PDA

View Full Version : New Machine...cantilevered design project log



corydoras
05-18-2007, 06:33 PM
I just started building my 2nd machine, and it is to be a servo powered cantilevered design.

The servos are Sanyo Denki BL Super on the X and Y axis, and stepper for the Z axis, due to weight reduction.

Just finished the framework and the gantry, but still need to make the z akis and the belt transmision for the Y axis.

The X axis ia a Parker/Hauser linear drive HPLA 80 with a 4:1 reduction on a planitary gear.

The Y axis are 2 x 25mm THK linear guides and are to be belt driven.

Working area for the mill are gonna be 1850 mm x 1200 mm and approx 70 mm on the z axis

Here are some pictures from the build, and theres gonna be more.

Please feel free to comment the construction and keep em good ideas coming

Glidergider
05-19-2007, 06:55 AM
Wow, that's a cool design. I hope you have an overhead crane installed above it. It must weigh a bunch. The cantilever design will be very conveniently and provide lots of access to the cutting materials. Is this really a wood cutting machine, or do you plan another type of cutting head planned? The design makes me think of a plasma cutter. Its going to be a nice machine, no matter what kind of head you have installed.

corydoras
05-19-2007, 09:10 AM
Its gonna be a wood maschine, and the machine is made into severel parts that are boltet to make movable and to make it fit trough my basement doors.

Its not as heavy as it looks, but plenty rigid fo my kinda work.

The gabtry weighs aprox 120 lbs but I can move it on the linear guides with my little finger.

Yes the cantilevered design makes it very accessible.

Thanx for the comments

Cory

walter
05-20-2007, 11:15 AM
Hey that looks cool!

Glidergider
05-20-2007, 11:19 AM
Cory,
Just curious, what kind of work, professional wood carving or is this a hobby machine?

corydoras
05-20-2007, 04:45 PM
Just a hobby machine, needed something bigger and better than the first one I made.

http://www.cnczone.com/forums/showthread.php?t=23391&highlight=angel+2000


Painted tha parts and are going to start making the z axis now, as I wait for my timing belts and pulleys to arrive

Cory

cabinetcraft
05-20-2007, 05:49 PM
Good luck on the cantilever design. Make it a stiff as possible and then double that times 2. We have a large cantilever router at work from Europe and it does OK until you get out towards the end of the Y axis and then the quality of the cut starts to go down. I'm used to 3/4 red oak at 500 ipm+ but not on this one armed bandit as once your past center on the Y axis there is too much flexing and everything goes to H**l. Make it very rigid and then some. Looks really good though, what you going to use for a spindle?

Glidergider
05-20-2007, 06:12 PM
There are thousands of guys who would love to have your first cnc machine. Very nice. This second one is going to be even bettter. I have the JGRO for about 2 months now, and I already want a bigger cnc.

I see you are a RC airplane guy. Do you cut molds with your cnc? Someday, I plan to do that, but of course I'll have to expand the cutting table first.

corydoras
05-21-2007, 03:47 PM
Did a test run on my x axis today (the parker/hauser linear drive)
Jogging back and forth with me STANDING on top of the carriage......about 180 lbs....and no problem at all.

I have a 4:1 planetary gear on it and i jogged at about 12000 mm/min

Hopefully this will be enough to convince me that my 120 lbs gantry will runn smoothely on rapid moves.

Will post more pictures as the build progresses

Cory

corydoras
05-30-2007, 02:21 PM
Well here are some more pics from the build.

Got the x and the y axis in place and have done some test runs aon 65000 mm/min....man that is fast...and will of course not be able to cut at that speed, but the rapids are fun.

Video coming later today.

Not completely happy with the gantry motion some small vibration and backlash....hmm I am using AT-10 timing belts and 4:1 planitary gear..

Cory

Glidergider
05-30-2007, 03:22 PM
Looks like you could build seats on top of that gantry, charge $5 and give a pretty exciting ride on that thing. I'd pay. Wow, I'm impressed with the speeds.

corydoras
05-30-2007, 04:34 PM
Thanx Glider..

Mabye I´ll use it for that if it doesn´t get accurate enough.

Maby I´ll have to rebuild the gantry....its awfully heavy an maby not as rigid as I hoped for.

