New THC Feature....options [long]

[Torchhead]

Let me preface this by saying I don't normally discuss upcoming products or features before they are ready to ship, trying to generate presales marketing. I have watched others take their beatings over announcing a product that is months out and may never actually be affordable or producible.

I have been working on this project (along with others) for several months. At first, it started out as a way to do a better job of generating the ARC OK (aka Arc XFR; Okay to Move) signal from plasma units that don't have those signals available. The idea is to make an easy-to-install DC current probe that would snap around the workclamp cable and trigger a relay at a preset level. It rapidly (well, sorta rapidly) became clear that since we could measure the DC/PWM current precisely and actually read it out using the same circuitry we use to read the Arc Gap volts for our Digital Torch Height Control (DTHC), it then opens up all kinds of possibilities for added control of the plasma process.

The simplest and easiest to implement is the presentation on the screen next to the Torch Volts of the Torch Current in real time. That brings us to the crux of this post:

1. We can make the Arc OK trip point (value where it sees a valid arc) a parameter in our Cut Profiles section so it can be automatically changed to match the type material you cut. You would define that parameter just like the other presets in the middle and lower part of the popup window.

2. We can make it so the Torch Current can have fault points above and below the recommend current and define what actions the fault triggers (i.e Operator Alert, Motion pause and torch off, complete STOP etc). It would be a way to manually warn the operator they forgot to change the current after switching tips or material types.

3. We can, on a future version, add the ability for the software to actually set the cut current to match the one on the Cut Profile (that is a manual process now) That needs a lot more engineering (and possibly the cooperation of the Plasma cutter companies). One idea I had was to make a small mechanized "knob" to replace the current set knob on any plasma unit using a small RC type servo. [open for ideas on this one] Basically a motorized pot.

4. And now for the biggest question/challenge. We can make the system "smart" enough so that it would vary the current based on specific cut factors. We can measure the actual feedrate and cut direction from MACH3 and predict when a feature that needs lower current (corners, small circles, etc) is being cut and the DTHC would actually adjust the current on the plasma as it cuts. Now, I have never cut with a machine that could vary the cut parameters while cutting, so I have no frame of reference. I don't know what the range of current change should be and if the Arc Volts need to be adjusted upward (?) in sync. I do know the cut varies based on where the leading edge of the arc is contacting first.....

I need some feedback from the field. We as engineers, sometimes get dazzled with technology (ooh SHINY!!) and design stuff that is what we think the market should have, not what it can really can use! I call it the "Solution Looking for a Problem" syndrome. I'm too old and grizzled to fall too far into that pit. I have one foot that occasionally slips in.

So, if I could impose on the esteemed members to do a less than formal vote as to what features that you consider most important then we can use that to set priorites. We have 1 and 2 pretty much nailed but 3 & 4 are in the argument stage.

If you want to comment to me directly via e-mail then that is okay too.

NOTE:
This is not a stand alone product and is designed to plug into our new DTHC product line. If you have one of our new units (MP3000-DTHC or BladeRunner Dragon-Cut) and would like to be a beta tester contact me. We will need to switch out some of your cards and firmware.

TOM Caudle
www.CandCNC.com

[Karl_T]

I held back on responing cause I'm not a customer, nor a potential one.

This sounds WAY KEWL. But I'm an engineer too. I'll plan on adding a current sensor to my galil/camsoft THC. FWIW, my specialty in engineering was study of processes. For a given material, do you have a good feel for how to adjust current as velocity slows? How about a feel for optimum voltage as velocity slows? When mine runs (it will be a while) I'll study these two variables.

Karl

[Torchhead]

Well, It's one thing to be coo, but another to add a valuable function. That is where you have to remove your Engineer Hat and put on the Marketing one. In a normal company the field engineers that are at the customer's site come back and talk to the head of engineering about a problem or feature that needs to be addressed. Then the head engineer has to decide if the feature can be done without changing the cost structure too much. If it's an add-on (option) then he/she needs to get an estimate of cost. Then the sales and marketing guys get involved and they are asked what the market is and the level of resistance on price. ("what can you sell this for?"). If the numbers don't work it gets dropped. Sometimes it works the other way. A sales or marketing person either engages with a customer or they lose a sale to a competitor because theydon't have the option/feature. That causes massive pressure downward to engineering to come up with a solution and at a cost that can compete.

