and the full length of the gantry? Do you have a CAD model of the gantry and mounts? STEP export?

[peteeng]

Hi All - If someone here has a spectrum analyser could they analyse Gary's video and tell us what his vibration freq is? Ta Peter


Spek – Free Acoustic Spectrum Analyzer / Spectrogram Viewer

Gary you could try this one. Or anyone interested out there.... will be a couple of days before I can...



https://www.youtube.com/watch?v=XkmgMkDKAyU here's a great video using sand. But as our vib is back and forth you can't stick sand on a vertical surface!! But by sliding your screwdriver along the gantry you may be able to find the nodes. Then once we know the node spacing we can identify the frequency. Next is to figure some damping strategies.

[GME]

Hi Gary and others with this sort of gantry - I have downloaded the 80x160mm section and created a model 2m long. How long is the Gantry Gary? I have run a modal analysis which determines the vibration modes of the gantry by itself. Structures have global and local vibrations. Global is when the entire element or part moves, local is when a small part of a part moves. Here they are with comments:

122htz global banana thru thin direction
218htz global up and down
261htz local at hole end
327htz second order banana back and fwds
362htz torsion global
428 local at hole in end
430 local at hole in end
446htz local at hole in end
450 local at hole in end

Gary if the holes in the ends are threaded screw a bolt into them to stop the small local vibs. There are literally hundreds of them associated with these holes. They are what's called a "free edge" and they are prone to vibration. Now I know the rpm vs speed I'll try to figure the possible driving frequencies. Peter your motors are 200step/rev?

We are not in mid-band resonance territory Gecko say this is 60-120rpm area. So the motor is doing 5.09rpm. This means that the mechanical steps are tapping the structure at 5.09*200/60= 16.97htz and the usteps are tapping at 5.09*2000/60= 170htz so it would seem we are in the territory. (EDIT corrected the maths)

if you disengage the rack and run at the 5ipm and the structure vibrates then the vibration path is thru the pivot. This would be a good clue. Can you check the 5.09rpm please...

https://www.geckodrive.com/support/s...stability.html

One thing I'll investigate was there were a few "rigid body motions" This means the whole part is shaking vs vibrating. A small technical difference. In this case have to neglect these as the gantry is connected to the structure so shaking depends on the entire structure not just the part as I have modelled. But rigid body motion (or shaking) may have something to do with your machine.

Gary what is distance between your columns and the full length of the gantry? Do you have a CAD model of the gantry and mounts? STEP export?

Thank you for all your work, Peter. However, I have to admit that it's all Greek to me. What follows is the information you are looking for.

The gantry is 1879.6 in total length.

The motors are 200 steps/rev.

The drive is 3.2:1 gearing

The rack has 20 pitch, 20 degree pressure angle.

The structure vibrates with the R & P drive disengaged from the rack at 5.09 rpm.

Distance between the gantry mounts is 65" (1651mm). That's the distance between the interface plates. I measured that distance, because there are bolts that pass through the outside edges of the interface plate and into the bottom of the extrusion. So, the restraining elements are not limited to the bracing attached to the tops of the interface place.

I don't have a CAD Model. I modeled the bracing/mounts and the interface plates, but didn't add the gantry beam. From what I had, I knew it would fit.

The holes in the ends of the extrusion are threaded. I needed 16, and as luck would have it, I had exactly 16 8mm bolts. I then ran at 5 ipm. Astonishing! (To me anyway). The vibration diminished significantly. There is still some, but not nearly as much. So far, every change has brought about improvement.

