Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: Davek0974 on July 29, 2014, 10:51:10 AM
-
Hi all
after doing some nice work on my little CNC plasma table, I have now got a job which uses all of its travel capacity.
The steel is a bit bowed, probably max 4mm variation so not too bad across 625mm.
What happened on the first cut was that it started out perfectly then the cut started losing penetration intermittantly, my guess was that the THC was not tracking properly the bow in the sheet and i ended up with too high a cut.
I am stuck between it being the cheapo Proma THC i bought, or a setting issue.
In Mach i have the THC anti-dive at 90% and THC Z rate at 7%.
The rate was low because before it seemed to be overshooting, i.e. the THC loop could not keep up with torch travel and was adjusting too much, too late.
I think the Z axis screw is 2mm pitch and direct drive from a small Nema23 motor.
Maybe the rate is too low, the THC is crap, or something else is not right.
As this is my first table, I have no idea how much travel a THC is supposed to compensate for so I'm not sure where to pin the blame.
Any tips would be much appreciated.
-
Hmm, just found this on a site called cncarena....
"
Once you have it working you will be disappointed in the ability to cut the thin stuff. It's just not capable of responding or moving fast enough to keep up with warps while you cut at over 200 IPM .
The pure speed of MACH THC keeps practical response down to about 60 IPM in Z speeds on a fast Z. Factor in the added delay of the Proma that uses relays on the Up and DOWN and you will be luck to get stable operation at 30 IPM or less. Think about flying a drone at 6 ft above the ground over hilly county and your feedback is delayed by 1 sec!
The system will work on very gentle rises like on thicker material and at much lower feedrates."
Now, 30ipm is only about 762mm/min and I'm cutting way above that, could this really be the issue?
-
Well it looks like the THC is just no good, if i tune it for slow cuts it works well, but could never manage a cut at 3300mm/min, pretty certain if up the tuning to go faster it will bunny-hop on slow cuts due to overcompensating. :(
Anyone recommend a THC ?
I have tried contacting CandCNC but had no response yet.
I had this reply from Proma....
"THC 150 has about 150ms (0,1 - 0,15s ) processing time, o don't sure what many time need mach for reaction.
total time isn't very well for it. I think that better is use model SD (step dir) or full system MD+DV. both systems works alone (without software) and total delay time is about 70-80ms - for high speed still it's not very fast but better that compact THC."
Sounds like speed is not their thing?
-
I have a Proma THC and it keeps up to my machine when it works. I don't cut faster than 120 IPM, but if I can get Mach3 to move the Z the THC keeps up great! My problem is getting Mach to move the Z every time THC commands it. It work intermittently. What exactly did you setup in Mach? Where is the THC anti dive & THC Z rate settings? I've played with this for months already but maybe I'm missing something?!! I hope you find something that works reliably and keep us updated! Thanks, Brian
-
Hi
I have rate set at 9 and anti dive at 90%
I tried lowering anti dive but makes no difference, might try turning it off.
-
i have never regretted the lcthc from candcnc. basic height controller that just works and not very expensive.
-
But does it work at speed too?
I have tried contacting them for two days now but no response at all.
No objection to buying one if it works but can't do anything without a reply ;)
-
Still no reply from CandCNC, are they still going?
I did get one from a guy going by the name of randyray but he said he could not help interface it to my torch as it's not for automation.
-
i cut up to 240 ipm on 12 ga .100 thick . sheets are 170x60. your speed of the height controller is set in mach3 as a percentage of the z jog speed. my z jog is set at 200ipm
and is set at 7%. oxy flame cutting is set at 30% as height control is being controlled with a joystick. while you are at it get the feathertouch ohmic contact sensor.
candcnc probably is on vacation.
-
Thanks, just got back from a break away, will be picking this up again soon..
-
A THC loop of 150ms is pitifully slow. I can't see you getting much in the way of performance. All feedback loop systems operate behind the curve but apart from anything else - at 3300mm/s feed you're going to be a massive 8.25mm behind. I know you don't have the right system to use my THC but just to add perspective - at 3300mm/s feed mine would be just 0.055mm behind the curve (1 ms loop).
-
1ms is amazing!
They do another unit with a response of around 80ms I think which controls the z motor directly and not through Mach, the issue with using that is that you lose the antidive feature in mach3.
What was the reason I couldn't use your system?? (Can't remember)
-
There appears to be a belief with some THC manufacturers that Mach's internal THC functionality presents some sort of bottle-neck and hence they tout "standalone" as some sort of remedy to this. Some even blame Mach's "slowness" for causing "head bouncing". This is NOT the case. Mach's internal functionality simply responds to the up down signals sent by the THC hardware and it does this at KERNEL FREQUENCY so with even the lowest kernel of 25KHz it's responding within 40 MICRO seconds - orders of magnitude faster than any THC loop that I know of.
The problem is that SOME THCs are firstly slow and secondly not very smart or even smart at all in how they "condition" the signal. Mach will just do what it's told and if its not told very well it's hardly fair to blame Mach. It's "worrying" is it not, that some THC manufacturers out there clearly don't really understand how Mach's THC functionality works.
Note - this is not to totally dismiss "standalone" THCs - they have their uses - the main one being that they can more easily be made system independent - i.e. they'll work across say Mach3/4/LinuxCNC/whatever - but siting Mach's "slowness" is not one of them and is plain wrong. (I'm currently debating whether to do a standalone version of my system for this reason of system flexibility).
You can't use my system because your Hypertherm model doesn't support the Hypertherm robotics interface.