Here is a video of some movements at 50000 mm7m

http://youtube.com/watch?v=3MoU7KRB1AA

enjoy, and please comment

Cory

Glidergider
05-31-2007, 08:01 AM
Wow, the speeds. :) Hope the flex is within your limits. I saw another guy build a cantilever here in the zone. To solve his flex issues, he spread the rail base out. This effectively reduced his X travel.

corydoras
06-18-2007, 02:16 PM
Not much progress on the build, due to very nice weather:)

Started building the electrical cabinet,and doing some work on the z-axis.

One question I keep asking myself.....if this so called "one armed bandit" doesent get rigid enough.....does anyone have any idea of how to reinforce it or drawings for a new gantry?

I have have about 20 meters of aluminium sized 80x40 mm about 3mm thick that i can use for a new design.

Would it be possible to make another linear rail on the other side and still have the timing pulley on one side?....or will it yaw?

Keep em good ideas coming ...pleeeease

Best reg
Cory

zippy
06-18-2007, 04:59 PM
Where's the lack rigidity? Is it the bearing interface or resonance in the square tube? It looks like you have way more experience then me and better tools etc... but I saw a reference in the polymer concrete monster thread and then found and read a PhD thesis about controlling the vibration in steel tubing he tested lots of methods and published the data so if its vibration in that long arm I think that could help you. It looks like something that long would ring like a bell. I'll see if I can hunt up the link.

zippy
06-18-2007, 05:22 PM
I went back it looks like the beam is an extrusion so I don't know if this will help but its a good read.

http://www.mech.utah.edu/~bamberg/research/PrinciplesOfRapidMachineDesign/Principles%20of%20Rapid%20Machine%20Design.pdf

corydoras
07-12-2007, 09:06 AM
Thanx for the links Zippy, but the Cantilevered design is is stopped!!!!!!!!!!!!!

I figured out after many nights of no sleep, and to many hours constructing that I was never gonna be satisfied with this design.

I ran some circles, and the accuracy of start and stop pos where fine, but the circels cut edge was a little bit "wave shaped", and that was on the inner pos of the beam....dear not think of how it would be on the end...

This was only making a 2 mm cut in plywood

I started rebuiding the machine yesterday, and are moving to the traditional gantry/portal design.

The main frame is almost ready, and the gantry is due for tonight, but some problems occured,....the cabinet for the lectronics was designed for fitting under the machine, but with the moving gantry and the servos and pulleys underneath there is just not enough room.

I think Ill put the cabinet under the workbench in the shop.

The cutting area was a little rediced due to reconstruction, and are now1800x1150x250 mm, but I think it will have to be enogh fo now, but i still have 2 pc of THK rail lying aroud so if i ever want to enlargen it, it coud be 3600mm for the x axis, but no room in my shop for that.

Ill post some photos of it soon, if anyone are interested in following the build

Cory

Glidergider
07-12-2007, 10:21 AM
Cory,
Good luck with the rebuild, but I know you won't be needing luck with your kind of skills. I'm envious of your large table and super fast speeds. Once you get the gantry redesigned, it will be an awesome machine.

corydoras
07-12-2007, 10:47 AM
Thanx Dave,

Hopefully, it will be as accurate as I want it, I am also thinking of making a low vacum bed for my 4" balsa sheets when starting to cut some wing ribs, and one for my depron sheets to make some indoor models.

As you may have understood, my other passion except cnc is rc model airplanes,....and correct me if I´m wrong......your nick sommhow tells me that this takes up some of your spare time as well

Cory

Glidergider
07-12-2007, 12:29 PM
Yup, I do RC gliders. Although I will admit, this CNC gig has put a serious dent in my flying time. I'm the president of the North Alabama Silent Fliers. We are a a small AMA club that is devoted to RC sailplanes.

I'm real interested in your vacuum bed design. I will be needing that too. About a week ago I cut my first set of balsa ribs using a .032 Down-cut bit. Here's a link to my pictures.
http://www.cnczone.com/forums/showpost.php?p=317357&postcount=196
http://www.cnczone.com/forums/showpost.php?p=317485&postcount=198

Can you discribe in general terms how you plan to build the vacuum bed?

harryn
07-12-2007, 03:18 PM
I am still just a beginner, but AT 10 belts are not all the same. There are some that are "standard grade" and others that are for precision work / no backlash. The same goes for the pulleys.