Unfortunately, I'm not sure whether to attribute the lower vibration to the bolts. This evening, I installed a portion of my 18mm Baltic Birch spoilboard base layer. I didn't have it completely bolted down, but had 7 our of 16 bolts in. The piece of BB is 30" x 60" (762mm x 1524mm). I'm using 2 pieces rather than a full sheet. I happened to have 2 half sheets on hand. I'm not sure it matters, but I am putting in 4 rows with 4 bolts to a row. I will be adding a 5th row, but I have to cut the holes by hand. I did the pocketing and holes on the CNC in so far as the gantry would reach. I didn't want to bother with tiling the toolpath. There are 4 bolts along the centerline and 4 bolts along the front edge. I thought this might be important, in case the BB has a damping effect. Anyway, I can't assess whether the decline in vibration is attributable to the bolts in the ends of the gantry extrusion, the BB ply or a combination of the two.

FWIW, I also have two 60" (1524mm) lengths of extrusion that is part of the frame with threaded holes at both ends. They are 3" x 3" (3030 in 80/20 parlance) profiles. I planned to screw on covers later. They take 5/16" (about 8mm) bolts, so I'll added bolts to them as well and see what, if any difference they make. It seems like there was a slight marginal improvement over what I observed after treading bolts in the gantry ends.

It's funny, but the noise from the steppers running at 5 ipm tends to belie what's going on with vibration. They sound like there should be significant vibration, but "feeling" for vibration reveals that it has become quite minor. Might that tend to distort and audio analysis? The steppers are growling away, but the frame and gantry aren't actually resonating much at all.

I have to admit that I am still quite amazed by the difference the screws made. I'm guessing that most folks on the Zone have no idea how such a small thing could make such a large difference. The whole rig is getting very close to being vibration free. It will be interesting to see what difference finishing the spoilboard and adding the gussets will make.

One other thing. The plate my Z axis is attached to also has a number of open threaded holes. At total of 16 are 8mm and 8 are 6mm. The mounting plate is about 15mm thick. The plate comes from Avid CNC/CNC Router Parts. Unfortunately Avid doesn't have a photo of it online. I did find a drawing of it in the Avid assembly instructions. Here is a link: Gantry Assembly - PRO Series CNC Machine Kit Assembly Instructions Unfortunately, there are other holes that I cannot reach without removing the Z axis assembly. Do you recommend threading bolts into the visible threaded holes?

Wow! Things are certainly moving along nicely. Without doubt, I would be in a very different place without your expert advice and assistance.

Thank you so much.

Gary

[peteeng]

Hi Gary - Excellent news. I believe the gantry is vibrating in its first mode and there are several harmonics going on especially with the extrusion ends. The hole ends are vibrating at twice the primary frequency ie primary freq is 122htz for a 2m beam. I shall correct this to your supplied length. 244htz is twice the primary and there are many hole vibrations around this point. Car companies spend a small fortune getting noise and vibration out of their cars. The columns can be considered as clamps so the apparent length of the beam will be somewhere between 1880mm and 1651mm. If we could analyse the frequency we could then back calculate the apparent length. I decided to check the FE by hand so looked up the vibration eqn. So I've attached it here for anyone who can deal with it. The hand calc says 123htz and the FE says 122htz so all good.

Changing my spreadsheet to 1700mm long gives 170htz (eureka its the motor freq!!) and 1880mm gives 139htz.. The FE local results (end holes) will be length independent. So onto a damping strategy. One way is to make the beam asymmetric. So it can flex one way but not the other. A common approach is to make a "stand" in the middle of the beam and put a wire across it and preload slightly. So its stiffer on one direction vs the other. In your case you could bolt on an extrusion on edge at the back to make it more out of symmetry. But we'll digest where we are and see how things travel. Peter

Re: holes in construction extrusions. From the FE work I'd say if you can, do it. Seems they are the source of lots of bad vibes. Wouldn't have guessed that, s'pose that's why we do FE. Peter

Also - the vib is transmitting thru the pivot so if there is an opportunity to use a rubber fixture, plastic washers or anything that isolates metal to metal at the pivot you will drop the transmission rate further. I suspected that the vib would not be transmitted via the belt as its rubber.

