Hello Guest it is August 24, 2019, 08:04:18 AM

Author Topic: Physical buttons for plasma  (Read 87438 times)

0 Members and 1 Guest are viewing this topic.

Offline beefy

*
  •  138 138
    • View Profile
Re: Physical buttons for plasma
« Reply #90 on: January 26, 2016, 05:08:27 PM »
The Modbus in the THC is just a way of getting the Up/Down. Arc Ok etc to/from Mach instead of going via the parallel port or  motion controllers I/O.

As said I asked about it as it would have been an easy hook up but the guy said it wasn't that fast and not really suitable for thin material, so I never investigated it further.
Hood

I asked Jim Colt about sampling / THC signaling rate and he said most good THCs sample at least 25 times per second. I'm guessing you'd have to use a brain or macro pump with Mach3 to use Modbus so that limits the sample rate to a max of 10 times a second I believe, much slower than what Jim said. I'm working on my own THC and I have the sample rate at 32 times/sec which I could make a lot faster if I wanted. That doesn't seem fast in the world of electronics and microprocessors but at the end of the day the UP/DOWN output is controlling a mechanical device (Z axis) so it's own weight / inertia will limit it's reaction time. I hooked up the output to a terminal window and when you see the volt readings spewing on the screen at 32 times/second, and you imagine your Z axis ATTEMPTING to change direction at that rate, I don't think you need any faster.

My guess would be that the important bit would be once your system receives the THC UP or THC DOWN signal, how fast can it move. Mach is a problem there because the THC moves have no acceleration / deceleration in them so your THC RATE  setting is only a percentage of your Z rapid speed.
Keith.
« Last Edit: January 26, 2016, 05:12:41 PM by beefy »
Re: Physical buttons for plasma
« Reply #91 on: January 26, 2016, 05:17:36 PM »
  Mach is a problem there because the THC moves have no acceleration / deceleration in them so your THC RATE  setting is only a percentage of your Z rapid speed.
Keith.

That is interesting.
How would a stepper motor deal with that without missing steps ?

Just curious

Offline beefy

*
  •  138 138
    • View Profile
Re: Physical buttons for plasma
« Reply #92 on: January 26, 2016, 05:27:23 PM »
It deals with it simply be having the THC RATE set a lot lower than the rapid speed. I think my THC RATE is set at 15 percent.

A lot of people don't seem to understand that the Z axis design is also very important for faster THC speeds (with Mach3 and no accel/decel on the Z). You want as little inertia (weight) as possible so that the stepper is moving as little weight as possible. Linear bearings want to be very free moving, no stickiness. A weight counteract spring for the UP movements should help. At the end of the day you are telling the Z axis to move UP then having a change of mind and saying now go down, all as fast as possible. A heavy Z axis won't fare as well like that.

I'd hazard a guess a servo motor may help matters a little because I THINK they allow for some error from commanded movement to actual movement and will increase the current in an attempt to make the motor catch up. Not so with a stepper, try and make it do too much too quickly and it just squeals and stalls.

Keith.
Re: Physical buttons for plasma
« Reply #93 on: January 26, 2016, 06:10:51 PM »
Terry, thanks for that on the RS485.  (yes its just for the torch amps and pressure, yes I'm probably being lazy as that can all be adjusted on the front panel, but you could in theory include it within the gcode as a #var for example so that if you use say the gcode files intermittently, all your settings would be there from last time).

I did suspect that the main board just interfaces with RS232 to the interface board, which has a MAX488 or MAX490 on it which provides 4 wire modbus + ground tap https://www.maximintegrated.com/en/products/interface/transceivers/MAX488.html#popuppdf



Keith,

THC.... (I know Terry pulled me up on another post a little while ago (reply 18, in my response to reply 16 here http://www.machsupport.com/forum/index.php/topic,31649.10.html) about this but I still can't get this one out of my head as potentially being a better idea.... bear with me a little before shooting me down [I couldn't keep up with the speed of responses on that one to provide a response and put forward my thoughts below]).

I've always had a problem with THC shunting the Z axis up and down and also the potential for the Z axis to stall and this obviously provides the THC limit of z axis motion.  I can appreciate what Terry said, and your post above, but as you've pointed out your maximum Z axis velocity with instant (infinite) acceleration is 15% of your potential Z axis velocity.

Below is an extract of a macropump that I set up to try to increase DRO25 which is where the THC speed parameter is held, and consider this as a simple way to increase the Z axis velocity so that if the torch is not able to accelerate fast enough upwards (after 0.1 seconds [10hz macro pump speed]), DRO 25 is increased given the velocity of the z could increase further.