Ger21 pointed me toward these guys in the past

http://www.brecoflex.com/

Their web site is an amazing resource. Look under the section for linear motion.

corydoras
07-13-2007, 06:08 AM
Here are some snapshot photos taken with my phone cam, and the frame and the gantry is now almost ready.

Need to make belt holders in each end, mount the cable tracks and and put some paint on Her.

Then I´m moving here home for final assembly and electronics

Maby sometime next week I´ll have Her moving around.

(I think maby thats the best since Ivé spent alot of days in of my summe vacation on this, and the family needs some time too)


Ill keep you updated

Cory

Glidergider
07-13-2007, 08:13 AM
That's an great looking machine. You flexibility issues should be gone forever.

Glidergider
07-15-2007, 03:52 PM
Thanx Dave,
I am also thinking of making a low vacum bed for my 4" balsa sheets when starting to cut some wing ribs, and one for my depron sheets to make some indoor models.
Cory

Cory,
You inspired me to build a vacuum bed. I'll post pictures later tonight. Its 48 inches long, 10 wide. It'll handle an 8 x 48 sheet of balsa or foam. I haven't cut anything on it yet so I don't know if I'll have to add clamps yet. The vacuum is great for sucking the balsa down flat. I just don't know yet if it will prevent sliding. I might have to add a little clamp too.
dave

corydoras
07-17-2007, 04:47 PM
Hows the vacum bed coming along Dave?

I have now reassebled teh machine in my shop and ran the same circle cut that i did with the cantilvered design, and there are still some wave shapes on the cutting edge.

As you can see from the picture, it always appears on the same spot.

The picture show one deep and one shallow cut, showing the error in the same spot

I have a little backlas in the Y axis, but from the photos attatched, would you say it is from this or the X axis?

Could it be one of my linear blocks that are not good on my thk rails?
The feel pretty good, and I have dismantle them and, cleaned them and lubricated them all over


BTW the cutting was done at 5000mm/m


Best reg
Cory

Glidergider
07-17-2007, 05:52 PM
Cory,
Is that a vee bit? How much deflection are we talking about?

From the evidence I see in the photo, I'm guessing its your Gantry flexibility is the issue. The spinning cutter is pullling your router bit in the plus X direction. The gantry deflects in torsion allowing the bit to move in the X due to cutting bit forces. That's just my guess.

In your photo label the direction of travel and rotation of the bit. Studing those aspects might help you determine the forces.

Glidergider
07-17-2007, 05:59 PM
Cory,
The Vacuum Bed is still untested. Tonight will be it's first test. Hopefully. If I can drag myself away from watching the baseball game on TV. So we will see.
Here's a link to a photo.
http://www.cnczone.com/forums/showpost.php?p=320522&postcount=216

corydoras
07-18-2007, 02:15 PM
I think I may have found the problem, I am almost suere that it comes from bad carriers on the linear rail.

Whwn I try to move them across the rail i feel some small resistance, like I can feel everi ball inside the carriage flip throu the turn going to the back side if anyone understands what i mean?

So I have to put some extra force into the push of the carriage, and the it suddenly moves in a "bump" until the next ball is about to turn.

These rails were second hand and free from a replaced machine, but has travels just a small deistance due to machine operation.(mostly stand by)

Maby thas just as bad, if it has some vibration aon the machine?

The problem I have is if i am to buy 4 new carriages, and then to find out that this is not the problem, then I have a problem as these carriges are about 350USD a piece here in Norway.

I have not been able to find out what brand they are,´they are not marked anything only H30, and Ive been sweeping the internet to try to find someone that looks like this (see attached photo)

Size? aprox 115 mm long and 90 mm wide, and the guide is close to 30 mm (maby that why they are called H 30?)

Ive looked at THK, INA, SKF, STAR/Bosch Rexroth, but no luck so far

Anyone got any other ideas of how to make this machinework ok, without these rails?, skateboard bearings?

I got about 100 of the available for free.