-
Thanks, I would never infer that Mach is the bottleneck, and also If i were to produce a unit like a THC i make damn sure I knew how Mach would react when connected.
Anyway, if i were to say change to a Hypertherm PM45 Would that help? I believe that is a "proper" cnc plasma cutter????
Having seen how useful this machine is, I am now looking at upgrades.
-
Yes - the P45 is the lowest power/price Hypertherm system that has the machine interface which my THC is designed to use. Just note that if you decide to go that route you'll also need the machine torch. See www.razordance.co.uk/THC.htm (http://www.razordance.co.uk/THC.htm) for more details.
That said, I note from an xml you posted here ages ago that unless things have changed your Z axis isn't the fastest on the block so that's going to limit you to some extent whatever system you use.
-
Thanks Stirling.
I think the z axis is on a 2mm screw pitch, it was a pre-built z axis system, presumed it was "normal" ?
According to hypertherm guru Jim Colt, at the same power level, on the same material, the 30XP cutter will give the best results, so for the present at least I am sticking with that system.
I will have a further play around with the THC system.
-
Hi Sterling,
I am getting a Hypertherm HD3070. Is your THC able to work with that? I am using an Ethernet Smooth Stepper and I use Ohmic height sensing with a floating head. any thoughts?
Thanks, Brian
-
Hi Brian
Never tried it on high def cutters but according to their manual (p246) the cnc interface on Hypertherm's high def cutters is completely different to their "standard" def cutters - so I'm pretty sure the answer is a no. TBH - if I were looking at high def - I'd be looking beyond ALL the "budget" THCs. Hypertherm's offering costs several grand for a reason.
-
Thanks Stirling.
I think the z axis is on a 2mm screw pitch, it was a pre-built z axis system, presumed it was "normal" ?
So just do the maths. How fast does your motor have to spin to handle the slope in the metal you want to handle at the XY feedrate you want to cut at.
-
Ouch, maths! ;)
What i think is the issue, is that the motor will spin at whatever rate i have in THC Rate (currently 5) for as long as the THC box sends the up signal.
But, with the bad loop-speed it has, it is telling the system UP for too long due to the loop response. This makes the system overshoot so it then tells it DOWN straight away and we are into see-saw country in no time at all.
To counter this I have set the THC Rate low BUT this now means the system only works for thick/slow cuts, going to thinner metal means more speed which means I need a faster THC response rate so I up the THC Rate setting a bit but then that only works on fast cuts.
I think it's just a crap THC box really, there is plenty more speed available in system, I just can not use it properly to my best advantage.
Does that make sense???
-
Thanks for the reply Sterling! The unit I'm getting is an older machine, only paid $2500 for it so a "couple grand" for a THC is kinda out of the books for now! I'll keep open for ideas, maybe come up with a solution yet! Thanks again, Brian
-
I forgot about the hysteresis setting in the THC too, I had it set for +/-1v but having just made some straight cuts without THC it seems the voltage does not settle at a fixd point for any length of time and +/-1v is not a wide enough envelope for it.
I now have it set at +/-2v which should give it some leeway hopefully, out of the box it comes set at +/-4v which seems a bit too wide.
A faster loop response seems to be the only way of getting full control though, but at least I'm getting a handle on how it all works together now.
-
@Dave - I was just attempting to get you to do the numbers so that you can actually predict what will happen. Every feedback loop needs smarts to cope with the delay. First approach should be to reduce the delay as much as poss but after that the longer the delay the more smarts are needed.
Back to the maths though for your system. That 5 in THC rate means max Z speed under THC will be 5% of what you have in motor tuning. So in your case 55mm/min. At (say) 3300mm/min feed just as an example, that means the max slope you could even hope for - all other things being ideal - is just under 1 degree.
@Brian - no worries.
-
Interesting, not sure I could figure out how to calculate that :)
The Z axis motor in motor tuning is set at 1187.8mm/min so with the 5% THC rate it's running at 55.89mm/min.
I think that gives me a z axis travel of 0.93mm per second, and with the other axes possibly moving at 55mm/sec (3300mm/min) I can now see that I have a height change possibility of 0.93mm over a distance of 55mm.
I think that equates to a warp factor of 5.28mm halfway along a sheet of steel 625mm wide.
What I don't know is is that bad, useless, reasonable etc?
5mm of warp seems quite believable on a plate that size.
If I am correct on all of my assumptions above, maybe I have been testing it on too extreme of an incline??
-
I think that gives me a z axis travel of 0.93mm per second, and with the other axes possibly moving at 55mm/sec (3300mm/min) I can now see that I have a height change possibility of 0.93mm over a distance of 55mm.
I think that equates to a warp factor of 5.28mm halfway along a sheet of steel 625mm wide.
Correct - or to put it another way - a slope of 0.967977387 degrees i.e. just under 1 degree ;)
What I don't know is is that bad, useless, reasonable etc?
5mm of warp seems quite believable on a plate that size.
If I am correct on all of my assumptions above, maybe I have been testing it on too extreme of an incline??
If you can cut what you need to cut - then your THC is great - if you can't it isn't... ;)
-
Good stuff, at least I'm on the right track.
I spent a fair bit of time playing this afternoon, using up scrap and doing repeat straight cuts with varying settings.
I think I have managed to find a set of figures that allow it to follow a a warp of 8mm across 620mm quite closely and at speeds from 900 to 3300mm/min so I think it will function ok if I keep an eye on warp limits etc
It's better than it was at least.
THC Rate is now at 9% and Hysteresis is +\-2v