One of the advantages in a company with a flat organization chart like mine (I am Head Engineer, Chief Design Engineer, Director of Production, VP of World Wide Sales, COO, CEO, Chief Technical Writer, and Manager of Customer Support...oh, and I clean up on weekends) is that not a lot of time is wasted on meetings and bickering. The analysis process is shortened. That can lead back to the earlier syndrome of designing a product that has no real demand. I have, over my career fulfilled most of the roles in other companies but typically only one at a time.

I know that often the end user does not understand the technical details of plasma cutting and the parameters that make the difference between a mediocre cut and a good one.

From doing support on plasma cutting for years, I do know that the more information you have, the easier it is to find and fix a problem. Since cut current is "assumed" to be what the dial is set to, things like bad workclamp connections or something wrong in the power section is difficult to diagnose.

I don't want to waste time on advanced features (like variable cut current or automatic current setup) if the end customer sees no value or it just confuses the issue.

TOM Caudle
www.CandCNC.com

[Alan Bradford]

Hi Tom,
Hypertherm has had a similar option on their bigger machine for years.
It is an input that reduces the current by 50%.
It was thought of as a better way to make corners, but we just thought of it as another "Screen Door on a Submarine". Cute but useless.
The reduced current caused the arc lag to get worse and ronded the corners worse than normal.
Mind you it was a 50% step and not tapered to speed.
I have never used it on a customer's machine, just in engineering tests (when I worked for them in the early 90's)
A tapered reduction may work.
I dont want to discourage you, but thats my 2 cents worth.

[jimcolt]

I have played with reducing current during cornering and have had some limited success with it...especially on machines with sluggish acceleration and de-acceleration. If the machine motion has sluggish decel entering a corner....expect a dramatically wider kerf, and low speed dross....reducing the current during this event helps somewhat....although in reality you would want to reduce the current while maintaining high arc energy density....meaning you should also make the nozzle orifice smaller!

On cutting machines with reasonably good acceleration it is probably unnecessary to adjust current during cornering.....and with today's available drive technology...and good sturdy table designs with low levels of backlash...it is easy to build machines with acceleration in the 30 to 100 miligee range....which is good enough for plasma applications.

I think the best feature that dramatically improves plasma cutting performance is with accurate initial height sensing.....before every pierce. This will ensure that the torch always pierces with the proper pierce height (as specified by the torch manufacturer)....allowing for the longest consumable life. There are some users of high definition class plasma systems that continuously achieve over 3500 starts on a single set of consumable parts.....while others only get 200 to 300 starts on a set......and this can be attributed to pierce technique.

One pierce too close to the plate will damage the nozzle orifice to the point that cut quality is affected....and in my experience with customers complaining about cut quality and consumable life...helping them solve their initial height sensing issues solves their problems in most cases.

Initial height is typically achieved using either ohmic contact with the plate, stall force sensing, or with inductive or capacitive probes that sense the surface. None of these techniques are perfect for all applications...the ohmic contact does not work well when the top surface has poor conductivity (rust, primer, underwater).....the stall force sensing often deflects thinner plate, and inductive/cpacitive sensors will vary sensing height depending on different material types and surface conditions.

Many of the high end (read expensive) torch height controls use a combination of ohmic sensing....with a backup stall force sensing (if the ohmic does not work due to surface condition...or in under water cutting applications)....this type of IHS works well.....but it would be nice to have better accuracy in initial height sensing.

Jim Colt

[Torchhead]

Jim, thanks for the feedback. That is what I am looking for in turns of setting my design priorities.

I can see where knowing the actual cut current can have value; especially in troubleshooting bad cuts. The auto reduction feature was just one of those engineering things that could be done but as you say, maybe should be left to someone that has the time to do may hours of test cuts using different values of cut current at varying feedrates and maybe even vary the arc volts with it! If there were a neat chart out there that gave me numbers than the project would be pretty easy.

I think at this point we will focus on the Digital Current Probe as it is and leave the door partially open for the more advanced auto-set and auto-cut options sometime in the future.

The IHS is essential. The trade off of doing it every pierce versus at a set distance/ frequency, is in time. Since different material can be setup to do the method any way you want it gives the operator the flexibility to use the best method. I have a different post for thin material vs thick in the CAM program. In the thin I have it "touch-off" before every pierce. On the medium I have it touch off every 20 " of X Y movement (that is not a lot of movement when you are talking about the sum of both axis). In thick material that tends to be flat and level I have that set higher.

