Help dialling it in please :)

davek0974 · Post by **davek0974** » Mon Jun 15, 2015 3:50 pm

Could it also be because my gantry is very light? Probably only weighs 5kg tops.

Maybe not enough mass to absorb transmission effects?

Just guessing

spring tensioners won't work here as the belts must drive both directions and spring tensioners only on the slack side - linear positioners like a CNC table have no slack side to the belts.

SeanP · Post by **SeanP** » Mon Jun 15, 2015 5:18 pm

I thought I had gone to tight with my belts once, I thought it was putting strain on the motors so backed them off a little.

Gee with that weight of gantry you should be zipping about no probs mine is at least 10 times heavier

One thing I have noticed on my motors is I get a humming vibration which does send a slight vibration through to the torch, just at certain speeds.
Could it be that causing those ripples Dave, I don't see it on your 45 cuts though.

beefy · Post by **beefy** » Mon Jun 15, 2015 6:06 pm

I always try to use elimination when troubleshooting.

Put a pen/fine marker in your Z axis and draw on a sheet of card/paper. Do you still yet the serrated effect. If it only happens with cuts then does it happen with THC off, or only with THC on.

Keith.

davek0974 · Post by **davek0974** » Tue Jun 16, 2015 2:44 am

SeanP wrote:I thought I had gone to tight with my belts once, I thought it was putting strain on the motors so backed them off a little.

Gee with that weight of gantry you should be zipping about no probs mine is at least 10 times heavier

One thing I have noticed on my motors is I get a humming vibration which does send a slight vibration through to the torch, just at certain speeds.
Could it be that causing those ripples Dave, I don't see it on your 45 cuts though.

It could be vibration, i think the 45 might hide it due to the arc being slightly bigger, the 30XP has a very fine arc.

davek0974 · Post by **davek0974** » Tue Jun 16, 2015 2:46 am

beefy wrote:I always try to use elimination when troubleshooting.

Put a pen/fine marker in your Z axis and draw on a sheet of card/paper. Do you still yet the serrated effect. If it only happens with cuts then does it happen with THC off, or only with THC on.

Keith.

Will have another go with the pen later, I am fairly confident that it's not the THC saw-toothing along as it can't move that fast that regularly, it would be buzzing all the time and I can see the motor shaft when it's cutting - it only steps up and down occasionally as expected.

davek0974 · Post by **davek0974** » Tue Jun 16, 2015 11:45 am

OK, here we go, results of a solid day messing around.

First, I checked the machine over, made a few adjustments, cleaned the rails etc.

Back to straight lines - 3mm steel, 45A, speeds from 3800 to 2800mm/min, 106v measured

I thought it odd that the straight lines came out perfect but with a shape we get dross. However it WAS low speed dross and fell off easily. At the lower speed, the centre hole was better and not quite so trashed at the ends.

I also ran many dozens of different acceleration tests and reached the conclusion that they made little difference! Yes, i got jerky when pulling 0.18G so I backed it off to around 0.09G

On to the fiddly tab part of my test.

This time I added loops on the externals and 1mm fillets to the internals. Results were much more favourable, front...

Edge of the above test...

Then I did a pen test, it does show wiggle as exaggerated on the picture, on the exit from the loops, see below for my reasoning.

On close examination, the head on my creation can be twisted about the gantry fairly easily - the 44x44 aluminium section I used while strong in the beam is weak in torsion so as the gantry slams to a lower speed for the corners it starts the head wobbling as seen in the picture above. There is no slack in the carriage, it is all torsion or twist.

The rails I am running the carriage on were next scrutinised and with the motor belt off, pulling the gantry back and forth showed tight spots, rough motion and generally not good feelings. Cleaning the rails heavily with scotch brite and readjusting the bearings helped but my design is poor here because the rails are at bed level and get showered as its cutting.

The good news is that I managed to run my backlog of jobs and with some cleaning, they were not too bad. The biggest issue is that ugly sawtooth effect which is pretty bad, much worse than with the 30XP, you can hear the arc ripping as it moves. Presumably worse with the PM45 because of the increase in power in the arc amplifying the flaws in my machine.

So, there we go, apologies for the long post but I'm just laying out there as it is.

Where do we go now??

Well, I know for a fact, based on the examples of work I have seen that the PM45 is a great machine, well capable of producing what I need. However, my machine has reached it's limit I feel, it is now the weak link in the chain and therefore must be dealt with I think.