I don't think Gecko can do much about this. If you had a driver that you could change the Usteps you could use 4000 steps and move the driving freq up. cheers - persistence pays

Also a piece of 1mm plastic or insertion rubber between the motor and the arm will help. Plastic is viscoelastic so transmits less energy. Peter

[peteeng]

Hi Gary - I think you suffered some extreme bad luck. You picked the exact clamp length and beam inertia that agrees with your stepper pulse at a speed that is noticed. But now we understand the physics, solutions will be forthcoming. Peter

What bearing is in here?

[padprinting]

Laser engraving is a subset of laser marking. In this process, lasers are used to engrave an object. The laser engraving process leaves marks on different types of objects which include the change in color because of molecular/chemical alteration, melting, charring, ablation and foaming.

[wmgeorge]

Deleted

[GME]

Hi Gary - Excellent news. I believe the gantry is vibrating in its first mode and there are several harmonics going on especially with the extrusion ends. The hole ends are vibrating at twice the primary frequency ie primary freq is 122htz for a 2m beam. I shall correct this to your supplied length. 244htz is twice the primary and there are many hole vibrations around this point. Car companies spend a small fortune getting noise and vibration out of their cars. The columns can be considered as clamps so the apparent length of the beam will be somewhere between 1880mm and 1651mm. If we could analyse the frequency we could then back calculate the apparent length. I decided to check the FE by hand so looked up the vibration eqn. So I've attached it here for anyone who can deal with it. The hand calc says 123htz and the FE says 122htz so all good.

Changing my spreadsheet to 1700mm long gives 170htz (eureka its the motor freq!!) and 1880mm gives 139htz.. The FE local results (end holes) will be length independent. So onto a damping strategy. One way is to make the beam asymmetric. So it can flex one way but not the other. A common approach is to make a "stand" in the middle of the beam and put a wire across it and preload slightly. So its stiffer on one direction vs the other. In your case you could bolt on an extrusion on edge at the back to make it more out of symmetry. But we'll digest where we are and see how things travel. Peter

Re: holes in construction extrusions. From the FE work I'd say if you can, do it. Seems they are the source of lots of bad vibes. Wouldn't have guessed that, s'pose that's why we do FE. Peter

Also - the vib is transmitting thru the pivot so if there is an opportunity to use a rubber fixture, plastic washers or anything that isolates metal to metal at the pivot you will drop the transmission rate further. I suspected that the vib would not be transmitted via the belt as its rubber.

I don't think Gecko can do much about this. If you had a driver that you could change the Usteps you could use 4000 steps and move the driving freq up. cheers - persistence pays

Also a piece of 1mm plastic or insertion rubber between the motor and the arm will help. Plastic is viscoelastic so transmits less energy. Peter

Hi Gary - I think you suffered some extreme bad luck. You picked the exact clamp length and beam inertia that agrees with your stepper pulse at a speed that is noticed. But now we understand the physics, solutions will be forthcoming. Peter

What bearing is in here?

Good morning, Peter. I don't know what bearing is in there. The R & P drive came with the pivot shaft already inserted. I didn't want to pull it an then potentially have difficulty reinserting it. Note: while my R&P drive is much like the one Avid sells, it is not an exact copy. It came with the kit I purchased from Fineline Automation and I reused it. The Fineline model is based upon Avid's design, probably because Fineline used to purchase its drives from Avid.

One thing that may be noteworthy. When I designed the plate to which the R & P mounts, I made it with too close a tolerance. When I installed the drive, I found that the end of the pinon gear just barely rubbed against the frame. Fineline machines were notorious for not have the best pinion engagement in the racks, so I wanted to correct the problem in my current design. Unfortunately, I slightly overcorrected by a few thousands. To adjust for the error, I placed a shim washer between the drive and the mounting plate. The washer moved the pinion far enough out to give me sufficient clearance. I mention this, in case the added washer might have an effect on the vibrations. I'm guessing it does, since you mentioned eliminating/reducing metal-to-metal contact. I'm thinking that replacing the steel shim washer with a nylon washer might help. It can't hurt.