IHS in our MACH based system (because MACH is in charge of all three axis) is set in the G-code out of the CAM. It's actually a reference (Home) move on Z. While we could use ohmic sensing, force feedback or capacitive sensing to trip the Home input action the simple floating holder with a micro switch is drop dead simple, very reliable and easy to hookup and troubleshoot. It does require a custom torch mount but it can even be fitted with an up bias spring that lets you touch off on thin material without deforming it. I have several customers cutting 24 and thinner gauge material using Fine Cut consumables and getting good results (on painted R-panels).

By having information on the toolpath out of the control software and the current position of the Z, you can easily pickup feedrates, Z height, total travel etc that let you interact through the THC. It no longer has to be set separately to do different pierce heights, pierce delays, IHS before every pierce, and the arc volts presets for a given material. As long as the material stays flat the pierce height will be correct even without a touch-off before every pierce.

Thanks again for the feedback. There is no substitute for experience!

Tom Caudle
www.CandCNC.com

[DISCONNECTED]

Tom I have a question for you. Do you ever get a chance to sleep with everything you have going on?

[WSS]

This thread has some interesting reading. I have never cut using a mechanized plasma. I have two questions (maybe it will jog or bring forth some obscure, faint idea?), What is the actual range of "ohmic value"? Can this be added to or subtracted effectively by a user/manual pot? Second, why do we want to slow down in a corner? It may seem ignorant, but it is an approach from the other direction. Again, I don't have a clue what I am talking about. I can hear a Victor 3 hose torch (in particular, although we use 2 hose Victor and Harris torches as well) make a faster rip and blow sound at speed in a corner and mellow out in a straight away, which means we use a corner to set our max speed. If we use a straight away as our set point, then the corner will usually sever/bevel inward at the apex.


From Torchhead:
"From doing support on plasma cutting for years, I do know that the more information you have, the easier it is to find and fix a problem. Since cut current is "assumed" to be what the dial is set to, things like bad workclamp connections or something wrong in the power section is difficult to diagnose."

We assume we are setting a voltage (or amperage or travel speed, etc.) based on a gauge, meter or scale on a dial and that is usually on the machine or power source. Actual tip voltage could be very different even from shielding or lack of, much less grounding and other tangible hurdles. The info on one end could be literally a hundred feet away from the source.

One of our weld heads has a Laurel (or laurate or something like that) amp meter that has a "response time" setting. When we set it to it's lowest or quickest read time we can actually "see" the grind marks left by a 50 grit grind wheel in amperage "up and downs". To keep an operator from going nuts we set it at .5 second reads. There it will only tell you if your weld surface is level (generally a 1 inch wide weld and the oscillation will vary the amps by 10 amps or so if not level). I could not imagine the algorithm that would have to go into the detail on this thread.

Happy Monday!

[Torchhead]

Things happen many times faster in plasma cutting than in welding or oxy-fuel. It's the difference in speed of a Piper Cub VS an F16. The MIG and TIG and Plasma are a constant current cutting operation. That means you set the current on the power unit and because of Ohms Law that states voltage, resistance and current are locked together in a proportional balance. Change one and the others change in relation to which side of the equation they are on. If you keep one constant then if a second changes the third changes in direct proportion. In Plasma the current is set (constant) and the resistance is represented by the length of the arc so the variable if the gap changes is the voltage. Part of the challenge is that the change in arc gap to maintain an optimum cut is relatively small (.010" or less) and the resulting change in gap voltage is small. A change of one volt can be significant. Consider that the voltage might be 140VDC and that you have to sense and react to anything above 141 and below 139. It means you have to resolve a change of less than 1%. You have to do all of that in an atmosphere of noise and outside influences.

The process of cutting with plasma assumes the arc itself (the resistance in the formula) is based solely on the gap....it's not. It's also based on the area over which it is spread and the type/thickness of metal being cut. Slow the feedrate down two things happen: The arc becomes "straved" for material and start to cut more on the sides (kerf gets wider) and the voltage goes up because air represents a higher resistance than the metal. The slowdown of a CNC machine on tight corners (a function of the torque/acceleration capability of the machine) creates a condition where the cut changes and the voltage starts to go up. The THC uses the voltage like a altimeter and if it comes back and says it's too high it's job is to move the torch down to balance the equation. In a properly designed THC this is done rapidly. It creates a condition where, if you are cutting at the designed height of .063, you can see the torch move very close to if not touching the metal.