Are machines made with these bearings accurate?

Please have some answers as I am dying to try out my new machine:)

Best reg
Cory

Glidergider
07-18-2007, 02:59 PM
Cory,
Vacuum bed worked great. I cut some balsa last night.

Regarding your problem. I don't believe you found the problem. Your bearing blocks cause resistance only, correct. Do you feel any radial looseness? If not, I doubt the bearing blocks are the problem.

If you grab your spindle with your hand (unplug first). Push and pull in the X direction. Put a dial indicator on the spindle. Measure the deflection. I'm guessing your will see a mm in deflection. Is that possible?

corydoras
07-18-2007, 03:01 PM
Looking around in the garage i found some 20 mm predrilled hardened steel rods with matching wheels.

Half of the weels have an offset center for making adjustments, Could this be something to build on.

Hopefully without any large rebuild, as I have already spent close to 250 hours on this machine, and dont want to disasemble it again and bring it back to work.


Any suggestions are greately appriciated.

Best reg
Cory

corydoras
07-18-2007, 03:48 PM
Great, is it possible for you to shoot some video of it?

What do u use for vacum?
Does the vacum loose its effect when you have cut some parts and made leakage?


Well its kinda hard to say, if i jog the gantry slowly it still shakes.

Same thing happen when i remove the motor and timing belt and push the gantry back and forth.

I put a glass of water on the gantry and it gets ripples on it when i move the gantry, well im not sure

best reg
Cory

Glidergider
07-19-2007, 09:59 AM
Cory
Don't hold your breath on video of the vac box. It would be the most boring video of all time. I use my Dust Collector (2hp Harbor Freight) (http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=45378) to power the vac box.

You never did tell us how much error you are talking about. Your pictures don't have scale, so I'm not sure if your problem is 1 mm or 0.1 mm.

The ripples your see with the glass of water can be a separate problem. If you are seeing 1 mm of error in your circles, I doubt if its the cause of the bearing roughness. Radial play in the bearing assembly would cause problems. Axial roughness shouldn't affect the cuts.

corydoras
07-19-2007, 11:21 AM
Good to hear that your vacum bed turned out great.

Its kinda hard to measure on the circle, but I would say close to 1 mm

I dissambled the carriers again and found dents in the carrier groove for the balls on 2 of the carriers. (see attached photo)

Very hard to see at the photo, cause I dont have macro function on my phone, but you get the picture where the dents are.

These dents are caused by the machine the it ran on was placed in the same pos for weeks, and it had some vibration

The problem is that I also mixed all of the balls when cleaning them so I dont know what balls to put in the 2 good ones.

Best reg
cory

Glidergider
07-19-2007, 11:33 AM
Cory,
It's not the carriers. Its the Gantry flexibility.

Do you own a dial indicator? (http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=93051). They are cheep.

Attach the vicegrip to the cutting bed and align the indicator to the spindle area. Then physically (with hands and arms) push in the X direction to measure the deflection. I'm guessing you will see a deflection, which is bad news, and likely the problem. Your gantry is pretty tall, and depending on the joints, I'm thinking that you are getting some deflections.

Repeat the process in the Y axis. Realign the dial indicator to measure Y deflections first.

I hope my meaning is coming through.

cncfoam
07-21-2007, 09:57 AM
I would look at backlash in your pulley system. You might be able to take out backlash with software.
EZ

corydoras
07-30-2007, 12:46 PM
Hi guys

I Have now lowered the gantry, and made a frame underneath, but sorry to say that i still have problems with deflection.

I went to my friends house and looked at his machine, and believe me he has several mm that he can flex his gantry, and still he cuts circels round as a ball at 5000 mm/m.

I can cut at 1000 mm/m and get a decent result, but not far as good as he gets at 5000.

I tried something else than circles and got the same problem when the y and x axis move at the same time (see attached photos)

I put the dial indicator on and found about +/- 3/10 of a mm (+/- 0.3mm)

I am going in the same direction as cnc foam on backlash (no offence Glider)....but dont know where to find it.,,,could it be the planitary gear?

I have tried to lock the motorcoupling with a spline and there is less movement, so mabye my servo motor isnt up to the holding torque?

