Building a new table, any tips?...

[acourtjester]

I have no issues with the all in one box there is no HV involved inside the controller just normal 110 volts AC for power and signal voltages.
I even have the feather touch mod inside the box.

[davek0974]

That sounds promising, i have room in the main cab to fit the PC, will just need an external USB hub for the keyboard, mouse, joypad and file loading, plus a video extension lead.


Sounds good to me.

[acourtjester]

The last all in one controller I built.

DSCN1732.JPG

[davek0974]

No filter media on the cooling fans??

Looks good though, there is very little heat in my main cab - the drives run practically cold so i doubt i'll need a fan in there, may add a small pc one to circulate the air though.

[acourtjester]

Filter on the input side of the air flow fans put cooling air through the controller

[davek0974]

Panel finished...

IMG_0996s.jpg

Mesh below the slats...

IMG_0997s.jpg

Bit busy up that end of the shop

IMG_1003s.jpg

Stuck the monitor on the wall, seemed a good place, still working on the keyboard and mouse/trackball, might be a slide-out under the front of the table. I couldn't fit the PC into the main cabinet - not enough space

IMG_1002s.jpg

One odd thing I noticed when motion testing with the proper PC, the X & Z axis worked perfectly but the Y & A needed the "step active low" setting reversing in ports & pins? They sounded very rough and notchy when set the same as the other axes.

This seemed odd as all drives are the same, wired the same and connected to the same BOB????

Any ideas why I need to run two axes reversed signal level???

One difference is that the test PC was very underpowered and only ran at 25khz, the work PC runs at 45khz, step and dir pulse length is 2us on both PC's

[acourtjester]

could you have the Dir + and - wires reversed on those drivers.
Are you using the same drive ratio on the Y & A as the X axis?

[davek0974]

could you have the Dir + and - wires reversed on those drivers.
Are you using the same drive ratio on the Y & A as the X axis?

I will double-check the wiring polarity tomorrow, but is unlikely i have messed two drives up.

Ratio is slightly different but was arrived at by using the auto tuning feature and a long ruler.

[acourtjester]

Also check the switch setting for the pulses per revolution on the driver this is not the same as steps per mm (inch). It is the pulses it takes to turn the motor shaft 1 revolution and if Y and A are different they would be fighting each other (notchy)

[davek0974]

yep, done that, all dip switches are identical

both motors were moving in the correct direction just not happy until i changed the step active low setting

[davek0974]

Its the pulse width, CandCNC advised i increased it to 5us from 2us and flip the active low settings back - this fixed the issue.

Just got to measure and set the z switch offset and we're ready for cut tests

[davek0974]

Ok, we have action

Mixed bag of results, sort of exciting and disappointing all at the same time.

The good parts...
It works, speed is no issue,
The ripple effect in my cuts has disappeared,
It does not shake and wobble all over the place,

The bad parts...
The voltage is still way off of book specs, no change,
Can't get a clean cut on aluminium yet,
Need to work on divots,

Here we go...
Pierce & Cut heights are all as per book and this was checked several times by stopping mid cut and measuring the gap.

35mm washer, 45A 107v, 4000mm/min. Nasty divot as no cut rules yet just perpendicular lead in and no lead out.

IMG_1004s.jpg

Add a 2mm overburn and turn torch off 2mm before end - divots now pretty good.

IMG_1005s.jpg

Same speed, acceleration as listed, small shape again - 35mm sq. Same 2mm overburn etc but divot is terrible again????? There is a marked improvement between the 1st and 2nd cuts but little change between 2nd and 3rd so i set the acceleration at 1500mm/s/s. I can improve the internal corners a bit by setting Mach to stop CV on angles > 89 degrees but don't know if that is a good idea or not??/

IMG_1006s.jpg

3mm aluminium - 6400mm/min 45A 106v, lots of top dross, tried slower but made little difference. Had better cuts on old machine

IMG_1008s.jpg

1.2mm stainless is not too bad, 7900mm/min 45A or 4800mm/min 30A 109v

10mm steel, 0.9s delay, 117v, 500mm/min 45A.

Had one nozzle blow out completely, was getting poor cuts and found the orifice practically missing, no idea how that happened at all, as i'm testing I have been very careful of wrong pierce heights etc.

So, as i said, a mixed bag really. The divots are concerning, I need to get that bit nailed as well as the aluminium.

If anyone wants to chime in with some useful settings, please do. Is there a group test file that those with some spare minutes could cut for comparison???

[SeanP]

Glad its going well Dave, well it's looking good to me at this stage.