As Terry said in his post, "acceleration in King".... unfortunately DRO25 is not able to make acceleration king as you only have one chunk of acceleration.... from zero to 15% of full z axis velocity in.... but after 0.1 seconds if the z is still going up.... you have no more acceleration.

say like my Z your maximum velocity is 1500mm/min (60"/min) .... not sure what everyone else is running but curious to know [my z is a 2mm pitch acme leadscrew in a delrin nut].... thinking about changing it to a 10mm pitch ballscrew.

at 20%, the maximum z axis THC velocity is 300mm/min (12"/min)..... or  5mm/sec.... say my steps / unit is 200..., that gives a pulse width per step of 1milisecond. ... so as soon as the THC receives a signal for THUP, it will step at 1kHz (5x200), and after 1 millisecond the z will be moving at 5mm/sec.... the catch is after 10 milliseconds it will still only be moving at 5mm/sec and travelled 0.5mm.

However, if we increase DRO25 to say 50% of Z axis maximum velocity for the next 10miliseconds (macropump loop), it will now be doing twice the speed after 11 milliseconds than it was doing after 10miliseconds.... hence after another 10 milliseconds it will have travelled 1.5mm .... and not just the 1mm that the z would have travelled without the "improvised acceleration curve".

Code: [Select]
THCCorrDRO = GetOEMDRO(25)
THCCorrDROOld = GetOEMDRO(1100)
THCCorrDROMin = GetOEMDRO(1101)
THCCorrDROMax = GetOEMDRO(1102)
THCCorrDROInc = GetOEMDRO(1103)

TorchEn = GetOEMLED(852)

THCEn = GetOEMLED(24)

THCLedOk = GetOEMLED(36)
THCLedUp = GetOEMLED(37)
THCLedDn = GetOEMLED(38)

Dim THCLedUpOld
Dim THCLedDnOld
Dim THCUpDn
Dim SwitchCase


' >>> add in if statement to check current THC UP & DN pins status
If THCLedUp And THCLedUpOld Then
THCUpDn = 1
ElseIf THCLedDn And THCLedDnOld Then
THCUpDn = 1
Elseif Not(THCLedUpOld) And Not(THCLedDnOld) Then
THCUpDn = 1
Elseif Not(THCLedDn) And Not(THCLedUp) Then
THCUpDn = 0
Else
THCUpDn = 0
End If


' >>> Main Loop
If isEStop() Or isStopped() Or Not(THCEn) Or Not(THCUpDn) Then
SwitchCase = 0 'set THCDRO to Min Speed Setting

' increment THCDRO up to Max Speed Setting, by increment amount.
ElseIf THCUpDn And THCCorrDRO < THCCorrDROMax Then
SwitchCase = 1

' set THCDRO to Max Speed Setting
Elseif Not(invTHCMovement) And THCCorrDRO >= THCCorrDROMax Then
SwitchCase = 2

Else
End If



Select Case SwitchCase
Case 0
SetOEMDRO(25,THCCorrDROMin)
Case 1
SetOEMDRO(25, THCCorrDRO + THCCorrDROInc)
Case 2
SetOEMDRO(25, THCCorrDROMax)
Case Else
SetOEMDRO(25,THCCorrDROMin)
End Select



' >>> Update Old Info With Current Info
THCLedUpOld = THCLedUp
THCLedDnOld = THCLedDn


 



Again, it would be interesting to know what everyone else is running z axis velocity wise (is 60"/min fast or slow?).... although with DRO25 (THC Max speed.... its obviously running slower than that, think mine stalls at about 45% so I run it at most about 30% off hand), never tested it for missed steps, although do they matter as all it is doing is tracking the Torch Voltage, so if it looses a step (or a few) it will just correct the other way and when you next touch off between cuts, the Z axis DRO is zero'd (G28.1 Z).

Thanks

Rob
Rob

Albert Einstein ― “If you can't explain it to a six year old, you don't understand it yourself.”

Offline Hood

*
  •  25,849 25,849
  • Carnoustie, Scotland
    • View Profile
Re: Physical buttons for plasma
« Reply #94 on: January 26, 2016, 07:04:38 PM »

Again, it would be interesting to know what everyone else is running z axis velocity wise (is 60"/min fast or slow?).... although with DRO25 (THC Max speed.... its obviously running slower than that, think mine stalls at about 45% so I run it at most about 30% off hand), never tested it for missed steps, although do they matter as all it is doing is tracking the Torch Voltage, so if it looses a step (or a few) it will just correct the other way and when you next touch off between cuts, the Z axis DRO is zero'd (G28.1 Z).

Thanks

Rob

Would be interested to hear that also.
I will be using a servo on my Z and it has a rated RPM of 4000. I have several options.
Direct drive on a 10mm pitch screw thus giving 40m/min
Direct drive on 5mm screw thus giving 20m/min
4:1 gearbox with 10mm screw giving 10m/min
4:1 gearbox with 5mm screw giving 5m/min.

I was going to go for the last option as I thought 5m/min would be plenty but maybe not?

