CAUTION this is full of tecnobabble and may cause pain and discomfort to readers not prone to wearing a helmet with a propeller on top Take in small doses
.
Accuracy is a rather nebulous term. Do you mean how close it tracks to the material surface?. Do you mean how fast it can do that? The Hypertherm has to track closer than some other types (.06 ) so you have less tolerance to play with. A volt of Arc volt change typically represents about .012 to .015 inches of change . To compare apples here you need to look at the MP3100 (an MP3000 with the new DTHCIV) the Ethercut is not just a DTHCIV but has a technology similar to the ESS built in for motion control. The DTHCIV does not even use the Ethernet for its interface to MACH. It uses an RS485 port. We use an A-D converter at the plasma that isolates and makes the voltage pickup noise immune. The DTHCIV has a resolution of .06 volts. (DTHCII has .25Volts) if 1 volt of change = .015 in theory .25 volt = ,00375 inches and .06 equals .0009. You have to realize the DC volts coming back off the torch are not nice and clean so it has to be sampled and averaged. The D part of the PID (derivative) is a way to average the changes. Already built into the system is averaging in the form of LC and RC filters Then you get into loop gain equations. The problem with the older DTHCII that used the internal MACH THC logic and its motor control was there was an excessive loop delay and NO PID so you were limited to about 30 to 60 IPM top Z speed under THC control. So can the DTHCIV hold .005 tolerance while cutting at 300 IPM? It could if the feedback voltage was that stable. It can move a stepper at over 1000RPM (200 IPM with a 5 TPI leadscrew) and servos even faster. It's up to 10 times faster and 4 times more accurate. Will it result in a better cut? Not on thicker flat material. Some where between the Promo with relays for the signals back to MACH on the parallel port (about 1/2 the speed of the DTHCII ) and the ultra high speed DTHCIV there is a point that the average user gets no added benefit..
The MP3000-DTHCII continues to sell and work well for a lot of users that do not have need for the ultra response and are interested in a THC that can make good cuts in a wide range of material from 16 ga to 1". We sell a lot of Ether-Cuts not because of the DTHCIV but because of the better interface (Ethernet) to the PC than the legacy 25 pin LPT port for the step & dir, And the wider options of OS versions it will support Will you get a better cut with an EtherCut than an MP3100-DTHCIV? Nope! Does it have more features Nope! Will it let you use a laptop 35 ft away (or more) from your controller? YOU BET! Wilt it let you use WIN 7 64 bit? Yes again!
The concept you can hold +- .0025 " of accuracy is hogwash (technical term for hard to measure and quantity). Heck, from speed changes alone the voltage can vary over a volt at the same gap because arc gap is not the only factor in Torch Volts. Try it some time. Turn the THC off, level the materiel and make some straight cuts at the exact .060 height. Now cut faster and slower on the same material and watch the voltage. Then throw in some arcs, corners and tight turns. If even a straight cut holds that steady you have a very flat and homogenous piece of metal!. Even the same piece of metal can have areas of different density Considering plasma cutting is the process of vaporizing metal with a 30,000 deg arc and blowing it out with air with a flame that is neither rigid nor straight sided.
I am surprised that you being a numbers and detail engineering type you did not pickup on the HyT-Connect RS485 option we offer to let you control your Hyperthem's Cut Current and air pressure. You can do things like a "soft pierce" on thicker metal at the pierce at a lower current and slow down the plunge to reduce splash and trauma on the consumables. Users report 3X life increases on consumables where JOBS there are lots of pierce points) or to do a "peck punch " with the plasma that just pecks a mark on the surface without blowing holes all of the say through for center punching
The reference to the ESS and the delay is from this:
MACH reads the G-Code and sends "trajectory commands" to the ESS that fills a buffer and the ESS runs the actual motion out of the buffer at its rate. Where MACH is and where the actual machine is are different. So if you read a special command to turn the THC on or off in GCode and it gets processed in "MACH time" (the ESS does not process M codes directly) it happens BEFORE you are really in the spot you want it too. The problem is the way the ESS works with MACH3. For a person that deals in tiny increments and great detail I would think that problem might hold your curiosity .This becomes obvious if you need dynamic precisely timed DTHC OFF and ON.during a cut
Another issue is feedhold and how the loss of arc is handled. Somehow the THC has to stop motion and either use feedhold or STOP. Stop dumps the buffer and loses position. Feedhold issued to MACH results in some crazy moves with the ESS because it is not in charge of the ESS buffer . None of this is a problem with a parallel port system because there is no delay between code execution and the pulses being sent to the motors
I am sure at this point the average CNCZoner has a glaze over their eyes. 95% of my customers are small to medium shops and are welders, fabricators and tradesmen that don't know a megacycle from a unicycle but they can sure weld a lot better than I can!. So raw specs don't mean much. in fact the guys that sell the single knob analog THCs, call their stuff an " Automatic Torch Height Control" too! too!. It has become a check list item on a table with NO details. . Some famous vendors sell the "dream" of what you can
make with your new table rather than give in-depth technical details. They understand their market and the potential buyers!
TOMcaudle
www.candcnc.com