I'm not sure about your question on Ohmic value. If you are talking about the IHS (touch off) the answer is anything from fractional ohms (dead short) to high resistance (painted or dirty surfaces) The variation across a sheet of material can be fairly wide so you have to define a range rather than a value. Ohmic sensing works well on clean, dry metal. It fails on rusty, painted or heavily oiled surfaces. Because of that you have to have a backup sensing system based on another method. The usually takes the form of torque (force) sensing on the Z. That works fine if the Z motor is under direct control of the THC electronics. It could be made (with more electronics to give a digital feedback to the control software but everything you add in complexity and parts raises the price.

5 minutes response curve is no response at all. Even .5 sec in plasma will cause instability in the THC control loop. It puts you .5 second behind where you really are in the cut so at 200 IPM you are reacting to something 1.5" back you have already cut! It creates what I call a "Random Z position control" aka "Head bounce".

The concept of having the current variable, is to compensate within a range for things like a change in feedrate (which as described changes the gap voltage and "fools" the THC altimeter reading) and would automatically adjust the current in small steps to balance it back out. Even bigger is the fact the plasma cuts differently based on which direction it is cutting. Right now the options are:

1. Have something tied to the feedrate and just ignore the THC commands when the feedrate drops below a certain amount (thus preventing the torch from diving). That requires you are integrated with the control software and know what the actual feedrate is. That feature has been in MACH3 now for several years. It's called anti-dive. It works well as long as the arc gap voltage change is the result of a change in feedrate.

You can let the electronics in the THC detect if the acceptable gap volts is outside a preset range and not adjust the height below a certain point.

In our new DTHC we use both methods so you don't get head dive at the end of a cut (where you are very close to or cross the lead-in cut) or if you run close to or over a void.

Setting target (preset volts) on the fly or from g-code is not a problem. I have customers where that alone would help because they are cutting assemblies of stepped thickness (1/4" plate welded to thinner/thicker material). We are not changing the gap volts directly but setting a new value for the THC to track.

Setting the current on the fly is more of a challenge since most smaller machines don't have a way to do that electronically. Given the difficulty of the task VS the benefit (unknown or of minor value) it's not something I will invest a lot of time in. I have other CNC projects that I know need attention.


TOM Caudle
www.CandCNC.com

[Torchhead]

Originally Posted by DISCONNECTED Tom I have a question for you. Do you ever get a chance to sleep with everything you have going on?

rarely!

CNC is my life at this point. I am blessed because I get to do something for a living that I love. The "Zen" of engineering is a great "drug". I am a CNC Addict! I have (almost) always worked in a technical field. From designing power systems to Information Technology. I quit my last job making an insane amount of money, because being in IT management meant I just managed people and corporate politics. Only having one or two days a week to be creative to feed my soul was causing pain. I built my first CNC plasma cutter in 2001. I started a decorative steel cutting business in 2003.

Make no mistake. Owning and running a small business in these times is not a walk through the Field of Dreams. If it were not for the fact I really enjoy the work I would go work at Home Depot!

TOM Caudle
www.CandCNC.com

[WSS]

Tom,
Thanks for the time to respond! I have a lot to learn.

Quote:
Things happen many times faster in plasma cutting than in welding or oxy-fuel. It's the difference in speed of a Piper Cub VS an F16. The MIG and TIG and Plasma are a constant current cutting operation.

Or, we might say a Peterbilt and a Corvette, The truck can carry 40k of gold bullion in one trip while the car will certainly get there faster but take a few more trips (Gold=profit). Our mig set-ups are set to run in CV mode, they will work in CC but the mig we do needs constant voltage.

I was thinking of a THC for our mig guns. I know the signal from the welder is amps, but I think a torite coil can be wrapped around the +leg to get a voltage signal. Has anyone done this?

Thanks,
WSS

[Torchhead]

I did a THC for a MIG one time (had to rescale the voltage since it's way below normal Plasma cut voltages). It was based on constant current. The IHS was a real bear. I Guess if you had a way to read the cut current and profile and how that varies with the height (gap) then it would work pretty much the same way a arc voltage sensing.

TOM Caudle
www.CandCNC.com