....and in the spirit of troubleshooting where all of us are trying to be positive about mechanical design issues, Nils still has a challenge that is not resolved. I have seen the cut issue up close and personal and it has me stumped.

I haven't seen this kind of repeatable chatter/cutter issue ever. Not on my bridgeport, not on a Mazak, not a multicam, not on a shopbot series 1, not a shopbot PRT, not a thermwood, etc.

Nils' Mechanically sound machine has almost identical components to two other MM's here in Orlando and I cannot reproduce Nils results. Thus, I am challenged to help him resolve it.

Any positive discussion will be helpful....Pissy banter just makes it harder.

Let's try to keep it nice guys...

So, if someone would really like to resolve the perceived backlash issue as the culprit of this cutting problem, then chuck up a bit in the collet, cut some diamond shapes in solid surface or hard material about 5" in size with 45 degree sides splitting the x/y plane and publish the results. Now that would be a helpful clinical trial instead of a heated debate on who's is better and bigger!

Science, not politics.

Awwww....come on Sean. I was enjoying it. :D

As requested. This is 3/4" mdf cut with a 1/4" carbide bit at around 60 ipm in a single pass. The other side didn't have a pattern but there was a small amount of chatter. I used chalk to help highlight the pattern.

Look at Nils picture in post #1. It looks almost identical.

Doug Thanks. your cut is in diagonal?

You can test 30ipm 60ipm 90ipm in the same cut line and the pattern is identical betwen chatter marks?

Yes. It was cut on the diagonal. I haven't tested at those speeds but I will do so. Give me a day or two.

I'm starting to wonder if maybe one of my wheels on the car is not fully seated in a small section of the rail. The reason I say that is because this pattern is not stepped. The short sections of the pattern are slightly raised. I could easily understand if the pattern was like a set of stairs. I'm wondering if the car is getting pulled back and forth slightly as a result of the changing amount of pressure applied as the motor is given a step but then rocked back in the other direction as the router approaches the point right before it is given another step.

Unlike Nils, I'm not that worried about it. If I can fix it, I will but I'm not going to spend a lot of time tinkering with it.

The raised section of the pattern is approximately .003" and are about 1/8" long. Easily sanded out.

...Now that's what I am talking about! :D Thanks Doug! That's ozarkian help :)

Wow.. a 19 mm thick mdf cut with a 6 mm cutter in one pass.....???? I have never tried that before. Don't even think my little 500 W Makita router will handle it.

I always thought that you should not cut deeper than half your bit diameter...therefore if I use a 6mm cutter I will not go more than 3 mm per pass. With my little router I prefer to cut about 1.5 mm per pass on a 6 mm cutter. Takes more time to cut, but it's not so hard on the machine.

...interesting. That pattern looks familiar. Thank you Doug - very helpful!

Doug,

Thanks for taking the time to do some test cuts.

I don't believe that I could cut 3/4" MDF with 1/4" bit in one pass. I will have to try it when I have some extra MDF. I also don't think I would have that little chatter.

I would like to see you cut the diamond pattern in a piece of solid surface material and see what kind of chatter you get. It it much harder than MDF.

I have a question. I looked over your build thread and I could not find where you listed what motors you are using. I think they are the geared PK296A2A-SG7.2 motors but I would like to verify.

Also you mentioned that you set your debounce setting to 500? You did that based on a J.R. thread. I searched his thread but could not find it mentioned there but I might have missed it. Why did you set the debounce level?

Also you changed your kernal to 35,000, Why?

Thanks for the input.

Over the week-end, I opened a thread on the Mach forum site were I briefly described my problem and asked if there was a way to test if the computer had a problem with regards to the kernal. I did not get an answer on that.

But I did get another comment about the settings for CV (Constant Velocity). It was suggested that CV can have an effect with regards to chatter. So I did some reading. There is a separate PDF document called CV Notes that briefly deals with some of the parameters.