I wouldn't worry about voltage, just find your own settings and work with those, seen a few people mention it on the 45, even once you get a baseline for a certain thickness I seem to have to tweak +3 or 4 volts as the consumables wear usually.

The way I get around that divot on a outside cut like that is to use a lead out an hold the piece as it's about to drop, I have a old hard drive magnet on a stick to catch parts.

For holes I use the torch off, but still find it better to still use a short 2 -3mm lead out just to be sure it's moving away from the part, no overburn but have the torch off around .065mm 'before end' using that seems to cover most thicknesses really.

[acourtjester]

do a search here on divots Jim Colt has posts giving info on what to do there.
Nozzle blow out may be a bad pierce.
Working with cut rules (path rules) should help with other things. As you say the wavy cuts are fixed you may need to just fine turn your operational configuration to get it where you are happy.

[SeanP]

Cut some 3mm yesterday 5500 was best for me

[davek0974]

Glad its going well Dave, well it's looking good to me at this stage.

I wouldn't worry about voltage, just find your own settings and work with those, seen a few people mention it on the 45, even once you get a baseline for a certain thickness I seem to have to tweak +3 or 4 volts as the consumables wear usually.

The way I get around that divot on a outside cut like that is to use a lead out an hold the piece as it's about to drop, I have a old hard drive magnet on a stick to catch parts.

For holes I use the torch off, but still find it better to still use a short 2 -3mm lead out just to be sure it's moving away from the part, no overburn but have the torch off around .065mm 'before end' using that seems to cover most thicknesses really.

Thanks Sean, I can't see why some users have the voltage disparity, not yet correlated it with what THC they are running though, might be interesting to find out. If its just a CandCNC thing, it could be adjusted but as you say its not critical, just causes issues when posting on the 'net as the first thing most will say is "your voltage is off....."

Yeah, i need to make some notes and pin it on my desktop re the divots and the cures, probably the hardest part of all.

You say you have torch off 0.065mm before end? That is a microscopic distance at speed, surprising it does anything at all really but will be trying that tomorrow.

[davek0974]

do a search here on divots Jim Colt has posts giving info on what to do there.
Nozzle blow out may be a bad pierce.
Working with cut rules (path rules) should help with other things. As you say the wavy cuts are fixed you may need to just fine turn your operational configuration to get it where you are happy.

Yeah its the worst part of getting a new table

SeanP, thanks, I gather thats at 45A??? Will try again tomorrow.

[tcaudle]

How far off from chart numbers are you when you cut at the right height and feedrate with the DTHC turned off? How are you getting the voltage from the 45? Things get calibrated to less than 2%
Do you have the PWM module version ? Always start out with manual (THC OFF) cuts to get a voltage baseline and progress to turn the THC on after a cut starts. Touch-off, pierce height and cut heights are important so doing a "air cut" with the plasma off and watching that is important

[davek0974]

HI

seems to be around 10-12v lower.

I am taking the voltage direct from the CPC port on the P45 into the PWM module.

I always get my baseline by running a straight cut on the flattest sheet i have, even holding it down to avoid warp, then run a cut maybe 200mm long at the chosen speed and monitor the voltage displayed, it generally levels out in the middle 80% of the cut and thats the part i average out for my voltage.

To verify cut height i do a mid-cut stop and check with feeler gauges, I always run book height of 1.5mm cut.

Next i feed the voltage i noted into sheet cam and turn the DTHC on then run another cut and this time monitor the THC adjustments in Mach, any serious lift or drop and I would nudge the voltage accordingly then save that voltage as my set-point.

[SeanP]

SeanP, thanks, I gather thats at 45A??? Will try again tomorrow.

Yes 45a

[davek0974]

Swing-out keyboard, seems to work ok...

IMG_1011s.jpg

IMG_1010s.jpg

Big success, thanks goes to SeanP for his settings on the 3mmAluminium...

IMG_1015s.jpg

Pretty clean I think.

IMG_1016s.jpg

1.2mm Stainless, not too impressed yet, might have to get some 30A consumables and try that out, not sure I can cut my stuff unless I can get this better.

IMG_1013s.jpg

IMG_1014s.jpg

Now, the last major issue - that damn unequal bevel, here is a bit of 3mm steel with a 3mm slot, these parts need to slot together. You can see one side is good the other is bad, this aligns with the Y axis.

IMG_1017s.jpg

Heres some more, 3mm again, placed on table as in picture. I have marked the angled edges.

IMG_1012s.jpg

Now, I could fudge this by leaning the torch over to compensate but that is a bodge and will only ever get me two bevelled edges. So, what can do this, the consumables are new and today I have checked the condition of them between every cut after yesterdays blow-out. Overall the cut quality is impressive - the new machine is smooth and this is reflected in the cuts but this problem migrated from the old machine so must be torch related?????