Please correct me here if I am being a fool but my plan is now to send the 30XP back as arranged, then build a new table, using this one to cut whatever is needed for the new build. This little table owes me nothing, it has earnt enough income for me to get the PM45 in a pretty short space of time so I have no worries about dismantling her.

Here's the poor old girl...

I now need to find some plans for a tried and tested design for a small table, I can re-use the Z-axis I think as it was a purchased assembly. The motors are pretty big, may need another for slaved drive on the gantry. The drives are ok, I have a spare for the slaved motor, will need a bigger PSU but the rest of the electronics are all new CandCNC stuff so will be re-used.

Rack and pinion drive??

Any tips for what to look out for?

I only have space for a 1m x 1m machine but thats a fair bit bigger than what i have, but will want it to perform as required for the PM45 this time

SeanP · Post by **SeanP** » Tue Jun 16, 2015 4:56 pm

Going off that test piece Dave you don't look that far away to me, looks better than what I cut really!

Like you say though on the thinner stuff you want it spot on for the faster speeds.

Gee one thing leads to another!

How are you finding the angularity on the 3mm parts? going off the pics they look decent enough to me but you can't always tell for sure with photos.

Post by **acourtjester** » Tue Jun 16, 2015 8:40 pm

Here is a cutting accident I had the other day cutting ¼” steel at 49 ipm with a PM65 running 45 amp with 45 amp tips. I had a hick-up and backed up the g-code to the start of the parts and it cut just off the edge of the original cut. It ended up 0.065” thick with very little angularity dross still on part of the cut. Also so a stick figure I cut a while back these are from my DIY table.

davek0974 · Post by **davek0974** » Wed Jun 17, 2015 5:08 am

SeanP wrote:Going off that test piece Dave you don't look that far away to me, looks better than what I cut really!

Like you say though on the thinner stuff you want it spot on for the faster speeds.

Gee one thing leads to another!

How are you finding the angularity on the 3mm parts? going off the pics they look decent enough to me but you can't always tell for sure with photos.

Thanks,

the angularity is a bit odd, being worse on one edge but my biggest issue is that damn ripple caused by my rubber-like machine twisting and flexing.

A steel frame build is the way up i think.

davek0974 · Post by **davek0974** » Wed Jun 17, 2015 5:09 am

acourtjester wrote:Here is a cutting accident I had the other day cutting ¼” steel at 49 ipm with a PM65 running 45 amp with 45 amp tips. I had a hick-up and backed up the g-code to the start of the parts and it cut just off the edge of the original cut. It ended up 0.065” thick with very little angularity dross still on part of the cut. Also so a stick figure I cut a while back these are from my DIY table.

thanks, that stick-figure is insane! I would have thought it would just have vaporised when cut.

Nice

Post by **BTA Plasma** » Wed Jun 17, 2015 8:51 am

Dave if your using mach go to general config and click "stop CV on angles greater than XXX" and put in a number less than 90. This will give you an exact stop in the corners. The reason you will want this is Mach is relying on your acceleration numbers for your axis tuning solely for directional changes (Making a corner) and will radius a corner. You need to make Mach stop dead in the corners. If you don't have a good THC with an anti-dive for corners it will drop the Z and may touch the steel blasting away your nozzle.

The wave on the outside is usually a loose Z or Y axis, the gantry is rocking or the torch mount is loose. It is not a loose X axis (the dual driven)

davek0974 · Post by **davek0974** » Wed Jun 17, 2015 9:21 am

Thanks,

I was advised early on to "leave all that CV stuff turned off" it's done well up to now when i need more speed. I can have a play with that.

The wobble appears to be my gantry twisting under the weight of the head/THC being thrown about, apart from belt stretch there is no slop at all, it just twists the gantry.

I could just rebuild the gantry and gain some quality probably but i also want to lift the rails up away from the mess and that means building a new frame - by the time i have done that i might as well have just built a new frame from steel

Rebuilding the gantry won't get me any more speed though.

davek0974 · Post by **davek0974** » Wed Jun 17, 2015 2:32 pm

One of tonights jobs - 3mm steel back plate for a house plaque, 285mm x 115mm, 45A, I found the results interesting...

The long edge - perfect, even considering my very low speed...

The short edge, both were the same - the dreaded ripples...

And on the back - two perfect dross-free edges and two drossy edges!

