HI Dave. Mach3 has anti corner dive built in. IF teh combined vel drops below the setting it locks out THCdown from working. You  still get the THC up function to prevent head collision. Most set it too low.

IF you set this too low teh head will dive due to teh THC trying to correct teh voltage as teh Vel slows down for a curve.  How far it will dive is a function of how slow teh Vel gets in that particular curve and where you have teh cutoff point set to.

Thanks, i did know that but maybe i need to revisit the settings, messing with sheet cam looks good on paper but i'm not convinced it even updates theTHC in time when running unless you switch THC off way in advance which is a bit pointless.

What sort of % is recommended in the anti-dive DRO?

Also, I gather there is no existing function that addresses my suggestion of stopping downward signals IF there is no corresponding fall in volts received???

HI Dave.

IF the arc volts go uP then the torch goes down

If the Arc volts go down the torch goes up

Anti corner lock blocks torch down ( Mach3 function)

Anti Gap lock blocks torch down across a void ( THC control function)

What else would you think it needs ?

I would start with the Corner lock at %90 it will vary as to your machine and it motion charectoristics

(;-) TP

Hi Hood - Hopefully I've worded this carefully enough to avoid a good kicking from Terry   - but here goes...

I kinda left it vague because it depends on what you want to do and what your THC can do. We can wax lyrical about speeds but ultimately you'll probably end up slowing Z down to match the capabilities of any (including mine) "budget" THC. Spend a few grand on one and the game changes entirely (so I'm told).

To give some perspective though, here's what the tables I've built in the past did (with my THC). I'm sure there are better (and worse) tables out there so I'm not saying this is good or bad - just what they were.

My THC samples at 1KHz. If it's set to do 1 step pulse per sample (which it is) then it can obviously drive the Z at 1000 pps. The Z was rack n pinion with a steps per of 42.44131816 steps/mm.

so Z moved at 1000 / 42.44131816 * 60 / 1000 m/min i.e. 1.4m/min.

Lets say then that it's cutting at 6m/min. It can therefore handle a slope of 6/1.4 = approx 13 degrees. I've yet to find this to be inadequate. Of course it means it couldn't cut corrugated - but it wasn't designed to - it was simply designed to cope with "typical" warpage.

Thanks for that Ian, not even sure what the sample rate of the THC I have is, so looks like whichever way I go it will be plenty fast.
One thing though, I wonder if I would maybe be better trying the built in THC of the CSMIO. I am not sure how it works but if it is totally internal to the CSMIO then it should be very fast. Will have to do some reading on it.

Quote from: Hood on January 27, 2016, 05:55:04 AM

With the Z I see no problems in setting an extremely high accel with the setup I propose as there will be little mass to it and the servo should  be more than capable of handling it even direct drive.

Remember with THC the Z accel is not used so it's irrelevant.

Yes I saw that mentioned earlier, what I was really getting at was I should be able to have a very fast acceleration with whichever scenario I choose for the motor/gearing/screw. So really there being no accel in Mach should not make a difference as the servo should be able to get to speed almost as quick as it is told to without falling foul of itself.
 For example I have a motor hooked up here, I command a RPM of 6000 at an accel of 6000 revs/s/s and it quite happily does it. There will be error of course but it still does it easily. Also there is no load so that would change things but 6000 rpm to -6000rpm at acccel of 6000rpm/s/s is fairly aggressive

Hood

HI Dave.

IF the arc volts go uP then the torch goes down

If the Arc volts go down the torch goes up

Anti corner lock blocks torch down ( Mach3 function)

Anti Gap lock blocks torch down across a void ( THC control function)

What else would you think it needs ?

I would start with the Corner lock at %90 it will vary as to your machine and it motion charectoristics

(;-) TP

Thanks TP, most likely my misunderstanding or the mixture of terminology.
I guess Anti corner lock is the Mach3 Anti-dive??
Anti gap lock sounds like the CandCNC Tip-Saver??
I know meant-dive is not that high so i'll play with that.

I was thinking of something that would sense the fact that the THC is asking for tip down BUT the volts are NOT changing - the torch has started riding on the plate, this can happen easily on 30A consumables as there is only a 1.5mm cut height so if you were maybe 2-3 volts or so off, it would drop onto the plate and then keep trying to drop the torch until it trips a limit or the torch falls off the mount. Just wondering

The THC cannot signal for torch down IF the voltage level has not changed. The voltage level is what drives the entire process . Mach3 has no idea what the voltage level is nor does need to know that . It simply watches the up/down signals and does what it is told to do(;-) .

Sometimes IF you get teh span volts(deadband) too wide the THC will get lazy and slow to respond . IT would be fine for thicker cuts but far too slow for thin springy sheetmetal that as you pass over a section it goes Boung  and springs straight up .500" or so. The tip will drag untill the THC can catch up.

(;-) TP

Ah, but what I mean is this...

Set volts to say 90v because thats where you reckon it should be...
Start the cut and the THC reads 98v so tells Mach "Send the torch down.."
Mach drops the torch but as it so close to the surface the volts only reach say 94v,
The THC shouts "I want more, drop the torch again..."
Mach says whatever and this loop continues with the Z dropping but the volts not changing anymore because they can't.

Sounds spurious but on thin material with the 30A consumables its very easy to do.
Mach or the THC should see the repeated DOWN calls AND the fact that volts is NOT changing and say "I'm not allowing that anymore", maybe even lift the Z a mm or so.

(;-) Mach3 cannot read a crystal ball and figure out exacty what YOU WANTED it to do. It can ONLY do what YOU TELL  it to do .  In this case it is an operator error that no machine can compensate for as you asked it to do something that it cannot physically do. I would have simply dialed  the setvolts to teh proper voltage level as soon as I saw it lower onto teh plate after the THC delay cleared.

IF you do not correct it it will continue going DOWN until it breaks the torch, trips a limit, Goes to teh THC MIN, or breaks away teh torch(magnetic holder).

I have done all of the above as well (;-).

You do WATCH new untested programs programed with SWAG settings don't you? (;-)

(;-) TP

OK That said MAYBE   there is a way.  

NOPE DON"T THINK SO as Mach3 has NO idea what teh arc volts really is in real time so there is nothing to compare to.

  TP

OK That said MAYBE  there is a way. 

  TP

You see, thats the sort of thinking I like

I know it's OP error, I just put my PLC systems programming hat on for the evening - you know, the one that tells me if it CAN happen it WILL happen AND the machine should block it if possible:)

The same as pressing the "dry run"button while running code - a simple fix in the end and a button that now does nothing when it shouldn't - the BEST kind of button in my mind

OK I don;t see HOW it could work as Mach3 has ZERO knowledge of Arc volts. It does not even know what it is. That is monitored from the THC AND the THC HAS zero KNOWLEDGE ABOUT the Z vel in Mach3.   

Nothing to compare .  Any com link between the two would be FAR too slow to react in time.

Interesting idea though, (;-) TP