I have set my servo amp to the highets gain possible and still running smooth with no vibration.

I have myposition loop gaim maxed out to keep it in pos.

i get almost like hickups at 12 and 6 o´clock on the end of the x axis move (when it changes direction)

This all points to backlash, and i´ve tried to put in backlash comp in mach 3 but tgis just messes up the whole cut.
When you look at the circle its almost like 2 of the sides (at 6 and 12 o´clock is a little "flat", that again points to backlash.

But howcome I then gett the strange wobbles when cutting the A letter?, the x axis motion is in one direction and should not siffer from backlash making a straightmixed cut (is this what they call interpolation?)

The first image is cut using a 90 degr v bit 2 mm cut at 5000 mm/m and the other one at 1000 mm/m

Well, I am running out of ideas soon, hope some of you guys may have something to light the fire again.

best reg
Cory

Glidergider
07-30-2007, 02:43 PM
Cory,
Your flexability looks good for a wood router. 0.3 mm is nothing. I wouldn't worry about flex any more.

Can I make another suggestion, and then I'll shutup. Use the dial indicator again on a X only run test. Drive in the +X direction a small amount to mesh the drive train.

Set the dial indicator to zero. Then drive in minus X a small amount (no greater then the dial indicator can record).

Stop and record the value. The dial readout should match the Mach3 command. This is a direct backlash measurement.

Repeat once more in the +X.

This will tell you what kind of backlash you have. Repeat in the Y axis.

Hope this gives you some ideas. This will let you confirm or elliminate the backlash as your problem.

Here's another idea for your the problem. Looking at your letter A, cut at 5000mm/m, could it be that your servo Position Feedback Loop is slow? That would explain the wavey lines. Could it be that you are loosing steps in the X but the Y is continuing, then the X catches up. I think you are already considering this possibility.

Good luck, hope this helps. Sometimes a wrong suggestion can trigger the correct solution.

Dave


Hi guys

I Have now lowered the gantry, and made a frame underneath, but sorry to say that i still have problems with deflection.
......
I put the dial indicator on and found about +/- 3/10 of a mm (+/- 0.3mm)

I am going in the same direction as cnc foam on backlash (no offence Glider)....but dont know where to find it.,,,could it be the planitary gear?

I have tried to lock the motorcoupling with a spline and there is less movement, so mabye my servo motor isnt up to the holding torque?

I have set my servo amp to the highets gain possible and still running smooth with no vibration.

I have myposition loop gaim maxed out to keep it in pos.

i get almost like hickups at 12 and 6 o´clock on the end of the x axis move (when it changes direction)

This all points to backlash, and i´ve tried to put in backlash comp in mach 3 but tgis just messes up the whole cut.
When you look at the circle its almost like 2 of the sides (at 6 and 12 o´clock is a little "flat", that again points to backlash.

But howcome I then gett the strange wobbles when cutting the A letter?, the x axis motion is in one direction and should not siffer from backlash making a straightmixed cut (is this what they call interpolation?)

The first image is cut using a 90 degr v bit 2 mm cut at 5000 mm/m and the other one at 1000 mm/m

Well, I am running out of ideas soon, hope some of you guys may have something to light the fire again.

best reg
Cory

harryn
07-30-2007, 04:20 PM
Since you are desperate for ideas, that allows me to make my uneducated observations - feel free to ignore.

I was looking at the pics of your trolley drive arrangement again, and just so I am clear, you have
- The drive motor
- An in-line grear reduction box
- The belt drive pulley attached to the gear box
- The belt itself - perhaps a kevlar re-inforced belt
- The belt tensioned on each end to ? lbs of force ?

I am particularly following your thread, as my own DIY router drive concept is reasonably similar. The good / bad is that each movement exactly matches the motor drive mechanism.

I was looking at the specs for the newst IMS stepper motors, etc, and noticed that the backlash on their standard gear reduction box was nearly 5 %. I am not sure exactly (5% of what), but it seemed worth looking at more before I bought one.

I am still just reading about stepper motors, but since each microstep can represent a significant distance on a trolley arrangement, it appears that I will need to run at fairly small steps (like 256 micro steps / motor step) to avoid visible steps. (if I skip the gearbox).