Also, I can improve internal corners by switching "Stop CV on Angles >89" on in mach3 settings BUT this then gives me a jerk at the start of every arc and radius cut, I think this can be seen in one of the pictures above. So I can't run that setting as a default and as very few jobs are fully square, its practically pointless. Can Exact-Stop be triggered from Sheetcam??? Is it really this difficult?? What do you guys do for sharp internals???

SO many settings and tweaks, doing my head in. I think I can crack the divot problem, a lead-out plus torch off before end of cut seems to be reasonable but I can tune this over time.

This bevel issue is bad though I think.

[SeanP]

Umm those lean angles sound all to familiar, I have the same happening on mine a fair bit, not sure what the answer is, I'm the same torch and material perfectly level and upright.

Using the cv setting, done that as well, trouble is I forget now why I stopped using it, think it was the judder on curved nodes.
I got better corners on mine by getting acceleration up as high as I could.

I have a few bits to cut tomorrow, if I get chance I will try that 35mm sq just for comparison.

Like you say it does your head in all this tweaking, rules etc.

The sheetcam rules are a godsend though, on that slot would you have had thc turning off in that area? any sharp corners I tend to have thc turning off just in case.

It would be interesting to see how a 30a setup would cut the stainless.

Hows the extraction system going so far?

[davek0974]

HI Sean,

I have not touched the THC on/off yet, i just tend to let it get on with things, I think the anti-dive % setting in mach helps block a lot of corner dive, at least i don't think i have seen a corner dive yet. The THC is still not set right yet, I cut a series of small parts yesterday when testing the slot widths and CV stuff, I found i had very variable dross on the back, turned the tHC off and no dross, may have been a lucky one but does seem to be very sporadic - it will be cutting clean at the start and when the THC kicks in i get dross. To counter this I have set my voltages even lower - as these are small parts there is no plate warp and the bed is flat so the THC should be doing nothing, i set the volts on the fly so the THC Adjustments DRO reads zero, its another couple of volts down.

I am checking the consumables very regularly at the moment, don't want to be adjusting stuff because of a blown nozzle etc, so far they are in perfect condition after the last 50 or so pierces. I even run a small job - a 3d ant and it came out very nicely, a quick clean up on the slots was all it needed to fit together, cuts are smooth and when cutting a 3mm slot i even get a slither of waste out!

The extraction, apart from being very noisy, is a 100% success, even with my small 625x625mm sheet slung randomly on the bed there is more than enough flow to keep all of the muck away. One lesson i did lear on the Ant job though is to rescue any drops from my mesh platform below the slats before starting the next cut - they get covered in a hard to remove gunk and spatter

So all in all, its working just as designed, i have the speed, i have reasonable acceleration - set at about 0.3G in Mach, the cuts are smooth, the bigger table space is very nice, just need to dial-in the last parameters i think.

I also tried fudging this angularity thing by deliberately angling the torch according to the results and it does seem to help a bit, there is still bevel but at least its more even all round, I'm not happy with that bodge though and would prefer to have the torch perpendicular but at least it improves quality for the present time.

I'm not happy with the thin stainless yet either, will be doing some more work on that i think.

[davek0974]

Yesterdays little fun test...

IMG_1019s.jpg

3mm Stainless, 108v, 45A, 3800mm/min, pretty clean...
The bottom 4 lines are on the new table, the top mess is the result i had on the old table which made me build the new one!!!

IMG_1020s.jpg

Test shape, old m/c vs new m/c...
I know the slot is buggered - this was because i used the test file for mild steel and forgot to tweak the settings on the slot part first

IMG_1021s.jpg

You can at least see the waste in the slot, that a 3mm slot and the waste is about 1mm.

IMG_1022s.jpg

So, not too bad today, also been messing with divots, internal holes. Tried a 5mm overburn with the torch off 1mm before the end, so the end of the cut is 6mm from torch off and you can still see a mark. Will keep messing with that one but at least she's up and running again.

I'm leaving the torch at the current angle - it's working so until i find a reason, i'll live with it.

[SeanP]

I have a few bits to cut tomorrow, if I get chance I will try that 35mm sq just for comparison.

Sorry didn't get a chance today, away on hols in the morning, you know what it's like

You look to be making good progress though one way or another, cuts look very good on the spider thingy to me.
A fair improvement on the stainless dross also, it's not that easy cutting small bits like that.

Glad the extractor is working well,,,,just the job