Hood

Offline beefy

*
  •  138 138
    • View Profile
Re: Physical buttons for plasma
« Reply #95 on: January 26, 2016, 07:32:37 PM »
Rob,

I like the concept but have no idea if it's feasible. I haven't had a great deal to do with macropumps and don't know if there is any reason it won't work. But like I said I like your way of thinking, and I'm going to keep that idea in mind.

Initially I was thinking, "Ah but my THC will run at 32 times/sec", then realised that won't matter. If the THC UP or down signal is still there after 1/10th sec, then give it some help.

I hope it works  ;D

Ah, just had a thought of a potential spanner in the works is when it stops at the higher speed. Let's say the macropump bumps the speed up then the SET VOLTS is reached, so the THC UP/DOWN signal is removed. Mach3 now will stop the Z without any deceleration, possibly causing the stepper to slip and lose it's holding torque and maybe stall. Worse case is it will stall and won't drive properly, or even when going down will keep going into the workpeice. Best case is you just lose position but the stepper keeps on doing corrections.

So this lack of acceleration may cut both ways. It limits your max instantaneous speed PLUS it MAY limit your max instantaneous stopping speed.

Something to keep in mind.

Keith.
« Last Edit: January 26, 2016, 07:46:12 PM by beefy »

Offline BR549

*
  •  6,886 6,886
    • View Profile
Re: Physical buttons for plasma
« Reply #96 on: January 26, 2016, 08:03:21 PM »
HI Keith , The problem with trying to lower teh amps is teh nozzles are based on teh rated current to establish a high density arc. Teh shap of teh nozzle cone and teh air pressure help to compact teh arc into a much denser arc great for cutting. When you just lower the amps the arc can get weak and lazy. High end cutter can vary teh gas % and gas pressures and modify teh charectoristics of teh current to keep a dense arc going. Low end plasma units like we use it cannot do that so you basically just end up with a weaker lazy arc. AND that is not what you really want.

IF you think about it Plasma cutting depends on a constant velocity to cut best. It is just the nature of the beast. With low end cutting the best you can do is set your anti corner dive to come in early to lock the THC down and just let it run through teh corner(slowing down naturally ) then pick back up after the corner.  Trying to manipulate all teh other variables on the fly WHEN the plasma unit was never designed to USE those variables is a beast at best and normally does NOT show an improved cut.

You would better off spending time to optimize teh Aceeleration curve of teh machine rather than relying on magic monkey dust.  

Next thing to work on is seeing IF there is an alternate cut setting that works at a lower feedspeed where your machine can better deal with corners without slowing down as much( % wise) as with high speeds.  

Then work on your tool pathing to optimize it for sharp corners.

(;-) TP

Offline beefy

*
  •  138 138
    • View Profile
Re: Physical buttons for plasma
« Reply #97 on: January 27, 2016, 12:43:51 AM »
Thanks Terry,

some time ago I tried using lower amps on Finecut nozzles and the cut quality was quite bad. Maybe that was an example of what you are telling me.

Keith.

Offline Davek0974

*
  •  2,577 2,577
    • View Profile
    • DD Metal Products Ltd
Re: Physical buttons for plasma
« Reply #98 on: January 27, 2016, 02:42:43 AM »
Love this thread, great minds at work here:)

I have spent many (too many) hours messing with cut current on my PM45 the end result is that i just leave it at 45A! It really cuts poorly at lower amps on the 45A nozzle, the 30A nozzle works best only at 30A.

I have however discovered I can drop my speed by over 1000mm/min on aluminium and still get good cuts so the window for a nice cut is quite wide it seems, this helps my machine a lot, even though I took time to design it as rigid as possible it sill wobbles on fast moves which shows in the cuts.

BTW my THC rate is 20% with a max Z of 1300mm/min
Bridgeport Mill, Mach3 V062, CSMIO-IP/A controller, AC Servo Drives

Offline stirling

*
  • *
  •  2,188 2,188
  • UK
    • View Profile
    • www.razordance.co.uk
Re: Physical buttons for plasma
« Reply #99 on: January 27, 2016, 05:36:20 AM »
Acceleration in THC controlled Z is not missing by fault, it's missing by design. Trying to put it back is missing the point.

The reason THCs have no acceleration is twofold. Firstly, if you have need or time or room to accelerate then you're already hopelessly out of position anyway so the game's already over. Secondly if you accelerate you also have to decelerate. THC MUST be able to CHANGE direction as quickly as is possible. The only way you can do that is to have NO deceleration ramp. If you have no deceleration then you CAN'T have acceleration.

Re: Speed design of Z: Let's assume your Z can move WITHOUT accel as fast as the fastest feedrate in the XY plane you'll ever need. In that case your system can follow a 45 degree slope in Z (maximum).

First question is: Do you need to be able to do this? - adjust accordingly.
Second question: Is your THC response fast enough to do this? - adjust accordingly.