I have no idea what it's telling me apart from there is an issue with the Y axis on my machine, but I knew that anyway, but why would such a little ripple cause so much dross when both sides are cut at the same speed?

I also ran some 1.2mm 304 stainless, 25A, 3800mm/min, again - held the speed down to help minimise the ripples. Very good result and got the job done.

Post by **acourtjester** » Wed Jun 17, 2015 4:16 pm

If your dimensions are correct on the cut part that tells you the calibration is correct but your speed has a problem. Sure looks like the motors move at different speeds that is what the dross is telling you.

davek0974 · Post by **davek0974** » Wed Jun 17, 2015 4:45 pm

thats what i thought a long time ago but dimensions are good and all cuts come out very close on size.

I tried very hard to make dross on my straight line tests and could not get it to do it, i am running at nearly half book specs and still no dross on the x-axis cuts but I get ripple and dross on the y-axis

maybe the ripple is upsetting the arc stability and sapping its energy density some how???

SeanP · Post by **SeanP** » Wed Jun 17, 2015 4:55 pm

I have noticed same on mine on some parts with dross on 2 sides Dave,,not sure why that happens though.

Post by **tcaudle** » Thu Jun 18, 2015 11:33 am

I have no idea how Mach3 handles corners, it's a bit of black art it seems and the best advice always comes back as "just let it handle it", however, I do know there is a relationship between speed and acceleration whereby things can go badly wrong - the internal corners turning into radii when you set acceleration too high for example

MACH uses CV....constant velocity. It looks ahead at the toolpath Like you would if you are driving looking at the road ahead) and tries to make the motion as smooth as possible and at the highest speed up to the commanded feedrate that it can. Like a driver when there is a corner coming up, it takes its current velocity and possible acceleration (from the settings) and says "can I make that corner at the angle called for at the speed requested or do I need to "cheat" the toolpath and slow down and round out the corner.?)
There are indeed factors and you can dink with the CV settings and improve a specific cut but make other ones worse. You can even drop out of CV mode to what is called Exact Stop where it runs up to the corner completely stops then changes direction. That can get really ugly on a curve because in exact Stop mode it stops and starts on every node. It can jerk like a sick dog following a curve Anything you do to the speed will have other reactions. lowering speed on a plasma cut cause the plasma arc to cut less metal and more air and to keep the current constant the machine raises the voltage. It also makes the flame cut more on the sides looking for metal and the kerf gets wider. As the voltage climbs the THC attempts to correct by moving the torch down (torch dive), if the volts change is rapid enough and the THC is slow enough you can use voltage based anti-dive sensing to lock the Z. You can also use velocity sensing anti dive but sometimes the deceleration trigger point is hard to predict and once again depends on the THC response rate .

Higher acceleration means sharper turns and less slowdown in and out or turns but your table can only achieve a given amount of acceleration on a axis based on the weight (mass) of the axis and the Force (derived from the rotary torque of the motors) That number is expressed in different ways as in G's (or milli G's) or IPS/sec or M/sec/sec. Exceed what your table can actually do and you get lost steps and skidding. The faster you are moving the more acceleration you need to keep the toolpath exact.

davek0974 · Post by **davek0974** » Thu Jun 18, 2015 11:56 am

Thanks Tom, useful insight there.

Do you have any theories on why there is dross on the wobbly edge of my job above??

BensPlasmaAu · Post by **BensPlasmaAu** » Fri Jun 19, 2015 3:59 am

Hi Dave, what are your velocity and acceleration settings at the moment? Im havng almost the same issues as you were. Im pretty sure my problem is ive got things set way too fast. My reduction drive is similar to your ratio at 3:1.

Ben

davek0974 · Post by **davek0974** » Fri Jun 19, 2015 4:21 am

Hi

when these jobs were done , i think it was set at 6000vel and 900acc, pretty certain that was it.

It was still wobbling so needs to go lower, the machine is happy at 1000-1500mm/min cuts but i need way more for thin stuff, turning down the current helps a little but not the right answer.