I am not at all clear on the real precision of each microstep vs the full rotation. The motors will clearly have build tolerances, so each and every microstep would seem to be slightly different than the others - no idea how much this is in reality.

If you are not using "way" oversize belts for the load, they will go through some slight stretching / retraction as you change directions. Steel re-inforced ones are better.

There are different tolerance belt teeth / pulley arrangements. Here is a link which discusses tooth based backlash.

http://www.brecoflex.com/?CATID=1&SCATID=2&SMENID=12&PUDE=72

Just curious, is there any relationship between the circle non-round areas and certain spots on the drive pulley or belts ?

Sorry for no real answer, just some more ideas to think about.

Good luck.

Harry

ger21
07-30-2007, 07:36 PM
Assuming you're using mach3, too slow of a PC could cause that. How fast is the PC you're using?

corydoras
07-31-2007, 12:43 AM
I´m running a Pentium4 2,5Ghz with 1Gb of Ram, so most likely this should not be the problem

Harry: no specific place on the belt, have tried both ends, the middle and same on the y axis

Best reg
Cory

Glidergider
07-31-2007, 08:15 AM
Cory,
After more thought, your problem is not backlash. Backlash would not cause the waves in the letter A at different speeds. Its not machine flexibility either as you've eliminated that possibility. Its not your bearing carriers either.

My money is on the position feedback loop. So how can you test that? What acceleration values do you have set on your X and Y servos?

corydoras
07-31-2007, 09:31 AM
Hi Glider

Ive tried all kind of acc settings from 150 mm/m to 1000, it hets a little better at 150 mm/m, but not much

I am going to try and put a large stepper motor on the x axis tonight and see what happens


Wish me luck that i then can find the source of this displacement

Best reg
Cory

Glidergider
07-31-2007, 09:36 AM
What are you running right now, servos with a feed back loop? That's what I read in your earlier posts. You are going to switch to a stepper on the x axis? Will the stepper have a feedback loop?

harryn
07-31-2007, 12:32 PM
If it is not backlash, then perhaps sort of a "dithering". Perhaps the stepper motors are not quite holding still in a micro step position ?

corydoras
07-31-2007, 02:57 PM
First of all, thank you guys for showing interest and bringing up ideas.

I´ll try to answer some of the q posted here:

Harryn: i think after a lot of testt, that flexibilty is important but not as much as i thought.

Backlash is also important, but in my case was not the issue.

Gilderglider: PLZ do not shut up, keep your posts and ideas coming,...after all this is what this forum is all about:)

I am running ac servos , but tried a friends stepper with GOOD result. the circles bacame very nice.

But at more rapid speeds this also had some problems holding in, but not as the servo in the6 and 12 o´clock position., more overall, loosing steps.

I run my servos without any encoder feedback to the computer, only resolver feedback to the drive.

I have encoder output on my drive, but doesn´t it take some kind of counting card in the oc, or is it possible to get into mach3 via paralellport?

The stepper that was tested had no feedback loop either.

So now i have some new Q´s.

Mabye i should get a steppermotor for my Y axis?

Are there any good stepper, driver boards out there for a reasonable price?

I already have a drive/BOB that i use for my Z axis, but it only takes 2,5Amps pr channel, and for a large stepper this is not enough.

My friend stepper that i tried tonight was a hybrid stepper connected i series, so it would not fry my driver card, so it probably did not have the holding torque specified....?

This was a 650 Ncm motor:
http://www.arceurotrade.co.uk/en-gb/dept_144.html

so I should at least shoot for a stepperlarger than this or what?

Any suggestions.

I should also mention that Postage to Norway is extremely expencive, and the customs are crazy over here.

I am visiting the US in sept, so mabye this is a good time to pick up some stuf(althogh i DO NOT think I can wait that long)

And by the way price matters, but I want a fully functional machine capable of running at least 7-10000 mm/m cutting, and faster rapids....at least this is my aim.

Please post some links to stepper7driver supplier on the web that you are happy with, and I´ll be greatfull.