In Mach, if you click on the Config tab and go to General Configuration, you will see on the right side, a General Configuration area and a CV Control area.

Under General Configuration (GC), there is a "Look Ahead ___ Lines" with the default of 20.

Under CV Control Area, there is "CV Dist Tolerance ____ Units" with the default of 180.
There is also a "Stop CV on angles > _____ Degrees with the defalut of 0.

Based on my reading of the CV Notes, I made the following changes:

CV Look Ahead 200 lines.
CV Dist Tolerance 100 units
Stop CV on angles > 90 degrees.

After restarting, I cut my Diamond at both 35 and 70 IPM and the cuts were smooth. But when I cut at 100 IPM, I still had chatter.

I have since changed the CV Dist Tolerance to 50 Units. I recut the Diamond at 100 IPM and no change in chatter.

I found the results most interesting and I am going to post my results on the Mach forum site. I also wanted to share these results with you.

Nils,

Those are the motors that I'm using. I don't recall resetting the debounce nor why I would have done that. If I remember correctly, I set the kernal at that setting because my testing showed that the computer would allow it. I'd love to cut some solid surface material for you but I don't have any. I'm sure you could cut 3/4" mdf at 60 ipm in one pass since you are using the same motors and router as me. However, I can hear it slow the router motor at points in the process. I'm sure Sean will be rolling his eyes when he reads this and I know I'll probably have to replace the bearings in the router a little sooner but that's okay. Oh yeah, and that single pass cut also included a sheet of .030" aluminum.

Doug, I would never roll my eyes! LOL. With the right bit and feed rate, not a big deal. Of course, you have about a billion CFM of air rushing past your bit on the way to the cyclone to help keep the cutting area clean and cool! :D

Nils,
With the new settings in mach, do you think you are past the rough angle cutting now? Or are you still tweaking. Myabe some comparison pics of the differences?

Heath,

I still am having a problem. This morning, I increased my router speed because I had lowered for some thing else and I cut a straight piece and things were smooth. In fact, I pulled a Doug and cut the .5" solid surface material with a .25" bit at 400 IPM and it was very smooth. Less I forget, I cut it in one pass!

I thought this is a very good sign. So I cut my Diamond pattern and the chatter was there. The Mach people are thinking that it might be bit deflection.

I am not doing any thing else until I have received and installed the new ABEC 7 bearings that I have ordered for my router. I want to see what affect that has on the results. Since I already have 4-5 thousands runout with my router. I think the new bearings will tighten things up.

I will be happy to post my results with pictures.

Thanks Nils,
I am sure your outcome will be very userfull to everyone. Have you changed the router bearing before? I am eager to hear how it goes for you.

No, I have not change the router bearings before.

Nils,

Paco has a great site/blog detailing bearing removal for the Milwaukee router.

http://pacosarea.blogspot.com/search/label/router

Thanks Greg,
I was just looking for that for Nils. :)

Greg J and Heath,

Thanks for the paco site. I have seen his site before but it will come in handy when I replace my bearings.

I am replacing my bearing with a better rated bearing than with the standard factory bearing. The bearings they use are not rated as high as the highest motor setting on the router. Chopper suggested and it made sense to me to use bearings that are rated higher. The lower larger bearing, I am replacing with a ABEC 7 rated bearing and the top bearing is rated for 34,000 RPM. Which is much higher than the router can handle. It has a higher tolerance and should help with runout.

I will take some pictures and post my progress.

Thanks again for the info.

Here are some pics of the cuts I made today they are in aluminum
6061, cut at a 45 degree angle to the machine
//chopper

Hi Chopper, may I ask how thick that aluminum is and what were your cut parameters? Specifically, inches per minute, and depth per pass. That cut looks pretty good to me.

Chopper,

That looks as good as a mill would cut. How about more details.