I agree with you about Gecko. This does not appear to be stepper drive problem. Your calculations, coupled with the changes we've successfully made point to something other than a fixable stepper driver issue. Gecko does made a drive that allows a variety of microstep options. I've eyed it, but since I may go to servos at some point, couldn't justify buying it. As much to the point, with the gearing I have, the resolution I get has generally satisfactory for my purposes.

I'm off to shop for nylon washers. When I pull the R&P drives to swap out the washers, I'll take a look at the pivot shaft bearing.
:cheers:
Gary

[GME]

Hi All - If someone here has a spectrum analyser could they analyse Gary's video and tell us what his vibration freq is? Ta Peter


Spek – Free Acoustic Spectrum Analyzer / Spectrogram Viewer

Gary you could try this one. Or anyone interested out there.... will be a couple of days before I can...



https://www.youtube.com/watch?v=XkmgMkDKAyU here's a great video using sand. But as our vib is back and forth you can't stick sand on a vertical surface!! But by sliding your screwdriver along the gantry you may be able to find the nodes. Then once we know the node spacing we can identify the frequency. Next is to figure some damping strategies.

Hello, Peter. I agree; the video is great. I had been having difficulty understanding the node thing. The video explained it well (and without any dialog).

I downloaded the analyzer and ran the audio from the left and right stepper videos through it. I don't know how to interpret the results, but here they are:



Left side



Right side.

To my untrained eye, they look remarkably similar. I suspect that is to be expected.

I will finish installing my spoilboard base layer tomorrow and will rerun at 5 ipm to see what difference, if any, the BB plywood makes to the resonance. I am also picking up nylon washers, and bolts for the Z axis plates, in the morning. I will isolate the motors and pivot shafts with the washers to add a layer of isolation, as you suggested. I'm going to testing incrementally. I will run at 5 ipm with the only new change being the spoilboard base layer. Then, I will add the washer isolation and run again. Then I will add the bolts to the Z axis mount and test again. I still have to cut the aluminum gussets to see if adding additional bracing helps, but that's for later. I need to finish the spoilboard, which will require a good deal of cutting and gluing. It will take at least a few days.

I have to alternate work on outside projects, like revising parts of my sprinkler system, with working on the CNC. I wish I could just get the CNC done in one shot, but it's not to be. I thought when I retired a few years ago, I'd have more time in the shop. It hasn't worked out that way. Outdoor projects in the summer and indoor in the winter. Snow in the winter is not much of an issue here, but we get a lot of rain coming off the Pacific. Not as much rain as you get, but more than I like.

Gary

[peteeng]

Hi Gary - I interpret these as we have frequencies everywhere. I was hoping to get a spike at 170htz and 340htz but the image covers nearly everything. I'll read the manual for spek and see if I can interpret. But I think it's moot. We know the source is the motor and everything that can be done to isolate the structure from the motor or damp the structural elements is to be done. Peter

[peteeng]

Re: The spectrograph. The RHS is the dB of the noise. The colour defines the dB level. dB is air pressure which correlates to volume. The LHS is frequency from 0 at the bottom to 22kHz at the top. The bottom is time so 0-0.4 maybe 0.4min? So to interpret, you have heaps of very load high freq noise. Not sure where that comes from. There is a 10kHz band at 40-50dB (yellow) but down where we want it we don't have much definition. This could be because the microphone can't register <200htz. Look at your mic spec? is it a phone camera? Look at the SPEK and see if you can filter the bandwidth. You maybe able to tell it to only listen to sub 1Khtz??

I looked up phone mics and they generally are 20htz to 20khz. But some have 2 or 3 mics and run noise cancelling software which maybe would get rid of loud low noises?? I don't think we have to dig into this much unless your interested. We know the motors fundamental vibration is 170htz....Peter

5.09*2000/60= 170htz rpm*usteps/min/sec = usteps/sec = htz fundamental which agrees with the first vib mode of the gantry section at its clamped spacing
1st harmonic 2*170=340htz
2nd harmonic = 510htz