Thanks for all your effort and ideas, keep up your hobby cnc-ing

best reg cory

Glidergider
07-31-2007, 04:53 PM
Cory,
Your desire to cut at 7-10000 mm/m is so far beyond my understanding. I like it. :) I would highly doubt if the 650 Ncm motor would get you there. My calculations say that's equal to 923 oz-in of torque. I'm running Keling's 425 oz-in and I can't go anywhere near that speed. My voltage is just 24 volts though.
Again, I'm out of my league talking that speed.

Madclicker
07-31-2007, 05:15 PM
I think you need to tune your servo drives. Tuning the PID loop is the key to optimum performance and accuracy.

paulC
07-31-2007, 07:29 PM
Could we see a picture of your z axis.
If you were to change the depth of cut when tracing your letter A is there any difference? i.e. Does it vary with the load on the router.
I'm thinking this could be due to some sort of oscilation of the router mount.
Paul

keyne
08-05-2007, 08:20 AM
Hi,

you are using AC servos with only the resolver? that is a quite coarse way to determine the motor's position, especially at low speeds. The resolver alone is good enough for speed control only.
I suggest you get a proper position feedback loop with an encoder!
If your drives have encoder connections, and operate with step/direction from Mach 3, you only need to connect the encoders to the drives (the positon feedback loop is built into the drives).
Going for steppers is a big step backwards, IMO.

Regards
K

corydoras
08-05-2007, 09:10 AM
Hi

PaulC: here are some photos of the Z axis

There is no difference if i make a deep or a shallow cut.

I think as someone mentioned in an earlier post that it could relate to my PI regulator settings.

I´ve played with them, and it still has some positioning problem.

If i put to much I on the regulator it starts jumping and vibrating, so there propably is to little power on the P gain to get it in the right pos when it is really close to the right pos.

Then when mach3 puts in more pulses in ne direction, it increases the P, and overshoots the pos, therefrom the waves?

Keyne: The servos have a buildt in encoder with 2000 pulses/rev.

These are also obtaiable from the servo amp if i wanted to use them for anything else in the process.

This is kinda frustrating, but maby putting a belt/pulley gear reduction of say mabye 5:1 infront og the 4:1 planitary gear i will get a miss placement of 20% of what I have now, right?

Of course this will slow down my rapids a little, but accuracy is more important.

Any ideas+

best reg
COry

paulC
08-05-2007, 04:21 PM
Your Z looks fairly solid. So that is bring us back to the servos again.
Other belt users have ended up with a 20:1 or lower gear ratio.
How many steps per mm does your system work out at and what are the max rpm of the servos.
Paul

corydoras
08-05-2007, 05:30 PM
I think the z should be sturdy enough, it doesnt show on the picture, but the 16 mm alu piece that you se passing through the cat track is going to a 20 mm THK liear rail suited on the top of the gantry, so it shoul be sturdy enough.

The motors are able to run 3000 rpm´s, and i have set the steps/mm to 60.5.

When I started up the drives i first ran at 22.2 steps/mm, but the drive has a build in electronic gear, so i changed the dip switches to get more pulses pr unit. And i still have more than enough speed.

The Y axis run without any problems, but of course this one is a lot lighter and much less mass to move around.

Same drive and motor and gear on both axis.

Mabye i should try some more reduction, although drives and pulleys are extremely expensive here in norway, im propably gonna use one of the advertisers on the forum from NY.

The next problem is that the cost of sending this to Norway is twice as high as the parts itself., about 100usd.

I´m visit Florida next month, mabye i should make em send it to my apartment there?......(if I can wait that long:):))


Best reg
Cory

paulC
08-07-2007, 10:24 PM
I wonder if it could be stretch in the belt. Not sure what could be done to prove or disprove this possibility.
Possibly cut a saw tooth pattern, increasing the speed with each tooth and see what you get. Make the teeth big enough to allow any stretching to end before deacelerating at the other end of the tooth.
And how tight is the belt? I understand you should be able to pluck it and get a note. It needs to be real tight.
Paul

harryn
08-07-2007, 11:01 PM
There are some good links for belt info on breccoflex.com.

As far as "tightness", it needs to be as tight as you are putting tension on the belt when driving the motors.

dan dimock
08-24-2007, 10:48 AM
Or plans for sale this type of machine or where may I go to find more infro.
Dan