Bigdogbro1 · Post by **Bigdogbro1** » Thu Aug 27, 2015 10:02 am

I run a belt drive system. I designed it to use belts because the machine sits in a pole building where the humidity gets high and the inside of the building gets wet. I didn't want the machine gantry parts to rust so it's all built from aluminum and stainless and plastics. We run a HT PM45 T45V hand torch using T30V torch consumables and normally cut 16ga CR steel. The plastic rollers I use have a bit of looseness in them side to side NOW but it can be easily fixed using a thinner inner spacer to clamp the wheel and bearings tighter together. This causes a bit of the same serration you see in one direction change from Y to X because the Z wheels needs tightening. We are very happy with the results and have been running 30A, 75ipm, 86v. .04" gap, .10" pierce with the CandCNC Bladerunner Ethercut DTHCIV, 4x320 steppers. I like the detail we can get and the 100% duty cycle at the 30Amp setting. We plan to run near 125+ once all is fixed up. Giving some time for new mechanicals to settle in and too busy right now to tear down.

Check your bearings for same and other mechanical movement, X axis gantry cocked on the Y beams and not rolling smooth...chatter, debris on bearings etc.. Doesn't take much to cause that wiggle. Over stretching the belts too tight can cause the belt to have different belt cog pitch and not match the pulleys. The belt tightness also affects the mechanical distance travel per stepper motor STEP. I have made custom belt tensioners recently to the Y-axis so I can tune the tension AND the left and right rail tracking position easily. I have my Y and its slave-A reduction drives tied together with a shaft the runs thru the X-axis beam. I have yet to come up with a reliable method of setting the belt tension. Hope this helps.

I included a video link to a sign cut. There are a few other videos showing the machine on my channel as well.

https://www.youtube.com/watch?v=DCGO1P2gQUY

davek0974 · Post by **davek0974** » Thu Aug 27, 2015 10:37 am

Thank for that,

things have moved on a bit now, I have nearly built the new 4x4 and the old one will be dismantled after thats up and running.

I will get some 30A consumables though, could be worth a try.

Bigdogbro1 · Post by **Bigdogbro1** » Thu Aug 27, 2015 11:41 am

Dave,

If you are cutting thin 18-16ga CR steel primarily then the T30V torch 30A consumables will give you better details. You can run slower and run the Hypertherm PM45 at 100% duty cycle (30A) which is great for those larger cut jobs that take over five minutes. The T30V non-shielded consumables won't last as long as the shielded PM45 but we are happier running slower IPM with finer details at the moment. According to the Mach3 maintenance meter, we are getting about 8 to 9 hours of run time but we're not sure exactly how that's counted. We think it's the machine movements and not the actual torch ON time. Regardless, the amount of money coming in more than pays for a few extra electrodes and nozzles.

T30V parts: Needed for installing on a PM45 T45V hand torch (You will no longer have the T45V ohmic sensor capability, unless you make your own modification)
.032" kerf, NON-SHIELDED fine feature consumables
Electrode: 220478
Swirl ring: 220479
Nozzle: 220480
Retaining Cap: 220483

I noticed on your earlier picture of the table, you have a shaft drive for the Y and slave side similar to mine originally. I was getting torsional twist in that 3/8" diameter hardened shaft. I accommodated in my design to add a A-slave motor to help lock down that end in case I had the twist problem. After adding the extra A-slave motor on the Y-axis all that wiggle went away. I still have the shaft installed and all works much better. I wanted to try a single Y motor to reduce the mass. The extra motor was an easy add because I had the plan "B" ready to go. Go with the double Y axis drive motor system, you will be glad you did. Go lightweight, STRONG and powerful to get the 350 IPM target. My overhead gantry comes in at about 48 pounds and I could have made it lighter weight by sizing down the X beam to a 2"x2" (6 lbs.)from the 2"x4" (12lbs.) used currently and dropped the Y/A driveshaft. Could even use some G10 fiberglass plate material on some of the side plates instead of the 1/4" aluminum plate.

davek0974 · Post by **davek0974** » Thu Aug 27, 2015 2:09 pm

Yes i do suffer twist in that cross-shaft, but I'm not worried about it anymore, heres a link to a video of my new build...

https://www.youtube.com/watch?v=n34DDnw8org

Almost ready now

I have twin slaved 4Nm motors on the Y-axis and a single 4Nm on the X, its like chalk and cheese compared to the old one, and easily hits my 9000mm/min target for speed.

In my main build thread I also have posts relating to the results of having an FEA analysis of my gantry, it seems i got the design just right

Bigdogbro1 · Post by **Bigdogbro1** » Thu Aug 27, 2015 2:45 pm

They always look so great when they're fresh and clean. Good luck on the new system...keep us posted!

Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)

Re: Help dialling it in please :)