Help required

[BensPlasmaAu]

I've been making cuts. I don't know exactly whats going on so I thought I would send it out to the experts. What is happening???

This is where the cut started and finished. Is it losing steps somewhere??

Slots aren't very even.

[KIDTech]

I had some cuts like that before and it was in my code, there was an "s10" code before every "M05" and it was making my machine stop motion for a split second then move again. I took out that line and it's fixed my issue.

[BensPlasmaAu]

Thanks KIDTech, will have a look at the code and see if its in there.

[tcaudle]

Kidtech: Must be using the wrong POST. S10 is a special DTHC OFF code in a POST for the DTHC and/or MP3000-DTHCII
Without the correct plugins and version of MACH3 ANY S code causes a slight delay.
If you are using a PROMO THC you need to use either the MACH Plasma POST or at most the MP1000-THC post that has the touch off

[tcaudle]

The bad cuts look to be mechanical related. It appears the outside cut is off from where i starts to where it comes back around and ends (should be the same spot) that could mean loose couplings. pulleys or gears or motors losing steps trying to move too fast

[KIDTech]

Yes, I have tweaked many different things so that they work on my table for now. I am running pretty well with the changes I have made to some posts. I had plans of building another table soon with your electronics but now will have to wait a bit as I bought a new cutter. I do know as soon as I can I will be upgrading to your height control. I'm learning the lessons of "pay me now pay me later" by going cheap in the beginning when I should have invested in better equipment.

[BensPlasmaAu]

The bad cuts look to be mechanical related. It appears the outside cut is off from where i starts to where it comes back around and ends (should be the same spot) that could mean loose couplings. pulleys or gears or motors losing steps trying to move too fast

Thanks Tom. Nothing is loose, it think its me trying to push the motors too fast and making them loose steps. I'm moving the table to a workshop space I now rent this weekend, so i'll have a go at making it slower. Hopefully that works. I've got a 3:1 reduction, MOD2.0 20 tooth gear rack. Roughly what speed/acceleration is about right?

[beefy]

Ben,

what do you mean by making the motors go too fast. You have given feedrates for metal types/thicknesses/nozzle sizes.

Do you mean you are using the biggest possible nozzle sizes so you can run faster ??

If speed is not an issue, then the smallest (= slowest) nozzle size that will cut a given thickness, should give the best cut quality.

Regarding losing steps, TEST your table at rapid speed. Start off at a known point on the table someone at the bottom left. Zero your X & Y DROs at this point. Make up a means of accurately checking this exact position. Make up a really crazing test run pattern zig zagging all over the place, etc. I had this run going all over my 2.4 x 1.5 table, doing sharp turns, little circular moves, anything I could think off, and all at RAPID speed. It lasted 15 or so minutes. When it was done I jogged my table back to the starting point and checked my XY DROs were back at zero, indicating there were no lost steps.

Now tell a lie, I did it with a function in Mach called VERIFY which uses your homing switches to do this, but the above will do the same thing if your table doesn't have this feature. If you want to test very accurately your start point could be spaced off your X & Y reference points and you could measure this space distance, then re-check it after the test run.

Keith

[BensPlasmaAu]

Hi Keith,

I mean i think i have the velocity and acceleration in motor tuning is set too high. Ill have to have a look tomorrow and post some screenshots etc of what ive got. Im also going to set up a test file like you have and see how it goes. Now ive got to remember how to stop the torch from firing. Ill remember when I get there.

Ben

Ben,

what do you mean by making the motors go too fast. You have given feedrates for metal types/thicknesses/nozzle sizes.

Do you mean you are using the biggest possible nozzle sizes so you can run faster ??

If speed is not an issue, then the smallest (= slowest) nozzle size that will cut a given thickness, should give the best cut quality.