Hi Nils, I know a computer swap was recommended for test purposes, did you every try that? The reason I ask is because I have had a problem before and believe it or not, each machine I tried (to resolve a different problem than you are having) produced a slightly different sound on my Mechmate which indicates to me the PC does play a role in your cut quality or should I say cut "characteristics". In all, I believe I have had 5 different machines with identical settings in Mach driving my gantry and each made a noticably difference sound when jogging at the exact same speed settings - some machines were smoother and some rougher. Don't ask me why but it is definitely worth a try on your trouble shooting adventure even if that means borrowing a friend's machine - P4 vintage works best in my tests. Well worth the time and effort. One thing I've learned working on this wonderful machine is that sometimes, the solutions that don't seem logical ironically turns out to be the fix. Keep an open mind and don't throw in the towel as the solution is probably right around the corner.

Hope you can get your proble resolved.

David,

I have not switched out computers. I don't believe my problem is with the computer but I appreciate you mentioning you results.

Today, I finished replacing the bearings in my Milwaukee router with higher tolerance ones. My runout went from 4-5 thousands to just under 2 thousands. When you run the router it does sound like it is quieter.

But I have actually gone backwards in the quality of my cut. I was cutting smooth at 35 and 70 IPM in solid surface material and having problems at 100. Now, I am no longer smooth at 35 and 70. I did not expect that I would go backwards. I am not sure what I am going to do next because I would have thought that changing the bearings would not have made things worse.

I am begining to think the computer might be a problem.

Nils, if you do try a different machine, whatever you do, don't go with a modern dual core. The power saving features built into the CPU (which can't be shut off) will not give you a smooth pulse train. This was the root of my problem. If you look in the trouble shooting section of the forum, there is an entire thread dedicated to this.

I am now using a Celeron 2.4 with 512MB of ram and to this day, not a single blip on the control side. My P4 3.06 with HT worked even smoother but, that machine is dedicated to my DVR system and I'm not going to change over unless I have to. Stay with P4 era PC if you do a swap or you will most likely experience the same issues I did where the steppers lock up during a jog or cut. I hope my experience can help you avoid a few pit falls. Now, I'm battling with interference from a VFD hooking up my Chinese Spindle. Good Luck!

the aluminum is 6061 and .250 in thickness,
the cut parameters were .020 removed per pass at 40 inches per minute on the feed rate, with no lube, 8,000 rpm on the power head 2 flute up spiral bit
//chopper

Chopper,

Thanks for sharing the cutting parameters. The cuts look very clean. It is clear that your system is very effective. I think cutting aluminum with a MechMate "on the diagonal" must be one of the most challenging jobs around. I assume that the other cuts I see on the aluminum were also made by the MechMate.

I looked back at your build pictures and I don't see the "power head." Did you end up with a spindle like you were discussing?

Any how, very nice work and thank you for sharing your results with us.

John,
no spindle, just a Milwaukee, router....
//chopper

I am making a quick visit to this thread, and will then stay away again.

The major marks shown on the work piece are maybe about 1/8" to 3/8" apart. There is zero chance that is caused by spindle/router bearings, gearbox backlash, spacers behind V-wheels, machine vibrations, or any mechanical defect. The fact that the pattern is repeatable and consistent indicates that the machine is being instructed to cut that pattern - in other words; the PC is telling the MM to cut the pattern.

When the PC was relieved of calculating load (paralleling the drives) the pattern changed. That proved that the PC was responsible, but somehow the opposite conclusion was reached.

Replacing the PC might not solve the issue, you may get a similar behaviour - Doug's test shows that.

When I point a finger at the PC, it includes hardware and software. I mean the "electronic pulse generating system". A PC can very easily generate disturbances onto smooth pulses if it is doing unnecessary tasks in the background. Thus, when I say to try another PC, I mean to try one that is set up differently. However, it could help to change the setup of the existing PC.

I believe that this thread has been sidetracked by focussing entirely on the mechanical aspects, and that tempers get inflamed by belt-drives vs gearboxes. I am 100% sure that a belt-drive will do zero for curing the particular problem mentioned at the start of this thread.