Regarding losing steps, TEST your table at rapid speed. Start off at a known point on the table someone at the bottom left. Zero your X & Y DROs at this point. Make up a means of accurately checking this exact position. Make up a really crazing test run pattern zig zagging all over the place, etc. I had this run going all over my 2.4 x 1.5 table, doing sharp turns, little circular moves, anything I could think off, and all at RAPID speed. It lasted 15 or so minutes. When it was done I jogged my table back to the starting point and checked my XY DROs were back at zero, indicating there were no lost steps.

Now tell a lie, I did it with a function in Mach called VERIFY which uses your homing switches to do this, but the above will do the same thing if your table doesn't have this feature. If you want to test very accurately your start point could be spaced off your X & Y reference points and you could measure this space distance, then re-check it after the test run.

Keith

[BensPlasmaAu]

Thanks again all. I've tried a bunch of different things over the past two weeks. I made up the magnetic torch breakaway, but due to not being the best at making things in a hurry, it did not work and caused plenty of problems. It was too loos on the mounts and caused the torch to move about kinda uncontrollably. Back to the drawing board on that one. So I mounted the torch solid to the floating holder bearing and got a heap better result.

I have now properly tuned my steps per. Over a 1000mm of inputted travel, it only went 978mm. So I tuned all the motors properly to these new steps and it goes the right distance. I increased the velocity to 10,000 and the acceleration to 2500 to see how it goes.

I read the Mach3 manual (a bit of a dry read I must say) and found the but where it showed you how to program a simple loop program. I programmed the X (and slaved A) to move to 500mm and back to 0 ten times and then remeasured the physical starting point. And it does go back to the exact 0 position.

Then it was decided that I get to cutting some 3mm scrap steel lying around the workshop. I made up a simple concentric circle file and set up up to cut. The file and g code works really well, pity the machine doesn't work so great. Again the same problem I was having where all the circles and start and finish were out of whack. I have now discovered the culprit. The pinned drive gear which was fine previously now seems to have slop. Its a Ben induced fault too, so i'm a bit annoyed at myself. In any case, the machine does great on straight cuts down the X axis, but not so great when the Y comes into the scheme of things. So it means unless I want to cut either straight lines and boxes, the machine will be a bit of a failure until I get the drive fixed. Time to do some parts ordering. I should probably wait for some money first.

Here is a few pictures of my test cuts and some observations I've made.

As can be seen in the corner of the triangle, the "lost steps" is actually the y axis failure. I think the velocity is a bit too high as the cut has a bit of a wave to it. I made 4 of these plates and every one was close to exactly the same. On a good note, they came out at exactly 150mm where they needed to be.

This is the cut down the X axis. It's not too bad really. Probably needs some tweaking to get it a bit nicer, but the dross was very limited and dropped off even with my finger nails.

This is the hypotenuse of the triangle, and as can be seen the edged has a huge bevel to it. Obviously its 50% X and 50% Y so it cased plenty of problems. Might be OK if it was being welded then we wouldn't need to prepare the bevel to get a good weld onto it.

This was a bracket I drew up quickly. It's got plenty of small holes and it clearly shows the issues I'm going to be having if I leave the Y axis untouched.

I want to really thank everyone for their input over the past year and a bit as I've gone from idea of building my machine to the start of cut testing. Everyone here has been encouraging, helpful and inspirational with sharing of pictures of their table and the stuff they have cut out. Keep it up everyone. When my table is dialled in and I'm cutting, you'll see a fair bit more of me with sharing and ideas. Again thankyou to all.

Ben

[Capstone]

I increased the velocity to 10,000 and the acceleration to 2500 to see how it goes.

I think Ben you should consider stepping back and really asking yourself if going super fast is really the best way to start. Considering how long it would take an average human to hand cut a piece of 1/8 in metal with a hack saw, is it really necessary to go any faster than the bare minimum, which is for the average part about 5-10 minutes start to finish. Going as fast as you've previously mentioned, even on a fairly light well built machine would only mean you'll finish in 2 minutes instead of 10?! What's the point if you continue to get crappy results? Why not start almost too slow and see how things go from there. I'd rather start with a heavy drossed piece cut perfectly. The MACH3 motor tuning section will tell you the "G's" for your settings. My "G's" work out to .2 with 350 for velocity and 80 for acceleration and I use 30amp consumables. I get nice tight accurate cuts with more dross than most, but I'd rather let a perfectly cut piece sit in my acid dip a little longer than waste more metal on bad cuts especially for parts that need to fit a mechanical fabrication. Your motor speed settings seem to be the real issue. too much speed; just because the machine can, doesn't mean it needs to go that fast. Decreasing dross is the last issue to worry about when you're making parts to size for auto racing applications. If you're cutting out 40 of the same artsy items on a single 4x8 sheet of 18ga, customers won't care really if they aren't identical one to the next, but the dross matters because it adds a lot of time to the work and profits, which is why so many here make a big deal out of it. Seems like you're only cutting brackets and pieces to weld up for mechanicals, so saving time doesn't help you much.

[BensPlasmaAu]

Thanks Capstone. I put it up because it seemed like a good idea. I will be dropping the velocity and acceleration back down when I get back out there, and once ive fixed the y axis problem. I also cant see the point in going too quick. It was just something I did to see what happened and how the machine behaved. Anyway a good experiment and it will be better slower anyway.

How does all this go say, with the feed rate of cutting when the tool says to cut at 8000mm/min and I have it set in motor tuning to 1000? Does it just completely slow down and force it to cut at 1000 and not where the recommended feed rate is. Im a bit confused by this. Probably not explaining myself very well at the same time though.

[Capstone]

Setting your motor's acceleration and velocity is somewhat separate from the cutting speeds. MACH will slow/adjust the motors to the correct cutting speed based on the recommended speeds from Hypertherm.

What you're adjusting is simply the top speed and how fast it gets to that top speed. So when cutting using those high settings you will likely cause some serious shudder-cuts as the torch attempts to change directions at speeds that can't be accurately made by the gantry due to inertia's that go beyond the mechanical rigidity of the components, like bearings and guides. I've learned first hand, it's a delicate balance and you will have to play with the settings as you deal with different thicknesses. There's a lot more technical explanation as it relates to acceleration and velocity here.

[BensPlasmaAu]

Awesome link there Phil. Just so I don't lose the link/can find it again quickly, I'll paste it in here. I do everything in metric because it makes sense, so I think a general acceleration for me would be 900 - 1000mm/sec/sec. That will give me 33,000 meters per minute............ that seems awfully quick to me now I think about it. And now the more I think about things, is your settings in inches or mm? 350ipm is equivelent to 8890mm/min. 80inches acceleration is equivelent to 2032mm. My settings (10,000 and 2500) are pretty close to yours. Ah isn't it great to have a standardised measuring system worlwide just to make it confusing.

"with motors on and locked go to each axis and try to move them on both directions, Then grab the end of your torch and se howfar it deflects when you put moderate sidways and back and forth pressure. The bearings on the Z are adjustable and need to be tight enough to no have slop . The settings you have may be okay. It depends on if you have the belt reduction option or it is direct drive.

Acceleration of 60 is high normal (Accel is in IPS/sec so multiply by 60 to equal ipm.sec. 60 means you can go from 0 to 2600 IPM in one second or to the commanded feed rate in a fractional second.

You might try lowering the accel to 35 and see if it makes any difference. If it does it STILL indicates something is loose mechanically , Lowering the accel just makes the moves less aggressive.

Another good way is to lay your hand lightly on the Z as it cuts and "feel " the moves. Motor tuning won't cause problems like this. That is overshoot in one direction then in the opposite....oscillation. i would be willing to bet the torch is wobbling or the z bearings are loose. It almost impossible to make a gantry move back and forth that quickly if if you try!

Velocity is the max seep you can go and is in IPM. Acceleration is how fast you can go from 0 to full speed or full speed back to zero.


Assuming the artwork is clean and not all ragged (would show up in SheetCAM ) then it probably is not electronics or software in nature"