An inexpensive experiment would be to copy a good test file to a thumb drive, download the EMC live cd
http://www.linuxcnc.org/content/view/21/4/lang,en/

boot your computer with it, and run the test reference file from the thumb drive.

EMC isn't horribly difficult to configure, and you don't need to swap computers, or wipe your current one.

Windows is fundamentally not a realtime operating system. I don't know how Mach gets around this constraint. EMC is working nicely for me with a *really* rudimentary computer.

http://bbs.homeshopmachinist.net/sho...0&postcount=26

On edit, I read now that you can't use a thumbdrive with mach... is that true?

Gerald,

Thanks for your input and for looking at this again.

The computer I am currently using with the machine is a Dell Pentium III. It is bare bones in that it has Windows XP Pro and Mach. I didn't add virus but I am going to check that today. I also did not add any other programs that should be running in the back ground. But I am going to look at Task Manager today to see what else is running.

I have been using my memory stick to load my GCode files and I know others do the same. I will try running from the hard drive because there could be a communication problem.

There is another MM builder that lives close to me that did not document his build. I recently visited his machine to look at his belt drives. I ran the diamond cut file on his machine at 100 IPM and he did not have a problem. I associated the smooth cut to the belt drives but it could also be the computer. He has recently offered to bring his computer over to try it on my machine. I hope that we will be able to do that on Tuesday. I will report back on the results.

Jeff, I have been trying enough things that have muddy up the water, I don't want to try another software package. I am well aware that windows is not a realtime operating system from my experience with simulators and realtime responses. But I appreciate you suggestion and for try to help me.

Nils,
Your windows installation might be stock but there will be lots of bloat that MS installs by default. These extra services have been known to cause problems with cnc activities. Also antivirus scanners can be a real problem, especially the real time scanners and when a scheduled scan starts. Your MM computer does not need to be connected to the internet so the need for antivirus is very small. You can look at Black Viper's windows services site to see what services can be turned off that you do not need. Look at bare bones setup for starters.

I checked and I do not have any virus software running.

I also ran my diagonal file with the GCode file loaded on the hard drive and not on the memory stick.

I opened task manager. In the performance window, the peak was about 15% activity while mach cut the file. I then opened the process window, Mach was the only process that ran and it peaked at 15%.

I don't believe any thing else is running but I am going to try another computer to verify that there is a problem with the current computer.

I will keep you posted.

I have just returned from Johannesburg where I was offered a position at a very large fruniture manufacturer managing the maintenance at two plants (turned it down even though the salary offered was astranomical by my standards - all the money in the world cant justify an abusive boss!)

But to the point...at this plant they have 9 big iron CNC's, two of them massive SCM Ergons with three heads and two huge extremely fast moving tables. The cut quality coming off these (R7 million) machines in chipboard was no better than what I get with my direct drive Mechmate when driven a bit fast. If I slow down, my cut quality is a bit better. I am still going to fit belt drives as I feel it will improve things but maybe we ARE expecting a bit much from the system when we try and get milling machine quality cuts from a gantry router. The fact that chopper is getting such good cuts in ally is heartening but there are just too many variable factors in computer make up, build quality and respective machine rigidity to make comparisons on this small scale.

Try another computer or two before chasing any more suspected problems.

Allan_C, all the more reason we need a sample test board to determine how well we built our machine. :D

p.s. I agree on the boss thing, best to walk away:)

:):):) Sorry for the thread drift

Today, I tried another computer that is being used on a MM that was built but not documented in the greater Orlando area.

Using the other computer we got the same cutting results on my machine. If you remember I did cut my diagonal file on his machine and it was smooth.

I thought my Z Slide was tight but he thought it was loose so I tightened things up but it did not help. On closer inspection it looks as though the Spider is flexing. I do have mamba or 3/16" Spider.

What is the best way to stiffen the spider?

Thanks

I "Stiffened" mine with some .375" rod top and bottom..

Dont forget to see this
Beefing up a thin spider